0% found this document useful (0 votes)
23 viewsStatistics U4
Uploaded by
Pavan ReddyCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
23 viewsStatistics U4
Uploaded by
Pavan ReddyCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
/ 38
() Small Sample Tests —t-Distribution — Properties and Applic:
‘Means, Paired t-Test.
(ii) Analysis of Variance — One Way and Two ANOVA (with and without Interaction),
(iii), Chi-Square Distribution — Test for a Specified Population Variance, Test for Goodness of
Fit, Test for Independence of Attributes.
After studying this unit, one would be able to understand,
Concept of t-Distribution and its Properties and Applications
{Test for One and Two Means.
Paired -Test.
Analysis of Variance (ANOVA) — One Way and Two Way ANOVA.
Concept of Chi-Square (2) and its Assumptions and Applications.
‘Test for a Specified Population Variance.
Test for Goodness of Fit.
oo & & & & & &
Test for Independence of Attributes.
is less than 30, Z-tests using normal
‘Small sample test is conducted when the size of the sample
sumptions on which they are
distribution is not applicable in case of small samples as th
based generally do not hold good. The sampling distribution of small samples follow student's
t-distribution. t-test includes testing for one and two means and paired t-test
ANOVA stands for analysis of variance. ANOVA test is based on variance ratios (or tes statistic
F), ANOVA refers to the separation and variance ascribable to one group of causes from the
variance ascribable to the other group. ANOVA test includes one-way ANOVA (or one Way
classification) and Two-way ANOVA (or two way classification). In one way ANOVA, only on€
factor is considered and its effect on elementary units is observed. In two way ANOVA, '0
independent factors are considered and their effect on the response variable is observed.
Chi-square (72) refers to standard normal variable with one degree of freedom (4.0) xf test
includes, test fora specified population variance, test for goodness of ft, test for independence a ;
ausample Test bia ‘aloe ;
when he size of the sale sali lesan 30, the Zest wing nol defo are not applicable
the assumptions ov we here base penerally do othod god incase ofall sarapes, The sampling
te ion of small samples follow student -istibaton, The student distribution has a pete dispersion than
iad normal distribution. AS ‘n’ gts larger, the tdistibution ie (i
st 2,
ibution
+4 nen population standard deviation (6) isnot
Yon udent’s' distribution) forte samp
(=n)
ie
‘known and the: Sample is of small size (i.e., $30), we use
Pies ling distribution of mean and workout variable as,
where,
gsampling standard deviation.
properties of Student's t-distribution
‘The properties of student's t-distribution are as follows,
|, The probability curve of is symmetric
3}, The ‘-lstribution ranges from —a toa just like in normal distribution.
3, The F-istribution is bell shaped and symmetrical around mean zero like normal distribution.
4, The shapes of the distribution changes as the sample size changes.
5, The variance of r-distribution is always greater than one,
6 The curve of r-distribution is unimodal with mean = median = mode = 0.
‘The distribution has a greater dispersion than the normal distribution. As n becomes larger, the ‘-distribution
approaches the standard normal distribution
8. There is a family of r-distribution one for each sample size whereas, there is only one standard normal
distribution.
Standard normal distribution
tedtrbution (say 0-15)
{-datrbution (say n=7)
5 3
Figure
Applications of Student's t-istribution
The following are some important applications of student's ¢-distribution,
Test of hypothesis about the single lone mean with population variances.
Test of hypothesis about the difference between two sample means with unknown population variances.
Test of hypothesis about coefficient of eorelaton.
4% SIA PUBLISHERS AND DISTRIBUTORS PVT. LTD. a2. ‘Explain the hypothesis test concerning
for small samples.
Answer =
Hypothesis Test Concerning One Mean
In case of test concerning one/single sample
¥ ah, (Sample mean)
11 =Population mean from which samples taken.
sy ta —¥)? (SD. sample)
Degree of freedom =n — 1
Decision
Ifcalculated value of | | table value for (n— 1)
GE. at 5% or 1% level of significance then reject mull
hypothesis and conclude thatthe difference is significant.
Ifcalculated value of |r| » tn this question, the changes i.
