Spencer Graves wrote: > As Peter said, the standard chi-square for a 2x2 table won't work for > your data. However, if you tabulate the numbers first and then ask for > the chi-square, you get a warning, not an error: Thanks for the solution of my problem! Best regards, Christoph