The hypothesis tested here is,
H0: Redemption level and location are statistically independent
Against
H1 :Redemption level and location are statistically dependent.
To conduct this test, the chi-square test of independence is used as described below.
Let;
Eij be the expected cell count for the observed cell count in row i and column j.
ri be the total count for row i
cj be the total count for column j and
n be the total count given by,
i=1∑3ri=j=1∑3cj=n
For the data above,
r1=218
r2=270
r3=112 and
c1=c2=c3=200
n=i=1∑3ri=j=1∑3cj=600
First, the expected cell counts are computed using the formula below,
Eij=(ri∗cj)/n
Therefore,
E11=E12=E13=(218∗200)/600=72.67(2 decimal places)
E21=E22=E23=(270∗200)/600=90
E31=E32=E33=(112∗200)/600=37.33(2 decimal places)
The chi-square test statistic is given as,
χc2=i=1∑3j=1∑3(Oij−Eij)2/Eij
χc2=(69−72.67)2/72.67+(97−72.67)2/72.67+(52−72.67)2/72.67+(101−90)2/90+(93−90)2/90+(76−90)2/90+(30−37.33)2/37.33+(10−37.33)2/37.33+(72−37.33)2/37.33=71.476(3 decimal places)
This statistic is compared with the chi-square table value with (r−1)(c−1)=(3−1)(3−1)=2∗2=4 degrees of freedom at α=0.01 significance level where r is the total number of rows and c is the total number of columns.
It is given as,
χ0.01,42=13.2767
Since the test statistic(χc2=71.476) is greater than the table value(χc2=13.2767), we reject the null hypothesis and conclude that sufficient evidence exist to show that the redemption level and location are related.
Comments