I would like to improve the quality of a statistic report.
Actually, if I want to report differences among groups I proceed as follow (for example in case of quantitative variable first I produce a table with a descriptive analisys and second a proc anova or a non parametric test):
1) PROC TABULATE DATA=XX;
CLASS YY / missing ORDER=formatted descending;
JJ='Mean'*(mean='Mean' StdDev='SD' min='Min' max='Max'
p25='' median='Median' p75='' N )/misstext='0.00';
2) title "ANOVA";
PROC ANOVA DATA=XX;
MODEL JJ = YY;
MEANS YY / HOVTEST=LEVENE ;
Do you have any idea to summarize the results in one table (for example the descriptive analisys together with the p-value)?
I think that you need some output from ANOVA in the form of a dataset. It is very easy to achieve with ODS output tables like this:
ODS output HOVFTest=HT;
proc anova data=sashelp.class;
The procedure creates a dataset HT containing p-value and other parameters of Levene's test. Then you can add/merge this dataset with descriptive statistics. The list of all available ODS tables for ANOVA you can find in DETAILS/ODS TABLE NAMES section of HELP for this procedure.