How to assess for EMM if Breslow day test doesn't work?

SAS49 — Sun, 01 May 2022 22:43:16 GMT

Hi all,

I am trying to assess for EMM between my exposure and outcome variables across strata of a 3rd variable. The exposure variable, color, has 3 categories within it (red, blue, green), and the outcome variable relatedness is ordinal and has 3 levels within it (low, medium, high). I am looking to see if there is EMM across the 3rd variable location which is binary (0, 1). When I run the following code, it outputs the Cochran-Mantel-Haenszel statistics, but no Breslow day test output.

proc freq data = my_data;
	table location*color*relatedness /cmh bdt;
run;

Does the Breslow day test not work in this case because I have categorical and ordinal exposure and outcome? Or is there something wrong with my code?

If the Breslow day test does not apply in this case, how would I statistically test for effect measure modification? Thanks for any guidance anyone can provide.

Re: How to assess for EMM if Breslow day test doesn't work?

sbxkoenk — Mon, 02 May 2022 12:34:50 GMT

Hello,

BDT is a CMH option.

( To request Tarone’s adjustment for the Breslow-Day test, specify the BDT cmh-option)

You need to ask for it like this : CMH(BDT)

proc freq data = my_data;
	table location*color*relatedness / 
  relrisk plots(only)=relriskplot(stats) cmh(bdt);
run;

Koen

topic How to assess for EMM if Breslow day test doesn't work? in Statistical Procedures

How to assess for EMM if Breslow day test doesn't work?

Re: How to assess for EMM if Breslow day test doesn't work?