11-17-2016 10:17 PM - last edited on 11-18-2016 03:03 AM by Reeza
If I am running a Weld Chi Square test along with a proc surveyfreq, that is:
proc surveyfreq data=malaria_data;
tables treatment*religion / wchisq;
Is the Pr>F the same as the p-value in that if Pr>F is greater than 0.05 (alpha value), we can accept the null hypothesis that the two variables are indepedent, leading us to believe that randomization into a treatment group has worked in regards to the characteristic of religion?
11-18-2016 06:01 AM
Yes. In SAS procedures the output column for the p-values will be labeled as "Pr > statistic" where statistic might be "t" for a t-statistic, "F" for an F statistic, "ChiSq" for a chi-square statistic, and so on.
Small values in that column indicate a rejection of the null hypothesis at the specified significance level.