Quantcast
Channel: Statalist
Viewing all articles
Browse latest Browse all 72773

Missing output in ologit?

$
0
0
Hi everyone,

I’m running a rather long ordinal logistic regression and am running into a problem: for some reason, not all of the potential options in my interaction term are showing up in the output. I have a continuous dependent variable (outcome), two categorical independent variables--sex (male, female) and diagnosis (DiagnosisA, DiagnosisB, controls)--and two continuous independent variables (var1, var2). When I enter the command:

ologit outcome i.Sex ib2.Diagnosis2 i.Sex#ib2.Diagnosis2 i.Sex#ib2.Diagnosis2#c.var1#c.var2

(there are a couple of other terms in there that I didn’t include because they are working the way they should and the command was pretty long!)

I get:

Sex
  • Male (beta, std. err. etc).
Diagnosis2
  • DiagnosisA (beta, std. err., etc).
  • DiagnosisB (beta, std. err., etc).
Sex#Diagnosis2
  • Male#DiagnosisA (beta, std. err., etc).
  • Male#DiagnosisB (beta, std. err., etc).
Sex#Diagnosis2#var1#var2

Etc.

In the Sex#Diagnosis2 group, where is Male#controls if the reference group is Female#controls? To that end, where are Female#DiagnosisA and Female#DiagnosisB?

That happened for each of my interaction terms that had 2 categorical variables. I didn’t get any error messages about anything being collinear, etc. I’ve tried this command both with and without all the i. and ib2.’s…every combination I can think of. Nothing changes (unless I remove all of the i.’s altogether, obviously then I just get single values for each term). I also looked this stuff up online but I think it’s a more complex topic than what I’ve been able to find…that or I’m just missing something obvious and no one else is as oblivious as I am…eek!

Any ideas on what am I missing?

Thanks so much for your help!

Caitlin

Viewing all articles
Browse latest Browse all 72773

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>