Dr. Yoke Muelgini, M.Sc. Jurusan Ekonomi dan Studi Pembangunan Fakultas Ekonomi dan Bisnis Universitas Lampung 2012 The Recipe for a Successful Thesis in Applied Economics Department of Economics and Development Studies, Faculty of Economics and Business, University of Lampung J u r u s a n E k o n o m i d a n S t u d i P e m b a n g u n a n F a k u l t a s E k o n o m i d a n B i s n i s U n i v e r s i t a s L a m p u n g
Dr. Yoke Muelgini, M.Sc. Jurusan Ekonomi dan Studi Pembangunan Fakultas Ekonomi dan Bisnis Universitas Lampung 2012 FEB Unila Course on ESP 434 Monetary.
Jan 13, 2016
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Dr Yoke Muelgini MSc
Jurusan Ekonomi dan Studi PembangunanFakultas Ekonomi dan Bisnis Universitas Lampung
2012
The Recipe for a Successful Thesis in Applied Economics
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
B
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi d
an S
tud
i Pem
ban
gu
nan
Faku
ltas Eko
no
mi d
an B
isnis
Un
iversitas Lam
pu
ng
2copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
3copy Dr Yoke Muelgini MSc FEB Unila 2012
If you are the next Paul Samuelson and will wholly transform the field of economics pay no heed If you are the next Ken Arrow or Robert Lucas J and will invent a new branch of economics these notes are not for you
This paper provides tips of things to think of when writing a script or a thesis or a dissertation in applied economics One of the most important ingredients in the recipe for success is to get an early start and to begin looking for data as soon as possible
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
4copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
5copy Dr Yoke Muelgini MSc FEB Unila 2012
What is ldquoresearchrdquo The traditional view is that research is ldquothe search for knowledgerdquo which reflects the idea that science should be completely objective The researcher simply collects the data which then ldquospeak for themselvesrdquo Figurly speaking knowledge is fruit on a tree in the forrest and all the researcher needs to do is to discover the tree and collect the knowledge (Greenlaw 2006)
Of course this is not entirely true What is missing here is a distinction between knowledge and facts Knowledge is our common understanding of how things work and if we do not know exactly it is our best guess Facts on the other hand are just data In other words knowledge is facts with meaning it is the researcherrsquos best interpretation of the facts and research is the creation of such knowledgeD
epar
tmen
t o
f E
con
om
ics
and
Dev
elo
pm
ent
Stu
die
s F
acu
lty
of
Eco
no
mic
s an
d b
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
6copy Dr Yoke Muelgini MSc FEB Unila 2012
Knowledge is a dialog between researchers as they strive to strengthen their interpretations of the facts Through this dialog weaker interpretations or arguments are gradually abandoned while the stronger ones are refined and made even better Weaker arguments are those with less supporting evidence Since the data are also the facts having weak support means that the argument does not fit the facts and that it must be flawed in one way or another
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
7copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
8copy Dr Yoke Muelgini MSc FEB Unila 2012
It is important to begin thinking about these items right away before you begin the actual thesis work so that once you begin you know what to do Do not just start to think about the question Research is a highly nonlinear process that must be carried out in an iterative rather than a step-wise fashion For example although you need to begin to think about the topic (step 1) and question (step 2) early you also need to think about the availability of data (step 4) If there are no data then you can just as well start looking for something else In fact as soon as you have just a partial idea of what to do you should begin to look for data Similarly in formulating the hypothesis (step 3) you need to know what kind of data you have (step 4)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
9copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topicbull There are at least three keys to choosing
a topic1 Try to pick a field that you find interesting
andor that you know something about
2 Make sure that data are readily available with a reasonable sample (we suggest at least 25 observations)
3 Make sure that there is some substance to your topicndash Avoid topics that are purely descriptive or
virtually tautological in nature
ndash Instead look for topics that address an inherently interesting economic or behavioral question or choiceD
epar
tmen
t o
f E
con
om
ics
and
Dev
elo
pm
ent
Stu
die
s F
acu
lty
of
Eco
no
mic
s an
d b
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
10copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topic (cont)
bull Places to lookndash Your textbooks and notes from previous economics classes
ndash Economics journals
bull For example Table 111 contains a list of the journals cited so far in this textbook (in order of the frequency of citation)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
11copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
12copy Dr Yoke Muelgini MSc FEB Unila 2012
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
13copy Dr Yoke Muelgini MSc FEB Unila 2012
Table 111a Sources of Potential Topic Ideas
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
14copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
2copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
3copy Dr Yoke Muelgini MSc FEB Unila 2012
