I'm a relative novice to SAS, and running a GENMOD analysis that includes a categorical covariate (maternal race/ethnicity), and the problem that I'm having is in interpreting the output under the maximum likelihood parameter estimates. Here's my output table:
I'm putting together a synopsis of results for my boss, and I'm not sure how to interpret or report the output for that covariate, since each category within the variable has its own estimate, CI limits, p-value, etc. Am I correct in assuming that each category within that covariate is being treated as a dummy variable, and a significant association is indicative of a relationship of belonging (vs. not belonging) to that category to the outcome (depression t-score)?
Thanks in advance!
When asking how to interpret some output it is a good idea to include the code used as well. Best would be to include the LOG with the code and any messages as sometimes you will get notes about the data that may affect interpretation and possibly whether some of the output should be considered reliable. The code is needed because options chosen do affect meanings.
Better may be to include all the output. The
Sure, that makes sense. Here's the log:
188 proc genmod data=ngc_asq.ASQ_StressDep_AllData_20231004;
189 where asq_done=1;
190 class mat_race_eth;
191 model asq8_tot_comm = mat_race_eth depression_tscore;
192 run;
NOTE: Algorithm converged.
NOTE: The scale parameter was estimated by maximum likelihood.
I believe in PROC GENMOD (SAS/Stat) the default encoding for CLASS effects = GLM.
It's confirmed in this blog :
Encodings of CLASS variables in SAS regression procedures: A cheat sheet
By Rick Wicklin on The DO Loop May 1, 2019
https://blogs.sas.com/content/iml/2019/05/01/encoding-parameterization-regression-sas.html
Without seeing your PROC GENMOD I do not know which parameterization for CLASS effects was used.
How to interpret the regression coefficients then?
Look here:
For completeness, I am also adding Part 2 & 3 of the series:
Koen
To answer briefly: yes, each category is represented in the model by a dummy variable. Your output makes it clear that you used the default dummy variable coding (GLM coding), so the parameter for a given category is interpreted as the difference in effect of that category compared to the reference category, which is the last category. The last category has a zero parameter because its effect is the difference with itself.
The usage note below --
http://support.sas.com/kb/38384
explains how to interpret the output for the GLM coding of class variables for various models, such as one-way ANOVA, two-way ANOVA, ANCOVA models. I hope you find it helpful.
Thanks,
Jill
Join us for SAS Innovate 2025, our biggest and most exciting global event of the year, in Orlando, FL, from May 6-9. Sign up by March 14 for just $795.
ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.
Find more tutorials on the SAS Users YouTube channel.