Degrees of freedom, d.f=n—1
=9-1=8
Note: Assuming level of significance as 5%]
Rejection region for d.f=8 and >= 0.05 is,
|< 1.860
Calculated value of ¢= 1.491
oo [aaa tepleag => Accepenal
Hence, we can conclude that the medicine has
no effect in increase of blood pressure.
Q10. A drug is given to 10 patients and the
increments in their blood pressure were
recorded to be 3, 6-2, -1, 4,6, 0, 0,0, 2.18
it reasonable to believe that the drug has
no effect on change of blood pressure?
Solution :
Null Hypothesis
of Let the null hypothesis be that the drug has no
fect on changing the blood pressure of patients.
A,
Ae,
As the sample size is small and (n = 30)-
Population variance is not known and sample are
dependent, paired t-test is used.
Ce an
Where,
Intthis question, difference in the blood pressure
(@)is given directly.
I z
106324
S = 2.86
er a CS ee
ich 7 SABO. O90
Vn 10
Degree of freedom, d.f=—1=10-1=9
[Note: Level of significance is assumed as 5%]
Rejection region for d.f = 9 and a = 0.05 is
| ¢| =2.26 for two tailed test, Calculated value of | ¢| = 2.
Since calculated value of ¢ = 2 is less than table
value ie., 2.26 at 9 d.fand 0.05 level of significance, H, is,
accepted and H, is rejected. Hence we can conclude that,
the drug has no effect on changing the blood pressure of
individuals. Q11, Whatis ANOVA? Whatare its assumptions
and applications?
Answer :
Meaning of ANOVA
ANOVA stands for Analysis of Variance. Analysis.
of variance is a statistical technique used to test whether,
the means of more than two quantitative populations are
equal ornot (i.c., to make inferences about whether those
samples are drawn from the populations having the same
mean). ANOVA test is based on variance ratios (or test
seh).
ANOVA test was developed by R.A Fisher in
1920's. An alternative procedure (to t-test) needed for
testing the hypothesis that all samples could likely be
drawn from the same population is ANOVA (or F-test)
Prof. R.A. Fisher defined ANOVA as “the separation
and variance ascribable to one group of causes from the
variance ascribable to the other group”.
Example
Five fertilizers are applied to four plots, each of
wheat and yield of wheat on these plots is given. We
are interested in finding out whether the effects of these
fertilizers on the yields are significantly different or, in
other words, whether the samples have come from the
‘same population. ANOVA answers this question.
Assumptions for ANOVA
The following are the assumptions of ANOVA
test,
1. Observations are independent.
2, Each sample is drawn randomly from a normal
population. The sample statistics reflect the
characteristic of the population.
3. Variance and means are identical to the population
from which samples have been drawn.
4. Environmental effects are additive in nature.
Applications of ANOVA
The applications of ANOVA are as follows,
1. ANOVA test is used in education, industry,
business, psychology fields mainly in their
experiment design.
2, ANOVA test is used to compare several
Population means simultaneously,
ANOVA is used to test the linearity of the fitted
regression line and correlation ratio.
TRI
SAMPLE TESTS, ANALYSIS OF VARIANCE AND CHI-SQUARE DISTRIBUTIO a
Q12. Explain in detail one-way class;
of ANOVA.
Answer :
‘One-Way ANOVA
In one way classification (or Single
ANOVA), only one factor is considered and "to
on elementary units is observed. i.e, data oy
according to only one criterion, BSSif
Example: Yield of crops is affected py
We of
seed only.
Steps in Conducting One-Way ANOva Test
The steps involved in conducting one
ANOVA test using shor-cut method areas foggg’®) |
‘Step-1 |
Setnull hypothesis (#7, and alternative
Pot
(4) thes
EO =i ay
Null hypothesis states that arithmetic mean o
Populations from which samples are taken are gant
equal.
Hy :u, AMF,
Alternative hypothesis states that arithmeti
means of populations from which samples are taken an
not equal,
Step-2
Caleulate the sum of all the items ofall samples,
GT= Ex, + Ex, +Ex, Ee,
Where,
T= Sum of all the items of all samples
Step-3
Calculate Correction Factor (C.F)
(GI?