If you are the next Paul Samuelson and will wholly transform the field of economics pay no heed If you are the next Ken Arrow or Robert Lucas J and will invent a new branch of economics these notes are not for you
This paper provides tips of things to think of when writing a script or a thesis or a dissertation in applied economics One of the most important ingredients in the recipe for success is to get an early start and to begin looking for data as soon as possible
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
4copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
5copy Dr Yoke Muelgini MSc FEB Unila 2012
What is ldquoresearchrdquo The traditional view is that research is ldquothe search for knowledgerdquo which reflects the idea that science should be completely objective The researcher simply collects the data which then ldquospeak for themselvesrdquo Figurly speaking knowledge is fruit on a tree in the forrest and all the researcher needs to do is to discover the tree and collect the knowledge (Greenlaw 2006)
Of course this is not entirely true What is missing here is a distinction between knowledge and facts Knowledge is our common understanding of how things work and if we do not know exactly it is our best guess Facts on the other hand are just data In other words knowledge is facts with meaning it is the researcherrsquos best interpretation of the facts and research is the creation of such knowledgeD
epar
tmen
t o
f E
con
om
ics
and
Dev
elo
pm
ent
Stu
die
s F
acu
lty
of
Eco
no
mic
s an
d b
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
6copy Dr Yoke Muelgini MSc FEB Unila 2012
Knowledge is a dialog between researchers as they strive to strengthen their interpretations of the facts Through this dialog weaker interpretations or arguments are gradually abandoned while the stronger ones are refined and made even better Weaker arguments are those with less supporting evidence Since the data are also the facts having weak support means that the argument does not fit the facts and that it must be flawed in one way or another
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
7copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
8copy Dr Yoke Muelgini MSc FEB Unila 2012
It is important to begin thinking about these items right away before you begin the actual thesis work so that once you begin you know what to do Do not just start to think about the question Research is a highly nonlinear process that must be carried out in an iterative rather than a step-wise fashion For example although you need to begin to think about the topic (step 1) and question (step 2) early you also need to think about the availability of data (step 4) If there are no data then you can just as well start looking for something else In fact as soon as you have just a partial idea of what to do you should begin to look for data Similarly in formulating the hypothesis (step 3) you need to know what kind of data you have (step 4)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
9copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topicbull There are at least three keys to choosing
a topic1 Try to pick a field that you find interesting
andor that you know something about
2 Make sure that data are readily available with a reasonable sample (we suggest at least 25 observations)
3 Make sure that there is some substance to your topicndash Avoid topics that are purely descriptive or
virtually tautological in nature
ndash Instead look for topics that address an inherently interesting economic or behavioral question or choiceD
epar
tmen
t o
f E
con
om
ics
and
Dev
elo
pm
ent
Stu
die
s F
acu
lty
of
Eco
no
mic
s an
d b
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
10copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topic (cont)
bull Places to lookndash Your textbooks and notes from previous economics classes
ndash Economics journals
bull For example Table 111 contains a list of the journals cited so far in this textbook (in order of the frequency of citation)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
11copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
12copy Dr Yoke Muelgini MSc FEB Unila 2012
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
13copy Dr Yoke Muelgini MSc FEB Unila 2012
Table 111a Sources of Potential Topic Ideas
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
14copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
3copy Dr Yoke Muelgini MSc FEB Unila 2012
If you are the next Paul Samuelson and will wholly transform the field of economics pay no heed If you are the next Ken Arrow or Robert Lucas J and will invent a new branch of economics these notes are not for you
This paper provides tips of things to think of when writing a script or a thesis or a dissertation in applied economics One of the most important ingredients in the recipe for success is to get an early start and to begin looking for data as soon as possible
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
4copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
5copy Dr Yoke Muelgini MSc FEB Unila 2012
What is ldquoresearchrdquo The traditional view is that research is ldquothe search for knowledgerdquo which reflects the idea that science should be completely objective The researcher simply collects the data which then ldquospeak for themselvesrdquo Figurly speaking knowledge is fruit on a tree in the forrest and all the researcher needs to do is to discover the tree and collect the knowledge (Greenlaw 2006)
