BookmarkSubscribeRSS Feed
Calcite | Level 5


I have a variety of related variables, the minority of which warrant analysis with a zero-inflated model. To facilitate presentation of test statistics and P-values, it would be nice if I could present results from the omnibus model and/or the main effects and their interaction for all variables. I'm posting this thread to see if that is possible with zero-inflated models. For this study, I measured vegetation on 3 locations over 2 months. Samples are independent across months and there are no repeated measures. I'm using the following code for this analysis:


proc genmod data=all;

   class Location Month;

   model AboveAvg = Location Month Location*Month/ dist=zinb;

   zeromodel Location;


When I include the class statement, test statistics and p-values correspond to pairwise comparisons with the variable level listed last in the dataset (top output, not complete but illustrates the point). When I omit the class statement, test statistics and p-values correspond to the variables and their interaction (bottom output), but all have 1 degree of freedom so I'm not sure if this is actually a measure of the main effects of Location and Month. Is it possible to test main effects or the omnibus model for zero-inflated models? Thank you for any help you can provide.

Calcite | Level 5

I figured it out. The model line should read as:


model AboveAvg = Location Month Location*Month/ dist=zinb type3;

Ready to join fellow brilliant minds for the SAS Hackathon?

Build your skills. Make connections. Enjoy creative freedom. Maybe change the world. Registration is now open through August 30th. Visit the SAS Hackathon homepage.

Register today!
What is ANOVA?

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.

Discussion stats
  • 1 reply
  • 1 in conversation