N
CF=
N=Sum of number of items in each sample
(or total number of observations)
Step-4
Calculate Total Sum of Squares (TSS)
TSS= £E xX) -CF
aH
Step-5
Calculate Sum of Squares Between Samples
(SsB)
re 2
ales. GT
Beer cs N
or
2 2
so. | 24) (2s) en} |
Step-6 my nm a
Calculate sum of squares within samples (SSW) within Samples
les. 7
sat goaloulated value of F table value of Fy the difference is taken as :
ete Tout two-way classification significant and we reject null hypothesis
wer!
fwoway ANOVA
Vay xo way classification (or two factor ANOVA), two independent factors are considered and their effect
onthe response variable of interest is observed.
Example: Yield of crops is affected by typeof seed as well as typeof fertilizer.
steps in conducting Two-Way ANOVA Test
‘The steps involved in conducting two-way ANOVA test using short-cut method are as follows.
step-1
Set null hypothesis and alternative hypothesis forthe two factors.
Say A and B
Hyg By = Varn
Hy: Ya = Haro Bae
null hypothesis stats that arithmetic mean of populations from which samples are taken are equal
Fg By F aon
Hg? Hey Fyn
Alternative hypothesis tates that arithmetic means of populations from which samples are taken are not equal.
Step-2
Calculate the sum of all the items of all.
GT = EEx,= ZR =BG
Where, :
Row
C= Column
Step-3
Calculate Correction Factor (CF)
2
CF= coe
Step-4
Calculate Total Sum of Squares (TSS) SMALL SAMPLE TESTS, ANALYSIS OF VARIANCE AND CHISTES "TT 7
4s,
Calculate Sum of Squares Between Rows (SSR)
Re eR? Re
a
Dg ott OF
Where,
n= Number of samples
Stop-6
Calculate Sum of Squares Between Columns (SSC).
eee. 3
2 SSC= sb 42 EOF
Calculate Sum of Squares due to Error (SSE).
SSE = TSS—SSR— SSC
ANOVA Table for Two-way Classification
_ SSR
MSR=7 ty
pad | PSE= =e)
MSC = Mean sum of squares of columns
MSR = Mean sum of squares of rows
MSE = Mean sum of squares of errors
Step-8
Compare the calculated value of F ratio with table value of F for degrees of freedom for between and within
samples.
Ifcalculated value of F table value of F, the difference is taken as significant and we reject null hypothesis,
[Note: Ifthe problem contains large numbers/values then coding method can be followed to simplify the calculations.
In coding method, any one number is picked (from the given values) and subtracted from the other values. The
resulting table is used for ANOVA test. The procedure is same as explained above.]
PROBLEMS ON ANOVA
ONE-WAY ANOVA
Q14. To assess the significance of possible variation In performance in a certain test between
the grammar schools of a city, a common test was given to a number of students taken at
random. From the following results, make out an analysis of variance. Jan./Feb-l, 050)
‘ire
OR
farning : Xerox/Photocopying of this book is a CRIMINAL act. Anyone found guilty is LIABLE to face LEGAL proceed! Jan-14, Q6(a)
arora enfin variation nthe performance of students in certain test
Ay
es
vate atenals rate bth et cetacean ee sila eS ie
feel
2-6) ts
Step-2: Calculation of Grand Total (GT)
GT =45 +50+60+65=200
$Step-3: Calculation of Correction Factor
n,=5,n,=5,n,=S and n,=5
N=n+ntntn,
N=S45+5+5=20
CR=-y— =2,420
Step-4: Calculation of Total Sum of Squares (TSS)
TSS= EE Xi -CF
a)
TSS = (421+ 558 + 824 + 875) - 2420
TSS = 2678 — 2420 =258
Step-5: Calculation of Sum of ‘Squares Between Samples (SSB)
zr? 2 60? , 652] _ 220)?