Of course this is not entirely true What is missing here is a distinction between knowledge and facts Knowledge is our common understanding of how things work and if we do not know exactly it is our best guess Facts on the other hand are just data In other words knowledge is facts with meaning it is the researcherrsquos best interpretation of the facts and research is the creation of such knowledgeD
epar
tmen
t o
f E
con
om
ics
and
Dev
elo
pm
ent
Stu
die
s F
acu
lty
of
Eco
no
mic
s an
d b
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
6copy Dr Yoke Muelgini MSc FEB Unila 2012
Knowledge is a dialog between researchers as they strive to strengthen their interpretations of the facts Through this dialog weaker interpretations or arguments are gradually abandoned while the stronger ones are refined and made even better Weaker arguments are those with less supporting evidence Since the data are also the facts having weak support means that the argument does not fit the facts and that it must be flawed in one way or another
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
7copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
8copy Dr Yoke Muelgini MSc FEB Unila 2012
It is important to begin thinking about these items right away before you begin the actual thesis work so that once you begin you know what to do Do not just start to think about the question Research is a highly nonlinear process that must be carried out in an iterative rather than a step-wise fashion For example although you need to begin to think about the topic (step 1) and question (step 2) early you also need to think about the availability of data (step 4) If there are no data then you can just as well start looking for something else In fact as soon as you have just a partial idea of what to do you should begin to look for data Similarly in formulating the hypothesis (step 3) you need to know what kind of data you have (step 4)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
9copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topicbull There are at least three keys to choosing
a topic1 Try to pick a field that you find interesting
andor that you know something about
2 Make sure that data are readily available with a reasonable sample (we suggest at least 25 observations)
3 Make sure that there is some substance to your topicndash Avoid topics that are purely descriptive or
virtually tautological in nature
ndash Instead look for topics that address an inherently interesting economic or behavioral question or choiceD
epar
tmen
t o
f E
con
om
ics
and
Dev
elo
pm
ent
Stu
die
s F
acu
lty
of
Eco
no
mic
s an
d b
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
10copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topic (cont)
bull Places to lookndash Your textbooks and notes from previous economics classes
ndash Economics journals
bull For example Table 111 contains a list of the journals cited so far in this textbook (in order of the frequency of citation)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
11copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
12copy Dr Yoke Muelgini MSc FEB Unila 2012
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
13copy Dr Yoke Muelgini MSc FEB Unila 2012
Table 111a Sources of Potential Topic Ideas
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
14copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
4copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
5copy Dr Yoke Muelgini MSc FEB Unila 2012
What is ldquoresearchrdquo The traditional view is that research is ldquothe search for knowledgerdquo which reflects the idea that science should be completely objective The researcher simply collects the data which then ldquospeak for themselvesrdquo Figurly speaking knowledge is fruit on a tree in the forrest and all the researcher needs to do is to discover the tree and collect the knowledge (Greenlaw 2006)
Of course this is not entirely true What is missing here is a distinction between knowledge and facts Knowledge is our common understanding of how things work and if we do not know exactly it is our best guess Facts on the other hand are just data In other words knowledge is facts with meaning it is the researcherrsquos best interpretation of the facts and research is the creation of such knowledgeD
epar
tmen
t o
f E
con
om
ics
and
Dev
elo
pm
ent
Stu
die
s F
acu
lty
of
Eco
no
mic
s an
d b
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
6copy Dr Yoke Muelgini MSc FEB Unila 2012
Knowledge is a dialog between researchers as they strive to strengthen their interpretations of the facts Through this dialog weaker interpretations or arguments are gradually abandoned while the stronger ones are refined and made even better Weaker arguments are those with less supporting evidence Since the data are also the facts having weak support means that the argument does not fit the facts and that it must be flawed in one way or another
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
7copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
8copy Dr Yoke Muelgini MSc FEB Unila 2012
It is important to begin thinking about these items right away before you begin the actual thesis work so that once you begin you know what to do Do not just start to think about the question Research is a highly nonlinear process that must be carried out in an iterative rather than a step-wise fashion For example although you need to begin to think about the topic (step 1) and question (step 2) early you also need to think about the availability of data (step 4) If there are no data then you can just as well start looking for something else In fact as soon as you have just a partial idea of what to do you should begin to look for data Similarly in formulating the hypothesis (step 3) you need to know what kind of data you have (step 4)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
9copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topicbull There are at least three keys to choosing
a topic1 Try to pick a field that you find interesting
andor that you know something about
2 Make sure that data are readily available with a reasonable sample (we suggest at least 25 observations)
3 Make sure that there is some substance to your topicndash Avoid topics that are purely descriptive or
virtually tautological in nature
ndash Instead look for topics that address an inherently interesting economic or behavioral question or choiceD
epar
tmen
t o
f E
con
om
ics
and
Dev
elo
pm
ent
Stu
die
s F
acu
lty
of
Eco
no
mic
s an
d b
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
10copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topic (cont)
bull Places to lookndash Your textbooks and notes from previous economics classes
ndash Economics journals
bull For example Table 111 contains a list of the journals cited so far in this textbook (in order of the frequency of citation)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
11copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
12copy Dr Yoke Muelgini MSc FEB Unila 2012
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
13copy Dr Yoke Muelgini MSc FEB Unila 2012
Table 111a Sources of Potential Topic Ideas
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
14copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
5copy Dr Yoke Muelgini MSc FEB Unila 2012
What is ldquoresearchrdquo The traditional view is that research is ldquothe search for knowledgerdquo which reflects the idea that science should be completely objective The researcher simply collects the data which then ldquospeak for themselvesrdquo Figurly speaking knowledge is fruit on a tree in the forrest and all the researcher needs to do is to discover the tree and collect the knowledge (Greenlaw 2006)
Of course this is not entirely true What is missing here is a distinction between knowledge and facts Knowledge is our common understanding of how things work and if we do not know exactly it is our best guess Facts on the other hand are just data In other words knowledge is facts with meaning it is the researcherrsquos best interpretation of the facts and research is the creation of such knowledgeD
epar
tmen
t o
f E
con
om
ics
and
Dev
elo
pm
ent
Stu
die
s F
acu
lty
of
Eco
no
mic
s an
d b
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
6copy Dr Yoke Muelgini MSc FEB Unila 2012
Knowledge is a dialog between researchers as they strive to strengthen their interpretations of the facts Through this dialog weaker interpretations or arguments are gradually abandoned while the stronger ones are refined and made even better Weaker arguments are those with less supporting evidence Since the data are also the facts having weak support means that the argument does not fit the facts and that it must be flawed in one way or another
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
7copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
8copy Dr Yoke Muelgini MSc FEB Unila 2012
It is important to begin thinking about these items right away before you begin the actual thesis work so that once you begin you know what to do Do not just start to think about the question Research is a highly nonlinear process that must be carried out in an iterative rather than a step-wise fashion For example although you need to begin to think about the topic (step 1) and question (step 2) early you also need to think about the availability of data (step 4) If there are no data then you can just as well start looking for something else In fact as soon as you have just a partial idea of what to do you should begin to look for data Similarly in formulating the hypothesis (step 3) you need to know what kind of data you have (step 4)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
9copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topicbull There are at least three keys to choosing
a topic1 Try to pick a field that you find interesting
andor that you know something about
2 Make sure that data are readily available with a reasonable sample (we suggest at least 25 observations)
3 Make sure that there is some substance to your topicndash Avoid topics that are purely descriptive or
virtually tautological in nature
ndash Instead look for topics that address an inherently interesting economic or behavioral question or choiceD
epar
tmen
t o
f E
con
om
ics
and
Dev
elo
pm
ent
Stu
die
s F
acu
lty
of
Eco
no
mic
s an
d b
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
10copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topic (cont)
bull Places to lookndash Your textbooks and notes from previous economics classes
ndash Economics journals
bull For example Table 111 contains a list of the journals cited so far in this textbook (in order of the frequency of citation)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
11copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
12copy Dr Yoke Muelgini MSc FEB Unila 2012
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
13copy Dr Yoke Muelgini MSc FEB Unila 2012
Table 111a Sources of Potential Topic Ideas
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
14copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
6copy Dr Yoke Muelgini MSc FEB Unila 2012
Knowledge is a dialog between researchers as they strive to strengthen their interpretations of the facts Through this dialog weaker interpretations or arguments are gradually abandoned while the stronger ones are refined and made even better Weaker arguments are those with less supporting evidence Since the data are also the facts having weak support means that the argument does not fit the facts and that it must be flawed in one way or another
