Chi Square Test for Independence  of Attribute

The chi square  test  can be  used to find out  whether two  or more attributes are associated or not. This  test helps  in finding the association or  independence of two  or more  attributes. In  order  to test  the independence  of two  attributes  you should take  the null  hypothesis that the two attributes  are  independent. The second  step under  null  hypothesis is to  compute the expected frequencies for  each  cell of  contingency table. In  case of  contingency table  the expected frequency  for ( A₁, B₁) will be  calculated as:

E ( A₁, B₁)  = (A₁) (B₁) / N

After computing expected frequencies you can  calculate the value  of chi square . the   calculated  value  of chi square  will be then compared with  table value at certain level of significance ( usually at 5% for  ( r- 1) ( c-1) degree of freedom. If  the calculated value  is lower than table  value then you will  accept the  null hypothesis  that two  attributes are independent. In case  calculated  value is greater than  the table value  then the  null hypothesis will be  rejected and you will  conclude that two  attributes are associated.