Gra6036- Multivartate Statistics with Econometrics (Psychometrics) Distributions Estimators Ulf H. Olsson Professor of Statistics.

Gra6036- Multivartate Statistics withEconometrics (Psychometrics)

DistributionsEstimators

Ulf H. Olsson

Professor of Statistics

Ulf H. Olsson

Two Courses in Multivariate Statistics

• Gra 6020 Multivariate Statistics• Applied with focus on data analysis• Non-technical

• Gra 6036 Multivariate Statistics with Econometrics• Technical – focus on both application and understanding “basics”• Mathematical notation and Matrix Algebra

Ulf H. Olsson

Course outline Gra 6036• Basic Theoretical (Multivariate) Statistics mixed with econometric

(psychometric) theory• Matrix Algebra• Distribution theory (Asymptotical)

• Application with focus on regression type models• Logit Regression • Analyzing panel data• Factor Models• Simultaneous Equation Systems and SEM

• Using statistics as a good researcher should• Research oriented

Ulf H. Olsson

Evaluation

• Term paper (up to three students) 75%• 1 – 2 weeks

• Multipple choice exam (individual) 25%• 2 – 3 hours

Ulf H. Olsson

Teaching and communication

• Lecturer 2 – 3 weeks: 3 hours per week (UHO)• Theory and demonstrations

• Exercises 1 week: 2 hours (DK)• Assignments and Software applications (SPSS/EVIEWS/LISREL)

• Blackboard and Homepage• Assistance: David Kreiberg (Dep.of economics)

Ulf H. Olsson

Week   hours Read

2 Basic Multivariate Statistical Analysis. Asymptotic Theory 

3 Lecture notes

3 Logit and Probit Regression 3 Compendium: Logistic Regression

4 Logit and Probit Regression 3 Compendium: Logistic Regression

5 Exercises 2 

6 Panel Models 3 Book chapter (14): Analyzing Panel Data: Fixed – and Random-Effects Models

7 Panel Models 3 Book chapter (14): Analyzing Panel Data: Fixed – and Random-Effects Models

8 Exercises 2 

Ulf H. Olsson

9 Factor Analysis/ Exploratory Factor Analysis

3 Structural Equation Modeling. David Kaplan, 2000

10 Confirmatory Factor Analysis 3 Structural Equation Modeling. David Kaplan, 2000

11 Confirmatory Factor Analysis 3 Structural Equation Modeling. David Kaplan, 2000

12 Exercises 2 

13 Simultaneous Equations 3 Structural Equation Modeling. David Kaplan, 2000

15 Structural Equations Models 3 Structural Equation Modeling. David Kaplan, 2000

16 Structural Equations Models 3 Structural Equation Modeling. David Kaplan, 2000

17 Exercises 2  

Ulf H. Olsson

Any Questions ?

Ulf H. Olsson

Univariate Normal Distribution•

•

Ulf H. Olsson

Cumulative Normal Distribution

Ulf H. Olsson

Normal density functions

2 21/ 2 ( ) /2 1( | , )

2

xx e

2

21( ) ,

2(0,1)

u xu e u

u N

Ulf H. Olsson

The Chi-squared distributions

2 2(0,1) (1)If u N then z u

21 2

2

1

, ,.... (1) var

( )

n

n

ii

If z z z are n independent iables

then z n

2

2

( ( ))

( ( )) 2

E n n

Var n n

Ulf H. Olsson

The Chi-squared distributions2

1 2

2 2

1

, ,.... (0, ) var

( / ) ( )

n

n

i ii

If u u u are n independent N iables

then u n

21 2

2 2

1

, ,.... ( , ) var

( / ) ( , )

n

n

i ii

If u u u are n independent N iables

then u n

2

2

( ( , ))

( ( , )) 2 4

E n n

Var n n

Ulf H. Olsson

Bivariate normal distribution

Ulf H. Olsson

Standard Normal density functions

2

2

2

1)(

u

eu

)2()1(2

1

2

222

)1(2

1),(

vuvu

evu

),....,('

;)2(

1),...,.(

21

)'2

1(

2/12/21

1

n

nn

uuu

euuu

u

uu

Ulf H. Olsson

Estimator

• An estimator is a rule or strategy for using the data to estimate the parameter. It is defined before the data are drawn.