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
7copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
8copy Dr Yoke Muelgini MSc FEB Unila 2012
It is important to begin thinking about these items right away before you begin the actual thesis work so that once you begin you know what to do Do not just start to think about the question Research is a highly nonlinear process that must be carried out in an iterative rather than a step-wise fashion For example although you need to begin to think about the topic (step 1) and question (step 2) early you also need to think about the availability of data (step 4) If there are no data then you can just as well start looking for something else In fact as soon as you have just a partial idea of what to do you should begin to look for data Similarly in formulating the hypothesis (step 3) you need to know what kind of data you have (step 4)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
9copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topicbull There are at least three keys to choosing
a topic1 Try to pick a field that you find interesting
andor that you know something about
2 Make sure that data are readily available with a reasonable sample (we suggest at least 25 observations)
3 Make sure that there is some substance to your topicndash Avoid topics that are purely descriptive or
virtually tautological in nature
ndash Instead look for topics that address an inherently interesting economic or behavioral question or choiceD
epar
tmen
t o
f E
con
om
ics
and
Dev
elo
pm
ent
Stu
die
s F
acu
lty
of
Eco
no
mic
s an
d b
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
10copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topic (cont)
bull Places to lookndash Your textbooks and notes from previous economics classes
ndash Economics journals
bull For example Table 111 contains a list of the journals cited so far in this textbook (in order of the frequency of citation)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
11copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
12copy Dr Yoke Muelgini MSc FEB Unila 2012
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
13copy Dr Yoke Muelgini MSc FEB Unila 2012
Table 111a Sources of Potential Topic Ideas
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
14copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
7copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
8copy Dr Yoke Muelgini MSc FEB Unila 2012
It is important to begin thinking about these items right away before you begin the actual thesis work so that once you begin you know what to do Do not just start to think about the question Research is a highly nonlinear process that must be carried out in an iterative rather than a step-wise fashion For example although you need to begin to think about the topic (step 1) and question (step 2) early you also need to think about the availability of data (step 4) If there are no data then you can just as well start looking for something else In fact as soon as you have just a partial idea of what to do you should begin to look for data Similarly in formulating the hypothesis (step 3) you need to know what kind of data you have (step 4)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
9copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topicbull There are at least three keys to choosing
a topic1 Try to pick a field that you find interesting
andor that you know something about
2 Make sure that data are readily available with a reasonable sample (we suggest at least 25 observations)
3 Make sure that there is some substance to your topicndash Avoid topics that are purely descriptive or
virtually tautological in nature
ndash Instead look for topics that address an inherently interesting economic or behavioral question or choiceD
epar
tmen
t o
f E
con
om
ics
and
Dev
elo
pm
ent
Stu
die
s F
acu
lty
of
Eco
no
mic
s an
d b
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
10copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topic (cont)
bull Places to lookndash Your textbooks and notes from previous economics classes
ndash Economics journals
bull For example Table 111 contains a list of the journals cited so far in this textbook (in order of the frequency of citation)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
11copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
12copy Dr Yoke Muelgini MSc FEB Unila 2012
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
13copy Dr Yoke Muelgini MSc FEB Unila 2012
Table 111a Sources of Potential Topic Ideas
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
14copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
8copy Dr Yoke Muelgini MSc FEB Unila 2012
It is important to begin thinking about these items right away before you begin the actual thesis work so that once you begin you know what to do Do not just start to think about the question Research is a highly nonlinear process that must be carried out in an iterative rather than a step-wise fashion For example although you need to begin to think about the topic (step 1) and question (step 2) early you also need to think about the availability of data (step 4) If there are no data then you can just as well start looking for something else In fact as soon as you have just a partial idea of what to do you should begin to look for data Similarly in formulating the hypothesis (step 3) you need to know what kind of data you have (step 4)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
9copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topicbull There are at least three keys to choosing
a topic1 Try to pick a field that you find interesting
andor that you know something about
2 Make sure that data are readily available with a reasonable sample (we suggest at least 25 observations)
3 Make sure that there is some substance to your topicndash Avoid topics that are purely descriptive or
virtually tautological in nature
ndash Instead look for topics that address an inherently interesting economic or behavioral question or choiceD
epar
tmen
t o
f E
con
om
ics
and
Dev
elo
pm
ent
Stu
die
s F
acu
lty
of
Eco
no
mic
s an
d b
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
10copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topic (cont)
bull Places to lookndash Your textbooks and notes from previous economics classes
ndash Economics journals
bull For example Table 111 contains a list of the journals cited so far in this textbook (in order of the frequency of citation)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
11copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
12copy Dr Yoke Muelgini MSc FEB Unila 2012
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
13copy Dr Yoke Muelgini MSc FEB Unila 2012
Table 111a Sources of Potential Topic Ideas
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
14copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
9copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topicbull There are at least three keys to choosing
a topic1 Try to pick a field that you find interesting
andor that you know something about
2 Make sure that data are readily available with a reasonable sample (we suggest at least 25 observations)
3 Make sure that there is some substance to your topicndash Avoid topics that are purely descriptive or
virtually tautological in nature
ndash Instead look for topics that address an inherently interesting economic or behavioral question or choiceD
epar
tmen
t o
f E
con
om
ics
and
Dev
elo
pm
ent
Stu
die
s F
acu
lty
of
Eco
no
mic
s an
d b
usi
nes
s U
niv
ersi
ty o
f L
amp
un
gJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
10copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topic (cont)
bull Places to lookndash Your textbooks and notes from previous economics classes
ndash Economics journals
bull For example Table 111 contains a list of the journals cited so far in this textbook (in order of the frequency of citation)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
11copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
12copy Dr Yoke Muelgini MSc FEB Unila 2012
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
13copy Dr Yoke Muelgini MSc FEB Unila 2012
Table 111a Sources of Potential Topic Ideas
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
14copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
10copy Dr Yoke Muelgini MSc FEB Unila 2012
1 Choosing your topic (cont)
bull Places to lookndash Your textbooks and notes from previous economics classes
ndash Economics journals
bull For example Table 111 contains a list of the journals cited so far in this textbook (in order of the frequency of citation)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
11copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
12copy Dr Yoke Muelgini MSc FEB Unila 2012
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
13copy Dr Yoke Muelgini MSc FEB Unila 2012
Table 111a Sources of Potential Topic Ideas
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
14copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
11copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
12copy Dr Yoke Muelgini MSc FEB Unila 2012
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
13copy Dr Yoke Muelgini MSc FEB Unila 2012
Table 111a Sources of Potential Topic Ideas
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
14copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
12copy Dr Yoke Muelgini MSc FEB Unila 2012
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
13copy Dr Yoke Muelgini MSc FEB Unila 2012
Table 111a Sources of Potential Topic Ideas
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
14copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
13copy Dr Yoke Muelgini MSc FEB Unila 2012
Table 111a Sources of Potential Topic Ideas
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
14copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
14copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
15copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The first step is choosing the dependent variable ndash this step is determined by the purpose of the research (see Chapter 11 for details)
bull After choosing the dependent variable itrsquos logical to follow the following sequence
1 Review the literature and develop the theoretical model
2 Specify the model Select the independent variables and the functional form
3 Hypothesize the expected signs of the coefficients
4 Collect the data Inspect and clean the data
5 Estimate and evaluate the equation
6 Document the resultsDep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
16copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Perhaps counter intuitively a strong theoretical foundation is the best start for any empirical project
bull Reason main econometric decisions are determined by the underlying theoretical model
bull Useful starting pointsndash Journal of Economic Literature or a business
oriented publication of abstracts
ndash Internet search including Google Scholar
ndash EconLit an electronic bibliography of economics literature (for more details go to wwwEconLitorg)
