7/25/2019 statistics-1221151261161613-8
1/52
Coverage Measures of Central Tendency
Mean Median Mode
Measures of Variability and Dispersion Range
Average deviation Variance Standard deviation
7/25/2019 statistics-1221151261161613-8
2/52
Introduction to Notations
If variable X is the variable ofinterest and that n
!easure!ents are ta"en#
then the notation X$ X % X & ' Xn (ill be used to represent nobservations )
7/25/2019 statistics-1221151261161613-8
3/52
Sig!a
Indicates * su!!ation of +
7/25/2019 statistics-1221151261161613-8
4/52
Su!!ation Notation
If variable X is the variable of interestand that n !easure!ents are ta"en#
the su! of n observations can be(ritten as
X i = X 1+X 2+ +X nn
i=1
7/25/2019 statistics-1221151261161613-8
5/52
Su!!ation Notation
X i = X 1+X 2+ +X nn
i=1
,pper li!it of su!!ation
-o(er li!it of su!!ation
.ree" letter Sigma
7/25/2019 statistics-1221151261161613-8
6/52
Rules of Su!!ation
su!!ation of thesu! of variables
is
(X i+Y i) = X i+ Y i n
i=1
n
i=1
n
i=1
the su! oftheir su!!ations
7/25/2019 statistics-1221151261161613-8
7/52
(a i+b i++z i)
= a i+ b i + + z i
n
i=1
n
i=1
n
i=1
n
i=1
The su!!ation of the su! of
variables is'
the su! of their su!!atio
7/25/2019 statistics-1221151261161613-8
8/52
Rules of Su!!ation
cX i = c X i
= c(X 1+X 2+ +X n)
n
i=1
n
i=1
If c is a constant then'
7/25/2019 statistics-1221151261161613-8
9/52
Rules of Su!!ation
c = ncn
i=1
The su!!ation of a constant isthe product of upper li!it ofsu!!ation n and constant c )
7/25/2019 statistics-1221151261161613-8
10/52
MEASURES OF CENTRALTENDENCY
Statistics in Research
7/25/2019 statistics-1221151261161613-8
11/52
Mean The su! of all values of the
observations divided by the totalnu!ber of observations
The su! of all scores divided by thetotal fre/uency
7/25/2019 statistics-1221151261161613-8
12/52
Population mean = X i N
i=1
NSample mean x = X i
n
i=1
n
7/25/2019 statistics-1221151261161613-8
13/52
f iX in
i=1
n
Mean in an ,ngrouped0re/uency
= (f 1X1+f 2X2+ +f nXn)
(here f is the fre/uency
of the occurring score
n
7/25/2019 statistics-1221151261161613-8
14/52
1roperties 2 Mean The !ost stable !easure of central
tendency Can be a3ected by e4tre!e values Its value !ay not be an actual value in
the data set If a constant c is added5substracted to
all values the ne( !ean (illincrease5decrease by the sa!e a!ountc
7/25/2019 statistics-1221151261161613-8
15/52
Median 1ositional !iddle of an array of data Divides ran"ed values into halves
(ith 678 larger than and 678s!aller than the !edian value)
7/25/2019 statistics-1221151261161613-8
16/52
If n is odd 9
M = X (n+1)!2If n is even 9
M = X n!2 + X (n!2)+1
2
7/25/2019 statistics-1221151261161613-8
17/52
1roperties 2 Median The !edian is a positional !easure Can be deter!ined only if arranged in
order Its value !ay not be an actual value in
the data set It is a3ected by the position of ite!s in
the series but not by the value of eachite!