SSB = ZL F- 4g a ae 6g) OO
= | 2025 , 2500 , 3600 , 4225] 48400
50 ~ [225 «250360 4S] 2"
‘SSB = [405 + 500 + 720 + 845] -2420
SSB = 2470 -2420= 50
i a)
“an SMALL SAMPLE TESTS, ANALYSIS OF VARIANCE AND CHI-SQUARE DISTRIBUTION
Calculation of Sum of Squares within Samples
SSW’ = TSS- SSB
SSW = 258 -S0=208
‘Step-7: Anova Table
Between 6.67
a ET Ee
Step-8:
Fao calculated = 1.28
F ratio from table when V, = 3 and V, = 16 is 3.24 at 5% level of significance
[Note: Assuming level of significance as 5%]
Since, F., < Fis,. we accept null hypothesis, i.e, there is no significant variation in the performance
students in a certain common test conducted between the different grammar schools of a city. of
Q45. A manufacturing company wishes to test the average life of the four brands of elect
bulbs. The company uses all brands in a randomly selected production plants. The recorgs
showing the lives (in “00” hours) of bulbs are as given in the table below:
Test the hypothesis that the average life for each brand of bulbs is the same. Assume alpha,
Solution : July-18, 09(6)
Step-1
Let the null hypothesis (H,) be that there is no significant difference in the average life of four brands of
A= 4, = Hy = Hy= Hy
Let the alternative hypothesis (H,) be that there is a significant difference in the average life of four brands
of bulbs.
Berm eH by
Step-2: Calculation of Grand Total (GT)
GT=86 +97 + 103 +74 =360
Warning : Xerox/Photocopying of this book is a CRIMINAL act. Anyone found guilty is Ll
ILE to face LEGAL proceeding) N=n,+n, +n tn,
Neadt5 45 ¢4miuee Sie
me cor = 7,200
pti Catculation of Total Sum of Squares (TSS)
eee TSS = E,2)x] -CF
TSS = (1870 + 1899 +2139 + 1374) ~7200
TSS = 7282 - 7200
TSS= 82
grep: Calculation of Sum of Squares Between Samples (SSB)
Dyin
SSB = i- GF
2 a 2
cn [9 eo? + Ope 4 403)? 74)? iE pot
94
2 [7396 + 2409. 10600 5476 | 1200
SSB = [1849 + 1881.8 + 2121.8 + 1369] - 7200
= 7221.6 — 7200
SSB = 21.6
step-6: Calculation of Sum of Squares within Samples
SSW = TSS — SSB
SSW = 82 — 21.6 = 60.4
Step-7: Anova Table
F-ratio calculated = 1.67
Value of F from table when V, = 3 and V, = 14 is 5.56 at 1% level of significance.
Since, F yg, We accept null hypothesis (H,), which means there is no significant difference im the
| average life of four brands of bulbs.
TWO-WAY ANOVA
Q16. Four technician analyzed three samples each for the moisture content in the sample. The
fesults are given below:
Analyze the di
< SIA PUBLISHERS AND DISTRIBUTORS PVT. LTD. ‘SAMPLE TESTS, ANALYSIS OF VARIANCE ANO CHLSQUARE DISTRIBUTION v
(Model Paper-, 09(b) | Jb /Auiary Ay
A two-way ANOVA technique will enable us to analyse the data and comment.
Here, the two factors are, Technicians (T) and Samples (S).
Step-1: Null Hypothesis
Hor ™ Hy = Hy = He = by ie, the mean moisture content is same according to all technicians,
Hg. = Hy = bly = My Le., the mean moisture content is same in all the samples.
Alternative Hypothesis.
H,,= Atleast two of 1,,H,,1,-and 1, are different.
H,.= Atleast two of jt, p,, 1, are different
bo Calculation fm of Items of all Sample:
GT=EEx,=3,R,= EC,
GT=30+33 +37+37=137
or
GT= 42+ 50+ 45 = 137
Step-3: Calculation of Correction Factor (C.F)
Here, G7= 137, N= 12,
Bape CD" 18769 156408
Step-4: Calculation of Total Sum of Squares (TSS)
TSS = 3,2,X3 -CF
TSS = (9? + 12? + 10? + 1? + 127+ 112 + 152 +12? + 9 + 10? + 12? + 14?) — 1564.08
TSS = (81 + 144+100 + 121 + 144 + 121 +225 + 144 +81 + 100 + 144 + 196) ~ 1564.08,
= 1601 ~ 1564.08 = 36.92
Step-5: Calculation of Sum of Squares Between Rows (i.e., Between Samples)
ROR? R 2
ssp= hy, Poteet tor
= [a2 , Go? , 4s)?