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
17copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
18copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After selecting the dependent variable the specification of a model involves choosing the following components1 the independent variables and how they
should be measured
2 the functional (mathematical) form of the variables and
3 the properties of the stochastic error term
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
19copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A mistake in any of the three elements results in a specification error
bull For example only theoretically relevant explanatory variables should be included
bull Even so researchers frequently have to make choices ndashalso denoted imposing their priors
bull Example
bull when estimating a demand equation theory informs us that prices of complements and substitutes of the good in question are important explanatory variables
bull But which complementsmdashand which substitutes
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
20copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once the variables are selected itrsquos important to hypothesize the expected signs of the regression coefficients
bull Example demand equation for a final consumption good
bull First state the demand equation as a general function
(32)
bull The signs above the variables indicate the hypothesized sign of the respective regression coefficient in a linear model
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
21copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A general rule regarding sample size is ldquothe more observations the betterrdquo
bull as long as the observations are from the same general population
bull The reason for this goes back to notion of degrees of freedom (mentioned first in Section 24)
bull When there are more degrees of freedom
bull Every positive error is likely to be balanced by a negative error (see Figure 32)
bull The estimated regression coefficients are estimated with a Precision
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
22copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
23copy Dr Yoke Muelgini MSc FEB Unila 2012
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
24copy Dr Yoke Muelgini MSc FEB Unila 2012
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
25copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Estimate model using the data in Table 22 to get
bull Inspecting the datamdashobtain a printout or plot (graph) of the data
bull Reason to look for outliersndash An outlier is an observation that lies outside the range of the rest of the observations
bull Examples
ndash Does a student have a 70 GPA on a 40 scale
ndash Is consumption negative
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
26copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Once steps 1ndash4 have been completed the estimation part is quick
ndash using Eviews or Stata to estimate an OLS regression takes less than a second
bull The evaluation part is more tricky however involving answering the following questions
ndash How well did the equation fit the datandash Were the signs and magnitudes of the estimated coefficients as expected
bull Afterwards may add sensitivity analysis (see Section 64 for details)
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
27copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A standard format usually is used to present estimated regression results
(33)
bull The number in parentheses under the estimated coefficient is the estimated standard error of the estimated coefficient and the t-value is the one used to test the hypothesis that the true value of the coefficient is different from zero on (more this later)
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
28copy Dr Yoke Muelgini MSc FEB Unila 2012
The following items are crucial when starting up a new thesis project1Choosing your topic2Do your econometric approach that fits the problem and that can be used to test the hypothesis of interest3Write Your Research Report
Dep
artm
ent
of
Eco
no
mic
s an
d D
evel
op
men
t S
tud
ies
Fac
ult
y o
f E
con
om
ics
and
bu
sin
ess
Un
iver
sity
of
Lam
pu
ngJu
rusan
Eko
no
mi P
emb
ang
un
an F
akultas E
kon
om
i dan
Bisn
is Un
iversitas Lam
pu
ng
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
29copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Most good research reports have a number of elements in common
ndash A brief introduction that defines the dependent variable and states the goals of the research
ndash A short review of relevant previous literature and research
ndash An explanation of the specification of the equation (model)
bull Independent variablesbull functional formsbull expected signs of (or other hypotheses about) the slope coefficients
ndash A description of the databull generated variablesbull data sourcesbull data irregularities (if any)
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
30copy Dr Yoke Muelgini MSc FEB Unila 2012
bull A presentation of each estimated specification using our standard documentation format
ndash If you estimate more than one specification be sure to explain which one is best (and why)
bull A careful analysis of the regression resultsndash discussion of any econometric problems encountered
ndash complete documentation of all
bull equations estimated
bull tests run
bull A short summaryconclusion that includes any policy recommendations or suggestions for further research
bull A bibliographybull An appendix that includes all data all regression runs and all relevant computer output