A3ected less by e4tre!e values
7/25/2019 statistics-1221151261161613-8
18/52
Mode Value that occurs !ost fre/uently in
the data set -ocates the point (here scores occur
(ith the greatest density -ess popular co!pared to !ean and
!edian !easures
7/25/2019 statistics-1221151261161613-8
19/52
1roperties 2 Mode It !ay not e4ist or if it does it !ay
not be uni/ue Not a3ected by e4tre!e values Applicable for both /ualitative and
/uantitative data
7/25/2019 statistics-1221151261161613-8
20/52
7/25/2019 statistics-1221151261161613-8
21/52
Range Measure of distance along the
nu!ber line over (here data e4ists :4clusive and inclusive range
:4clusive range ; largest score 2s!allest score
Inclusive range ; upper li!it 2 lo(erli!it
7/25/2019 statistics-1221151261161613-8
22/52
1roperties 2 Range Rough and general !easure of
dispersion -argest and s!allest e4tre!e values
deter!ine the range Does not describe distribution of
values (ithin the upper and lo(ere4tre!es
Does not depend on nu!ber of data
7/25/2019 statistics-1221151261161613-8
23/52
7/25/2019 statistics-1221151261161613-8
24/52
M" = # Xi $ X #
n
i=1
n
M%" = # Xi $ M #n
i=1
n
7/25/2019 statistics-1221151261161613-8
25/52
1roperties > AbsoluteDeviation
Measures variability of values in thedata set
Indicates ho( co!pact the group ison a certain !easure
7/25/2019 statistics-1221151261161613-8
26/52
Variance Average of the s/uare of deviations
!easured fro! the !ean 1opulation variance
7/25/2019 statistics-1221151261161613-8
27/52
?%
= ( X i $ )2
N
i=1
N
s % = ( X i & X )2n
i=1
n $1
7/25/2019 statistics-1221151261161613-8
28/52
s % = n X i2 $ ( X i )2n
i=1
n(n $1)
n
i=1
7/25/2019 statistics-1221151261161613-8
29/52
1roperties > Variance Addition5subtraction of a constant c
to each score (ill not change thevariance of the scores
Multiplying each score by a constantc changes the variance resulting in ane( variance !ultiplied by c %
7/25/2019 statistics-1221151261161613-8
30/52
Standard Deviation S/uare root of the average of the
s/uare of deviations !easured fro!the !ean > s/uare root of thevariance
1opulation standard deviation
7/25/2019 statistics-1221151261161613-8
31/52
? = ( X i $ )2
N
i=1
N
s = ( X i & X )
2n
i=1
n $1
7/25/2019 statistics-1221151261161613-8
32/52
@hy n2$ Degrees of freedo!
Measure of ho( !uch precision anesti!ate of variation has
.eneral rule is that the degrees offreedo! decrease as !ore para!etershave to be esti!ated
Xbar esti!ates ,sing an esti!ated !ean to nd the
standard deviation causes the loss of EN:degree of freedo!
7/25/2019 statistics-1221151261161613-8
33/52
1roperties > StandardDeviation
Most used !easure of variability A3ected by every value of every
observation -ess a3ected by Fuctuations and
e4tre!e values
7/25/2019 statistics-1221151261161613-8
34/52
1roperties > StandardDeviation
Addition5subtraction of a constant cto each score (ill not change thestandard of the scores
Multiplying each score by a constantc changes the standard deviationresulting in a ne( standard deviation!ultiplied by c
7/25/2019 statistics-1221151261161613-8
35/52
Choosing a !easure Range
Data are too little or scattered to Gustify!ore precise and laborious !easures
Need to "no( only the total spread of scores Absolute Deviation
0ind and (eigh deviations fro! the!ean5!edian
:4tre!e values unduly s"e(s the standarddeviation
7/25/2019 statistics-1221151261161613-8
36/52
Choosing a !easure Standard Deviation
Need a !easure (ith the best stability :3ect of e4tre!e values have been
