Is it possible to reduce variable level before running Chi Sq test of association

[mszommer]

Hello,

I wish to test the association between two categorical variables, one of which is Gender (1=female, 2=male, 3=diverse and 4=prefer not to answer).

Is it possible to drop levels 3 and 4 before running Chi Sq test of association (PROC FREQ)? If yes, how can it be done?

Regards,
MS

[Reeza · Accepted solution]

Add a WHERE statement to your PROC FREQ.

where gender not in (3, 4);