You can simultaneously fit the model to the training portion of your data and evaluate the fitted model on both the training and test portions using the PARTITION statement in PROC HPLOGISTIC. The following is a simplified version of the example titled "" in the HPLOGISTIC documentation in the SAS/STAT User's Guide (https://support.sas.com/en/software/sas-stat-support.html ). The ROLEVAR option lets you specify the variable in your data set that distinguishes the training and test portions. The output will show you fit statistics for both portions. Note also that instead of using the DESCENDING option, it is safer for you to always use the EVENT= option (either in the LOGISTIC, HPLOGISTIC, or GENMOD procedure) to be sure that you are modeling the level of the response variable that you consider the event level of interest.  proc hplogistic data=Sashelp.JunkMail; model Class(event='1')=Make Address All _3d Our Over Remove Internet Order; partition rolevar=Test(train='0' test='1'); run; ... View more

See PROC CAUSALMED in the SAS/STAT User's Guide https://support.sas.com/en/software/sas-stat-support.html . In particular, read the Overview, Getting Started, and Examples sections. ... View more

The GEE method available in PROC GEE (and GENMOD, but GEE is the recommended procedure) via the REPEATED statement is effectively absorption and clustering. It allows estimation of the effects specified in the MODEL statement adjusted for the repeated (clustered) measurements within subjects and avoids the need to estimate parameters for the individual subjects.  Using the example from the Details:Absorption section of the GLM documentation, the following statements produce essentially the same results: proc glm; absorb herd cow; class treatment; model y = treatment/solution; run; proc gee; class herd cow treatment; model y=treatment/type3 wald; repeated subject=cow(herd); run; ... View more

Because not all tests done are single DF tests on single parameters - for instance, type 3 tests or multi-DF tests you can construct in the CONTRAST or ESTIMATE statements. In general, for testing hypotheses on linear combinations of parameters with 1 or more DF, the Wald test has a limiting chi-square distribution. So, it is used throughout for consistency. ... View more

No worries, but actually GLIMMIX also runs fine with GLOGIT: proc glimmix data=dv2; model dv=income/dist=mult link=glogit s; run; ... View more

The values shown in the Estimate column are estimated logit (log odds) values for the associated Education level. Logits can have any real number. Since your response variable (Guide levels) shows 3 values in the table, I assume it has 4 possible levels resulting in 3 cumulative logits being modeled. The cumulative logits in an ordinal multinomial model like yours with response levels 1, 2, and 3 are, by default, log(p1/(p2+p3)) and log((p1+p2)/p3). That is, they divide the ordered response levels in the two possible places. Similarly for your 4 level response with 3 possibly divisions. The first column in the table identifies the logit by one of the response levels appearing in the numerator.   But I am concerned that your response levels do not sound like they are in logically ascending or descending order as they must be in order for the results to have any meaning. You should examine the Response Profile table to verify and then rename your levels as needed to assure that they are logically ordered.   Since you used the ILINK option, your table should also have a Mean column which gives estimated cumulative probabilities (p1, p1+p2, p1+p2+p3) for each Education level which are usually more meaningful than the estimated logits in the Estimate column. ... View more

Again, see that documentation for info on all options, but PARAM=GLM creates the typical 0,1 coded design columns that most people use to represent categorical predictors. But if your only categorical predictors are binary and you already have them coded as 0,1, then you can just omit the CLASS statement. ... View more

The CATEGORY= option is indeed supported in PROC LOGISTIC. Perhaps you are using a much older release. If you add the E option in the ESTIMATE statement in my example below, the resulting table shows that each estimate it produces uses, in turn, one of the intercepts so that each estimate in the resulting set appropriately gives the predicted probability for one particular response level. Also, if you change LOGISTIC to GLIMMIX in the code, and run it with just the MODEL statement and add the DIST=MULT option, it runs fine without having the response in the CLASS statement. ... View more

I assume that means each subject is one observation and is one of the 4 conditions. If that is the case, then you just need to sum the columns to get the cell counts of the 2x2 table and then enter those totals in a DATA step like the FatComp example in the note. ... View more

Those are the 4 cell count values used to create the FatComp data set in this note which illustrates, in various ways, how those and other statistics can be computed. So, you could use the same approach to create a data set with your cell counts and then use the SENSPEC option in PROC FREQ, as shown, to compute four of the statistics you want. Those statistics, and all of the other ones discussed in the note (including accuracy) can be computed as shown using PROC NLMIXED. Yet another method for computing accuracy is also shown.  ... View more