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
31copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background
bull You have been hired to determine the best location for the next Woodyrsquos restaurant (a moderately priced 24-hour family restaurant chain)
bull Objective
bull How to decide location using the six basic steps of applied regression analysis discussed earlier
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
32copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Background reading about the restaurant industry
bull Talking to various experts within the firm
ndash All the chainrsquos restaurants are identical and located in suburban retail or residential environments
ndash So lack of variation in potential explanatory variables to help determine location
ndash Number of customers most important for locational decision 1048774 Dependent variable number of customers (measured by the number of checks or bills)
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
33copy Dr Yoke Muelgini MSc FEB Unila 2012
bull More discussions with in-house experts reveal three major determinants of sales
ndash Number of people living near the location
ndash General income level of the location
ndash Number of direct competitors near the location
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
34copy Dr Yoke Muelgini MSc FEB Unila 2012
bull Based on this the exact definitions of the independent variables you decide to include are
ndash N = Competition the number of direct competitors within a two mile radius of the Woodyrsquos location
ndash P = Population the number of people living within a three-mile radius of the location
ndash I = Income the average household income of the population measured in variable P
bull With no reason to suspect anything other than linear functional form and a typical stochastic error term thatrsquos what you decide to use
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
35copy Dr Yoke Muelgini MSc FEB Unila 2012
bull After talking some more with the in-house experts and thinking some more you come up with the following
(34)
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
36copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You manage to obtain data on the dependent and independent variables for all 33 Woodyrsquos restaurants
bull Next you inspect the data
bull The data quality is judged as excellent because
bull Each manager measures each variable identically
bull All restaurants are included in the sample
bull All information is from the same year
bull The resulting data is as given in Tables 31 and 33 in the book ( using Eviews and Stata respectively)
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
37copy Dr Yoke Muelgini MSc FEB Unila 2012
bull You take the data set and enter it into the computer
bull You then run an OLS regression (after thinking the model over one last time)
bull The resulting model is
(35)
Estimated coefficients are as expected and the fit is reasonable
bull Values for N P and I for each potential new location are then obtained and plugged into (35) to predict Y
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
38copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The results summarized in Equation 35 meet our documentation requirements
bull Hence you decide that therersquos no need to take this step any further
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
39copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
40copy Dr Yoke Muelgini MSc FEB Unila 2012
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
41copy Dr Yoke Muelgini MSc FEB Unila 2012
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
42copy Dr Yoke Muelgini MSc FEB Unila 2012
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
43copy Dr Yoke Muelgini MSc FEB Unila 2012
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
44copy Dr Yoke Muelgini MSc FEB Unila 2012
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
45copy Dr Yoke Muelgini MSc FEB Unila 2012
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
46copy Dr Yoke Muelgini MSc FEB Unila 2012
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
47copy Dr Yoke Muelgini MSc FEB Unila 2012
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
48copy Dr Yoke Muelgini MSc FEB Unila 2012
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
49copy Dr Yoke Muelgini MSc FEB Unila 2012
bull The six steps in applied regression analysis
bull Dummy variable
bull Cross-sectional data set
bull Specification error
bull Degrees of freedom
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
50copy Dr Yoke Muelgini MSccopy Dr Yoke Muelgini MSc FE FEBB Unila 20 Unila 201212
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
Dep
artm
en
t of E
con
om
ics a
nd
Deve
lopm
en
t Stu
die
s Facu
lty of E
con
om
ics
an
d
an
d
Bu
siness U
nive
rsity of L
am
pu
ng
Bu
siness U
nive
rsity of L
am
pu
ng
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
51copy Dr Yoke Muelgini MSc FEB Unila 2012
Practice Makes
Perfect Why donrsquot
You do it
Learning to Use Regression Analysis
- Slide 1
- Slide 2
- Slide 3
- Slide 4
- Slide 5
- Slide 6
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Slide 12
- Slide 13
- Slide 14
- Slide 15
- Slide 16
- Slide 17
- Slide 18
- Slide 19
- Slide 20
- Slide 21
- Slide 22
- Slide 23
- Slide 24
- Slide 25
- Slide 26
- Slide 27
- Slide 28
- Slide 29
- Slide 30
- Slide 31
- Slide 32
- Slide 33
- Slide 34
- Slide 35
- Slide 36
- Slide 37
- Slide 38
- Slide 39
- Slide 40
- Slide 41
- Slide 42
- Slide 43
- Slide 44
- Slide 45
- Slide 46
- Slide 47
- Slide 48
- Slide 49
- Slide 50
- Slide 51
-
Related Documents
