I am not great with statistics so please be understanding. I have 17 pollinator sites that I collected data from and am comparing the effect of habitat type (2 classes) on several different response variables. Initially, I set up the model is SAS using PROC GLM as such;

proc glm data= gardens;

class Habitat;

model polldiver= habitat;

OUTPUT OUT=RESIDS R=RES;

Run;

PROC UNIVARIATE DATA=RESIDS NORMAL PLOT;

VAR RES;

RUN;

I am told that I need to add "site" as a random effect. However, I am unsure how to go about this.

My predictor variable is habitat type: urban or rural

My response variables are: pollinator richness, abundance, visits per flower, proportion of bombus, proportion of other bees, proportion of lepidopteran, proportion of other insects.

More Travis Watson's questions See All
Similar questions and discussions