SSR = teat a 1564.08
SSR = (441 + 625 + 506.25) ~ 1564.08
SSR = 1572.25 - 1564.08 |
SSR = 8.17
Step-6: Calculation of Sum of Squares Between Columns (i.e., Between Technicians)
Ge Gece G
ssc- 149494. 494 cpr
Eason 33° 37?
ssc= (30 +37 43F 3F — 1564.08
SSC = (300 + 363 + 456.33 + 456.33) — 1564.08
1564.08 = 11.58
SSC = 15 le values of F-ratio at 5% level of significance,
pe = 4.76
calculated F,,, < Table F.,
© ence null hypothesis is accepted i, there is no
frean moisture content in the sample,
Calculated Fi.) < Table Fs,
ull hypoinesay is accepted ice., there is no significant difference between samples. Mean moisture
all the samp!
Significant difference between technicians opinion about
‘i
ar.
| Operator _|
[Machine | 8, | 8 | BT 8 |
4 28 29
A, 24 22
A 20 72
A, 29 26
Perform analysis of variance of 0,05 level of significance to ascertain whether variability
in production is due to variability in operator’s performance or variability in machines’
performance.
Solution +
‘Two variable factors are,
A: Machine (Rows of the given table),
B : Operators (Columns of the given table),
Step-1
Null Hypothesis
0, = Variability in production is not due to variability in machines performance,
Ha =Ha, =H4 =Ha
+o, = Variability in production is not due to variability in operators performance,
a Ha, =He, =Ha, =Ha,
ative Hypothesis
44, Variabitty in production due to variability in operators performance ie Hy, # Hiya # yy # Hy
( #,4= Variability in production due to variability in machine performance i.e., Hy, # Has ¥ Hay ¥ Ha,
We SUA PUBLISHERS AND DISTRIBUTORS PVT. LTD. ) SMALL SAMPLE TESTS, ANALYSIS OF VARIANCE A CHI-SQUARE DISTRIBUTION
Coding of Data
‘To simpli
calculations, let us deduct 28 from each value.
‘Step-2: Calculation of Sum of Items of all Samples
GT=2Ex,=2,2,= EC,
GT=8+(-11)+28+37=62
or
GT=12+0+50+0=62
‘Step-3: Calculation of Correction Factor (C.F)
p
cr =
Ba)
eI = 240.25
Step-4: Calculation of Total Sum of Squares
TSS= 5,2, x]
TSS = (6) + (0) + (5)2 + (1)? + G) + CAP + E+ OF + CHP + CBP + 15)? + 44) Op
+(1P+ (IP +2) - CF
TSS = 2404 - 240,25
2 TSS = 2163.75
Step-5: Calculation of Sum of Squares Between Machines (Rows)
SSR = ping 0, Hi cr
SSR = Gee e et) cr
Buen ge 4
SSR = [36 +0 + 625 + 0] — 240.25
SSR = 661 240.25
SSR = 420.75
Step-6: Calculation of Sum of Squares Between Operators (Columns)
Per a
ssc= ep Se oe OF
8? , CAD? , 28) , G7"
SSC = ++ + -CF
4 4 4 4
SSC = [16 + 30.25 + 196 + 342,25] - 240.25 = $84.5 — 240.25
SSC = 344.25
‘Step-7: Calculation of Sum of Squares of Residual or Error (SSE)
SSE = TSS ~ SSR-SSC
= 2163.75 —421
144.25 = 1398.75
‘Warning ! Xerox/Photocopying of this book
INAL act. Anyone found guilty is LIABLE to f
XU
8 eat : ee eee ed
50P* sical Value Of Fyo, Computed Valueoff
3.86 0.738
Operators (2.3) =
Machines (9, 3) = 3.86 0.902
jsion
Operators
(Since the calculated value of (0.738) is less than table values (3,86) null hypothesis is accepted.
| Variability in produetion is not due to variability in operators performance.
