Thank you. So If I want to get my parameter estimates for the back-transformed outcome variable then I can run a new regression model using the back-transformed predicted values.   In this case: proc reg data=pred; model pred = height; run;   Right?     ... View more

That was a big help, thank you so much Rick_SAS!   In the second data step, P=new_pred would be P=new_weight?   Thanks!     ... View more

Thank yo so much Steve for helping me with this and for the valuable information.  Ok, I used the log statement before the "cards" statement so it worked. If I use the log-transformed Firmness in genmod with a normal dist.  I get AIC: 2.94 and BIC: 17.24. With a normal dist. and no transformation I get 738 and 752.  With a gamma dist. I get 898 and 913.  So I guess I will choose the log-transformed variable with the normal dist. in genmod.    To the second point on making predictions: What do I have to include in my model in order to fit a first degree linear model? What would be the code for that?   Thank you very much ... View more

Thank you Ksharp. I must be doing something wrong because I am getting the same results whether I log the variable or not.   proc genmod data=one; Where Variety='BL516' and Season=2021; Firmness=log(Firmness); class Wks; model Firmness= Wks/dist=normal; output out=out predicted=Firmness; run;   Also, how can I get the parameter estimates I need in order to make predictions like for instance "a 1-week increase in x is associated with ___ decrease in Firmness"?   Thank you very much! ... View more

Thank you Ksharp! It is strange because in the graph I posted before it looked more like a gamma distribution but if I compared the AIC or BIC, they are slightly lower when using normal (AIC=738 and BIC=752) than gamma (AIC=898 and BIC=913) distribution .  ... View more

Thank you Steve,   These 3 different options seem to work, but I couldn't figure out how to add the suboption xbeta to reschi option.   ods graphics on; proc genmod plots=all; model AvgFirm=Wks; run; ods graphics off;   ods graphics on; proc genmod plots=(predicted reschi); model AvgFirm=Wks; run; ods graphics off;   ods graphics on; proc genmod plots(unpack)=dfbeta; model AvgFirm=Wks; run; ods graphics off; ... View more

Thank you so much for your reply Steve!   I think this is what you mean?     Do you think that there is a normal distribution of residuals looking at the quantile residual plot?   Thank you Steve Caroline ... View more

Thank you for your reply Ksharp.   So instead of looking at the Shapiro-Wilk test, I should instead look at the distribution of residuals to check for normality?   Thanks for the link to test for trends, it is very useful.    proc iml; WksL={0 1 3 6}; contrL=orpol(WksL, 2); print contrL; quit;   proc glm; class Wks; model AvgFirm=Wks; contrast 'linear' Wks -0.545545 -0.327327 0.1091089 0.7637626; contrast 'quadratic' Wks 0.5128226 -0.170941 -0.740744 0.398862; run;   I used these codes and this is what I got:     Contrast DF Contrast SS Mean Square F Value Pr > F linear 1 27880.08352 27880.08352 1630.39 <.0001 quadratic 1 6793.04965 6793.04965 397.25 <.0001   Both, linear and cuadratic are significant. According to the graph, it looks to me like choosing a gamma distribution in genmod should be appropiate?   I would greatly appreciate your feedback Ksharp.   Thank you so much Caroline Thanks ... View more

I wanted to run a linear regression with a continuous variable (firmness) and a categorical var (weeks=0,1,3,6), but since my outcome variable is not normal (Shapiro-Wilk test significant) I tried it with genmod using a gamma distribution.  If I use lsmeans statement I can see if there is a significant difference between levels of the predictor variable (weeks) but Im still not able to tell the strenght and direction of the association between predictor and outcome to make a statement like "a 1-unit increase in the predictor is associated with an x change in the outcome var".  So my question is if there is a way to get this information when runing a reg analysis with genmod or if there is a more appropiate procedure when the outcome var (continuous) is not normal.   proc genmod data=one; Where Variety='BL516' and Season=2021; class Weeks; model Firmness= Weeks/dist=gamma link=log; output out=out predicted=Firmness; run;   Thank you very much! ... View more

Nevermind, I think I got it. This code gave me the graph I wanted.   ods graphics on; proc logistic data=one; class Wks Variety Season Harvest PercDFD / param=ref; model SeedGerm= Wks|Variety Season Harvest PercDFD; effectplot interaction(plotby=Variety plotby(rows)=PercDFD) / at(Harvest=all); store logimodel; run; ods graphics off; ... View more

Thank you so much for your quick reply and all the resources provided Ksharp. So only continuous variables are not included in the class statement? I tried this code to make a graph but I'm having trouble with the variable used in the "at" statement since it is a categorical variable (level 1, 6, and 8  ). All my predictor variables are categorical. This is the error Im getting: ERROR: For CLASS variables in the AT option, specify Harvest(coded)=value to indicate a coded design value, and specify Harvest='level' to indicate a CLASS level.   ods graphics on; proc logistic data=one; class PercDFD Season Harvest Variety Wks / param=ref; model SeedGerm= PercDFD Season Harvest Variety Wks; effectplot slicefit(sliceby=Season plotby(rows)=Variety) / at(Harvest=1 6 8  ) obs(fringe jitter(seed=39393)); store logimodel; run; ods graphics off;   I hope you can help me with this.   Thank you so much!   ... View more

That makes a lot of sense, I will do so. Thank you so much StatDave!     ... View more

Thank you so much for your great suggestions StatDave. Some varieties are only harvested in month 1, 3, 4, 6 and 8, unlike Hass that is harvested almost every month, so fiiting the model for each variety and season, and just testing for main effects without interactions solved the problem. So I learned the lesson on symplifying the model as much as I can. Also, the firth option worked wonderfully for the binary response variables. Is there a similar option for multinomial data?   Thank you! Caroline ... View more