• The search for good estimators constitutes much of econometrics (psychometrics)

• Finite/Small sample properties• Large sample or asymptotic properties

• An estimator is a function of the observations, an estimator is thus a sample statistic- since the x’s are random so is the estimator

Ulf H. Olsson

Small sample properties

( )

( )

Unbiased E

Biased E

1 1 2: ( ) ( )is more efficient Var Var

Ulf H. Olsson

Large-sample properties

: lim ( ) 1

.

nnConsistency P

for all

: lim ( )nnAsymptotic unbiased E

11

2

( ): lim 1

( )n

Varis Asymptotic Efficent

Var

for all

Ulf H. Olsson

Introduction to the ML-estimator

1 2( , ,......, );k i

Let be the data matrix

x x x where x are vectors

1 21

:

( , ,......, , ) ( , ) ( | )k

k ii

The Likelihood function is as a function of the unknown

parameter vector

f x x x f x L X

Ulf H. Olsson

Introduction to the ML-estimator• The value of the parameters that maximizes this function are the maximum likelihood

estimates • Since the logarithm is a monotonic function, the values that maximizes L are the same

as those that minimizes ln L

max in ( )

ln ( )0

ML

The necessary conditions for imiz g L is

L

We denote the ML estimator

( )L L is the Likelihood function evaluated at

Ulf H. Olsson

Introduction to the ML-estimator

• In sampling from a normal (univariate) distribution with mean and variance 2 it is easy to verify that:

22

1

1( )

n

ML iMLi

n

ML ii

x andn

x xn

•MLs are consistent but not necessarily unbiased

Two asymptotically Equivalent TestsLikelihood ration test

Wald test

Ulf H. Olsson

The Likelihood Ratio Test

.Let be a vector of parameters to be estimated

U RTwo ML estimates and

2

arg 2ln

( )

The l e sample distribution of

is d

R

U

LThe likelihood ratio is

L

Ulf H. Olsson

The Wald Test

1 2, , ( ) ' ( ) ( )If x N then x x is d

0

0

1

2

: ( ) ,

( ( ) ) ' ( ( ) )

( )

H c q

then under H

W c q U c q

is d

Ulf H. Olsson

Example of the Wald test

• Consider a simpel regression model

0 0

0

1 20 0 0

2

: ,

| |( ) ;

( )

( ) ' ( ) ( )

(1)

y x

H

we know z or ts

W Var z

is

Ulf H. Olsson

Likelihood- and Wald. Example from Simultaneous Equations Systems

• N=218; # Vars.=9; # free parameters = 21;• Df = 24;• Likelihood based chi-square = 164.48• Wald Based chi-square = 157.96

Assessing Normality and Multivariate Normality (Continuous variables)Skewness

Kurtosis

Mardias test

Ulf H. Olsson

Bivariate normal distribution

Ulf H. Olsson

Positive vs. Negative SkewnessExhibit 1

These graphs illustrate the notion of skewness. Both PDFs have the same expectation and variance. The one on the left is positively skewed. The one on the right is negatively skewed.

Ulf H. Olsson

Low vs. High KurtosisExhibit 1

These graphs illustrate the notion of kurtosis. The PDF on the right has higher kurtosis than the PDF on the left. It is more peaked at the center, and it has fatter tails.

Ulf H. Olsson

J-te order Moments• Skewness• Kurtosis

( ) , 1; ( );jj

Population central moments

E X j E X

X is continuous and random

42 2

2

: 3Kurtosis

31 3/ 2

2

:( )

Skewness

Ulf H. Olsson

Skewness and Kurtosis

1 2

0 0

2

.

: 0 : 0

and can be estimated from a sample

We can test H Skewnes and H Kurtosis

by z and tests

1 22,

var .

var : (( ) ' ( ))p

We can even estimate and test for multi iate kurtosis

Multi iate kurtosis E X X

Ulf H. Olsson

To Next week

• Down load LISREL 8.8. Adr.: http://www.ssicentral.com/• Read: David Kaplan: Ch.3 (Factor Analysis)• Read: Lecture Notes