Ok, thank you Season! ... View more

Thank you for your suggestion. Yes, the AIC and BIC values are correct (negative) . The estimates and SE from the glimmix model using beta dist are much lower than the ones using genmod and normal dist. I added to the glimmix model  output out=xxx resid=r;  but nothing happened.  Maybe I am missing something?   proc glimmix data=one; where Season=2021; PercDMp=PercDM/100; class Harvest Variety; model PercDMp=Harvest*Variety/ dist=beta ddfm=kr; output out=xxx resid=r; run;   Thank you very much Season! ... View more

Thank you so much Steve for keep helping me. Ok, I tried that code and the likelihood ratio test for variance homogeneity for Variety was significant so I didn't change the group. I am still getting the same significant p-values. I tried laplace but same thing.  I am still surprised that I am getting these significant p-values with whatever model I try (see asteriks in graph, it doesn't look to me that those differences in perc dry matter between the varieties are significant, specially when adjusting the p-values). 1,2 and 4 in 2024 are the harvest months.   ... View more

Thank you so much for your reply Season! Thanks for pointing out the link mistake. I changed to link=identity but still getting very similar p-values and still significant. When comparing the AIC and BIC values for the genmod and glimmix models I am getting huge differences:   proc genmod data=one; where Season=2022; class Harvest Variety; model PercDM=Harvest*Variety/type3 dist=normal link=identity; slice Harvest*Variety/sliceby=Harvest sliceby=Variety adjust=simulate(seed=1); run; /*AIC 3984, BIC 4287*/   proc glimmix data=one; where Season=2022; PercDMp=PercDM/100; class Harvest Variety; model PercDMp=Harvest*Variety/ dist=beta ddfm=kr; lsmeans Harvest*Variety/slicediff=Harvest adjust=simulate(seed=1); run; /*AIC -4342, BIC -4040*/ Do I have to use something like this to check the residuals? proc reg data=one; where Season=2022; model PercDM=Harvest/dw clb; run;   Thank you very much Caroline ... View more

Thank you very much for your feedback Steve and the rest of the folks helping me in this post! I analyzed the data with genmod using a normal distribution and also in glimmix using a beta dist and got almost identical adjusted p-values  for the differences between varieties within each harvest. I am just surprised that when comparing the percentage of dry matter between varieties I get so many significant differences at many harvest points even with adjusted p-values (see graph).   proc genmod data=one; where Season=2021; class Harvest Variety; model PercDM=Harvest*Variety/type3 dist=normal link= log; slice Harvest*Variety/sliceby=Harvest  adjust=simulate(seed=1); run;   proc glimmix data=one; where Season=2021; PercDMp=PercDM/100; class Harvest Variety; model PercDMp=Harvest*Variety/ dist=beta ddfm=kr; lsmeans Harvest*Variety/slicediff=Harvest adjust=simulate(seed=1); run;     ... View more

Thanks Ksharp, So what if it is a continuous percentage like 45.76 and dont want to convert my data, then I guess I will use normal dist in proc genmod or glimmix as StatDave suggested ... View more

Thank you so much for your reply StatDave. My outcome var is continuos (percentage of dry matter). I am using proc genmod instead of glimmix because I only have fixed effects. In that case I guess I would use a normal distribution as you suggested. In the case of a continuos outcome var that is an average (fruit weight), should I also use normal instead of gamma? Thank you StatDave! ... View more

Oh ok, but at least it helped me a lot with proc glm and when having a binary outcome in proc logistic. Thanks a lot Ksharp! ... View more

Thanks for your clarification Steve! ... View more

That's great, thank you very much Ksharp!  I tried this code (omitting the first part) when my outcome var is binary and I got pretty similar values for r-square. However, when I tried it for an ordinal outcome var I got pretty different r-square results (0.07172*0.07172=0.0051 vs 0.1022). Am I doing something wrong or does this only work for binary outcome variables? This is the code I'm using : proc logistic data=one desc; model hedonic=DM/link=clogit rsquare; output out=want3 p=pred; run; proc corr data=want3; var hedonic pred; run;   These are the results I get:   Response Profile Ordered Value hedonic Total Frequency 1 9 17 2 8 93 3 7 99 4 6 103 5 5 86 6 4 100 7 3 50 8 2 31 9 1 4   Probabilities modeled are cumulated over the lower Ordered Values.   R-Square 0.1022 Max-rescaled R-Square 0.1042 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 62.8777 1 <.0001 Score 60.9165 1 <.0001 Wald 60.5840 1 <.0001 The CORR Procedure     Pearson Correlation Coefficients Prob > |r| under H0: Rho=0 Number of Observations   hedonic pred hedonic   1.00000   4680 0.07172 <.0001 4664 pred Estimated Probability 0.07172 <.0001 4664 1.00000   4664 ... View more

Ok I will try that with y and pred y, but this applies only to continuous variables right? ... View more

Thanks for your clarification StatDave! ... View more

Thank you so much for your suggestion Ksharp. To use Pearson corr between two variables, neither of the variables can be the response or outcome var? ... View more

Thank you for letting me know Steve, that's good to know, I hope I can remember that when using Glimmix.  I have a question regarding regression analysis. I know that the higher the r-square value the better, but is there a minimum r-square value in order to make meaningful predictions?   Thank you Steve! ... View more