>Peter, I now see the original E & H table was based on matched pairs not the raw counts. I now understand this much better and have the syntax generates results that correspond with your results (and SPSS), Thanks again, Bob