Machines
(i) since the calculated value of F(0,902) is less than table value (3.86) null hypothesis is accepted.
Variability in production is not due to variability in machines performance.
oy
“sfixad
g18, What is Chi-square (12)? Explain its applications, test statistic and assumptions.
Answer
Meaning of Chi-Square
Chi-square (2) refers to standard normal variable variate with one degree of freedom (df). It is ex;
v= (454)
a random variable following normal distribution with mean ‘y? and standard deviation o. If
and standard
Where *y”
Jie» vy are independent random variables following normal distribution with means 41, Hy
UartOUS 6, Oy..46, then,
>_y(Xhy
eZ ("5")
7 is the sum of square of V independent standard normal variates following chi-square distribution. x2 is a
family of distributions one for each value of V.
Chi-Square () Applications of Distribution
The applications of chi-square distribution are as follows,
|. Chi-square test of goodness of fit is used to test deviations between observation and theory which may be a
result of change or inadquacy.
2. To test the independence of attributes.
3. To test whether population has a specified value of variance.
4. To test equality among several population proportions.
Chi-square Test Statistic
Chi-square i u .
(tithe type of populations. 2 describes the magnitude of discrepancy between theory and observations, Whether it can
ns. ‘
beattibuted to chance or it is due to inadequacy of theory to fit the observed facts. If? is zero, observed and expected
Fedvencies are equal ‘The greater 2 value more would be the discrepancy between observed andl expected frequencies.
| & SIA PUBLISHERS AND DISTRIBUTORS PVT. Li
_— Where, O = Observed frequency
E= Expected frequency.
Calculated value of x? is compared with table value of x? for given degrees of freedom at a Pesta,
of significance,
(@) If 72 (calculated value) > 72 (table value) the difference between theory and observation is Considereq,
significant (fluctuations of sampling not responsible). toh
(&) If z (calculated value) < 72 (table value) the difference between theory and observation is not conse
significant (fluctuations in sampling responsible)
Degrees of freedom have restriction of the total sample observations.
Assumptions/Conditions of Chi-Square (’) Test
There are five conditions to fulfill for a chi-square test,
Sample observation data must be independent of each other.
If individual frequencies are less than 10 then combine it with preceding or succeeding frequency,
Constraints on cell frequencies should be linear.
‘Sample size should have atleast 50 observations.
Given distribution must not be replaced by relative frequencies. They should be provided in origina yyy
4,3.1 Test For A Specified Population Variance a
Q19. Write a note on test of hypothesis concerning specified population variance,
Answer ;
CChi square test for the specified population variance is used when gives population is normal and hy,
specified variance oj. In this test null hypothesis is set up as H, : 0°= 9 Ifx,,x,,x,..... is a random sample
size n from the given population then,
yvaene
er) aS
Where, % %0
n= Sample size
S= Standard deviation of sample
‘94 = Population variance
Chi-square test for specified population variance follows 7? distribution with (n ~ 1) d.f at certain level
significance.
Tf 2,4 < 22 yyy» null hypothesis is accepted and if 72_, <72,,,, null hypothesis is rejected.
PROBLEMS ON TEST FOR SPECIFIED POPULATION VARIANCE ©
Q20. Arandom sample of size 25 from a normal population gives the sample standard deviatio
Of 8, Test the hypothesis that the population standard deviation is 12.
Solution :
Null Hypothi
Alternative : H,
Given that,
n=25,5=8> S= 64
Computing Test Statistic
H, : 6 = 12, population standard deviation is 12.
# 12, population standard deviation is not equal to 12.
— 25% 64 _ 1600 _
paeecias aday Tt
Degree of freedom is 25 — 1 = 24 at 5% level of significance
‘Tabulated value of x? at 0.05 for 24 d.f= 36.4
Conclusion i‘
Since calculated value (11.11) is less than the tabulated value (36.4), null hypothesis is accepted i
population standard deviation may be equal to 12.
