I am analyzing a temporal trend(yr) of certain exposure(a b & c) by using the NHANES. I draw a plot with the median of each exposure and visually see a trend.
My code is like this:
proc suveyreg data=a;
stratum stra;
cluster clus;
weight wt;
model a=yr/anova;
run;
I am not sure if this is an appropriate approach because the exposures are right-skewed. Can somebody tell me how to get the normality test of the residual in ANOVA while using proc surveyreg? Should I use log-transformed exposure in the model?
I don't know if I can use Kruskal-Wallis test here. Is there any survey-related procedure that could conduct the test? But I think maybe proc surveyregis more suitable based on the research question. Am I wrong?
Many thanks!
