I do not have SPSS 7.5 in order to test this...

Jay Rothman (jrothman@carbon.cudenver.edu) wrote: : I am using SPSS 7.5 to analyze a survey with 3 categories for each : question.

: When I do a chi-square goodness of fit with 3 categories and one of the : categories has no observed values (data A=0 B=20 C=28) SPSS (incorrectly) : ignores category A and does a 2-bin test with 24 expected in each category : - instead of a 3-bin test with 16 expected in each category.

- Do you mean to say that SPSS does not read your mind in order to determine that you have 3 categories, when you neither told it the range -- CHISQUARE=varx(1,3), say --

nor told it that there were *three* categories by specifying /EXPECTED=16,16,16 --

or do you mean to say that SPSS actually has an oversight in the programming for NONPAR ? (And, if so, could you be specific?)

-- Rich Ulrich, biostatistician wpilib+@pitt.edu http://www.pitt.edu/~wpilib/index.html Univ. of Pittsburgh