In Logistic Regression, the Type 3 Analysis tests are performed to test the statistical significance of each input. What exactly is the statistical test?
If you are asking what type of test the Type3 tests are, they are Wald chi-square tests if done by PROC LOGISTIC. If they are done by PROC GENMOD, they are likelihood ratio chi-square tests by default unless the WALD option is specified in the MODEL statement. But if the REPEATED statement is used in PROC GENMOD, then they are score chi-square statistics by default (again, WALD option makes them Wald chi-squares).
I don't understand your question. Do you mean pick up the most significant variables ?
proc logistic data=sashelp.class;
model sex=age weight height/selection=stepwise;
run;
@MLAC wrote:
In Logistic Regression, the Type 3 Analysis tests are performed to test the statistical significance of each input. What exactly is the statistical test?
In layman's terms, Type III tests are what you would use if there is no particular ordering of the predictor variables, each is considered to be entered into the model last, to see if it adds any predictive value. (I think) PROC LOGISTIC only uses Type III anyway, the other types of tests are for PROC GLM and similar procedures, and it is called Type III in PROC LOGISTIC just to keep the naming consistent with other modeling procedures. (A case where you would have a definite ordering is a polynomial model, where linear term is entered first, and quadratic term is entered second, and so on, in which case you would use Type II test; but you don't mention anything like this, so Type III is appropriate)
The specific output table you want says "Type 3" (or maybe "Type III") in the header above the table.
If you are asking what type of test the Type3 tests are, they are Wald chi-square tests if done by PROC LOGISTIC. If they are done by PROC GENMOD, they are likelihood ratio chi-square tests by default unless the WALD option is specified in the MODEL statement. But if the REPEATED statement is used in PROC GENMOD, then they are score chi-square statistics by default (again, WALD option makes them Wald chi-squares).
Are you ready for the spotlight? We're accepting content ideas for SAS Innovate 2025 to be held May 6-9 in Orlando, FL. The call is open until September 25. Read more here about why you should contribute and what is in it for you!
ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.
Find more tutorials on the SAS Users YouTube channel.