‘Warning: Xerox/Photocopying ofthis book ls & CRIMINAL act. Anyone found gulty ls LIABLE eo face LEGAL proces)
_— [pe weight of 10 parcels of cargo
42, 47, 55, 49, 45, 57, 50, 54, 5
we say that variance of di
ple of 10 parcels was drawn is,
40,
can
:
esis : H,: 0% =25, variance of d
2 ‘i oe
vo: H, : 3 > 25, variance of distribution
ee tion of Sample Mean by,
com
Ex _ 490 _
on, Oa
The tabulated value of 7? at 0.05 for 10 —
Conclusion
Since calculated value 1° is less than the tabulated value, null hypothesis is accepted i.e., variance of
distribution of weights of all parcels is equal to 25 square kg.
422 Test for Goodness of Fit
| 922. Explain briefly test of hypothesis concerning goodness of fit.
Answer :
Goodness of Fit
| % test help us to find out how well the assumed theoretical distribution fit to the observed data. When some
theoretical distribution is fitted to the given data, the statistician or managers will be interested in knowing as to
how well this distribution fits with the observed data. This can be known from Goodness of Fit method.
Ifthe calculated value of 7? is less than the table value at a certain level of significance, the fit is considered
tobe good one ice., divergence between the observed and expected frequencies is attributable to fluctuations of
Sampling. But if the reverse occurs, the fit is not considered to be a good one. In short,
9 dfis 16.9.
2
Xeat < Xiebie => Good fit
2
Xeat > Xirpie => Not a good fit.
If O, for i= 1,2,...n is a set of observed (experimental) frequencies and E, (i=
ial *
“of thcoretical frequencies then ?= yo with the condition that,
i
fat
= J SIA PUBLISHERS AND DISTRIBUTORS PVT. LTD. 3 uma sau samc Tess, amass oF YAMARCE ANO CH SOUAREDISTILTION
+ = DE, == Tee roguenc flows 7 Distribution with (n= 1) do,
=
asap pansion ore competed boned on poten tw o bincnial I when DO given in ye
‘Steps for Test of Goodness of Fit .
L Ti Hiypetaes God tex berwwen te thorctica distbution and given data
Atternative Hypothesis : No good fit,
i SS gies Page at hyptene 2> x2 with ydof ie, theoretical distribution is poe fy
S$. Computations :3°= $a,
6 Decision yposbesia if 22 i
:
ce, the theoretical distribution is a good fit tothe data .
23. Five coins are tossed 3200 times and the number of heads appearing each time jg
'As the end, the following results were obtained. Potag,
Pera; 2] s[{si
[20 [570] 1100] 900| 500 | 50
Test the goodness of fitto determine whether the coins are unbiased. Use 5% Los, :
(Model Paper, O94) | May/un.t, yyy |
1, = The coins are unbiased
Aktemative Hypothesis : //, = The coins are biased
In this problem the distribution of heads are following the binomial distribution, so we have to copy
‘expected frequencies as shown below.
Step-1: Calculate the Values of ‘p’ and ‘q’
In case of tossing a coin,
Probability of getting a head,
Probability of not getting a head, q = ‘
wn" Cr er
As there are 6 terms, ‘n’ is one less than the number of terms.
n= SN
Step-2
Compute expected frequencies using binomial probability law,
f= 3,200
pn="Cr4q""
Where, r = 0, 1,2, 3,4, 5 = 58.8
Level of significance (a) = 0.05
Degree of freedom=n~1=6-1=5
Table of ? at 5 d.fand 0.05 is 11.07
ion
Since, the calculated 13, (58.8) is higher than critical value (x? = 11.07), we reject mall hypotheris.
Therefore, it can be concluded that gina are biased and dsrbation f@ mot a good fe to data, i
Cuca eietas a given dats bow wean Sct XtiOD at 2 significance level of tag
Sotation =
Mall Hypothesis (H,) < The demand of 2 product is enifocmaly cesrioated
‘Alternative Hiypothesis (H,)
