Shalom. I have some problems in the BsProb procedure. I construct a full orthogonal factorial experimental design of 26 when the response is from random sampling from normal distribution with expectancy that depends on factor A (equal 2 when A=-1 and equal 0.5 when A=1). On primary analysis using GLM I accept (as expected) factor A with big influence. When Running the BsProb procedure I get posteriors probabilities that equal 0, on factor A and on all factors. I was trying to run it over a difference size of sample without any change on the result. What can be the problem? What can I do to solve it? Thank you, Respectfully, I attach the results of the primary analysis and the R output and program: lastsave7.txt - R output and program Sas output.html – the GLM procedure for y depend on effect A, you can see on the ANOVA analysis the large influence of this effect on y. [[alternative HTML version deleted]]