dee!ed acceptable Co!pare and correlate (ith other data
sets
7/25/2019 statistics-1221151261161613-8
37/52
FREQUENCYDISTRIBUTION
Statistics in Research
7/25/2019 statistics-1221151261161613-8
38/52
H HJ KJ H% 6& HK K% L% L LH JKH% HJ KL H$ 67 H6 K7 L$ L LK J$H% HH KK KJ 67 H6 6J L7 L% L6 LL
H% HH KK KJ 67 H6 K7 L$ L& L6 LJH& HL KL H7 67 H6 K7 L$ L& LK LJH& 6J K6 KJ 67 H6 HH L7 L% L LH
H& HJ KL H$ 6$ HK K% L$ L LH J%H& HJ KL H$ 6% HK K% L% L LH JH HJ KL H$ 6& HK K% L% L LH J
67 6H K& KJ H% H HH L7 L% L LH
Ra( data
7/25/2019 statistics-1221151261161613-8
39/52
67 6H K& KJ H% H HH L7 L% L LH67 6J K6 KJ H% H6 HH L7 L% L LH67 6J KK KJ H% H6 HH L7 L% L6 LL
67 K7 KK KJ H% H6 HH L$ L& L6 LJ67 K7 KL H7 H& H6 HL L$ L& LK LJ67 K7 KL H$ H& H6 HJ L$ L LK J$
6$ K% KL H$ H& HK HJ L$ L LH J%6% K% KL H$ H& HK HJ L% L LH J6& K% KL H$ H HK HJ L% L LH J
6& K% KJ H% H HK HJ L% L LH JK
Array
7/25/2019 statistics-1221151261161613-8
40/52
0re/uency Distribution Table
Class 0re/uency Nu!ber of observations (ithin a class f
Class -i!its :nd nu!bers of the class
Class Interval Interval bet(een the upper and lo(er
class li!its ie9 X upper li!it X lo(er li!it
7/25/2019 statistics-1221151261161613-8
41/52
0re/uency Distribution Table
Class Ooundaries True li!its of the class half(ay bet(een class
li!it of the current class and that of thepreceding5succeeding class -CO and ,CO
Class SiPe Di3erence bet(een ,CO and -CO
ie9 X ,CO 2 X-CO Class Mar"
Midpoint of the class interval average valueof the upper and lo(er class li!its ie) X upper li!it 2 Xlo(er li!it
7/25/2019 statistics-1221151261161613-8
42/52
Constructing an 0DT Deter!ine nu!ber of classes
Sturges 0or!ula Q ; $ &)&%% log n S/uare Root Q ; s/rt
7/25/2019 statistics-1221151261161613-8
43/52
Constructing an 0DT Deter!ine lo(er class li!it
-o(est class should not be e!pty !ustcontain the lo(est value in the data set
Deter!ine succeeding lo(er classli!its by adding class siPe C to thecurrent lo(er class li!it
Tally fre/uencies
7/25/2019 statistics-1221151261161613-8
44/52
0 / i ib i
7/25/2019 statistics-1221151261161613-8
45/52
0re/uency Distribution Table
Class Frequency LCB UCB RF
7/25/2019 statistics-1221151261161613-8
46/52
Ether Ter!s Relative fre/uency R0
Class fre/uency divided by nu!ber ofobservations ie) R0 ; f i 5 n
Relative 0re/uency 1ercentage R01 R0 ;
7/25/2019 statistics-1221151261161613-8
47/52
Mean fro! an 0D
X = f iX i'
i=1
f i'
i=1
(here X i ; class !ar" of the i th cl
7/25/2019 statistics-1221151261161613-8
48/52
Median fro! an 0D
M = * M + n!2 $ , M $1
(here * M ; lo(er class boundary of !edianclass
, M $1 ; less than cu!ulative fre/uency
f M
7/25/2019 statistics-1221151261161613-8
49/52
Mode fro! an 0D
Mo = * Mo + f Mo $ f Mo$1
(here * Mo ; lo(er class boundary of !odal classf Mo- f Mo$1- f Mo+1 ; fre/uency of !odal class class
preceding and class succeeding the
2f Mo $ f Mo$1 $ f Mo+1
M D i i f !
7/25/2019 statistics-1221151261161613-8
50/52
MD = f i #Xi $ X#n
i=1
n
Mean Deviation fro! an0D
(here X i ; class !ar" of the i th class n ;total nu!ber of observations# total
fre/uency ie) n = f i
7/25/2019 statistics-1221151261161613-8
51/52
s % = f i(X i $ X)2n
i=1
(n $1)
Variance fro! an 0D
(here X i ; class !ar" of the i th class n ;total nu!ber of observations# total
fre/uency ie) n = f i
7/25/2019 statistics-1221151261161613-8
52/52
s % = n f iX i2 $ ( f iX i )2n
i=1
n(n $1)
Variance fro! an 0D
n
i=1
(here X i ; class !ar" of the i th class n ;total nu!ber of observations# total
fre/uency ie) n = f