How to adjust degree of freedom in Chi-Square Goodness-of-Fit Test in SAS?

Question

I want to check p-value in chi-Square test, and this is my code in SAS.

data dataA;
  input Ball $ observed;
  datalines;
A 1
B 0
C 0
D 0
E 9
;
run;

proc freq data=dataA;
    tables Ball / chisq;
    weight observed;
run;

Here is a problem. This Chi-square test did not include when value is 0. Actually, DF is 4, but now it's 1.

The below calculation is what I want (Chi-square should be 31), including 0 values.

How can I include the category with 0 value so that DF becomes 4, not 1? and Chi-squared becomes 31, not 6.4?

Thanks!!

score 2 · Accepted Answer · answered Feb 01 '22 at 19:26

2

You need to use ZEROS option on the WEIGHT statement.

proc freq data=dataA;
    tables Ball / chisq;
    weight observed / zeros;
run;

answered Feb 01 '22 at 19:26

data _null_

8,534
12
14

2

In the old days we would add 1E-6 to the zeros. – data _null_ Feb 01 '22 at 19:40
Ah!! to avoid 0, you did put the very tiny number. That makes sense!! Thank you so much!! – Jin.w.Kim Jun 02 '22 at 19:22

How to adjust degree of freedom in Chi-Square Goodness-of-Fit Test in SAS?

1 Answers1