On 9 Apr 2002 14:04:34 -0700, [EMAIL PROTECTED] (May) wrote: > I need help with the spss software. I'm running a discriminant > analysis on just under 2500 cases with some missing values. There are > 2 supposedly very distinct groups. Each was made up from 3 different > types of people. > > How do I get SPSS to run a discriminant analysis which will create a > function that will predict the group a case belongs to? So far, the > analysis I have run use all 6 types of people instead of grouping them > into the 2 groups that I want. > > Any help will be greatly appreciated. Thanks
Discriminant function is usually more fun when based on 6 groups instead of 2, because 6 groups can have interesting plots in two or more dimensions. To get two values: I suppose you probably want to say (something like), RECODE grp6(1,2,5=1)(3,4,6=2) into grp2 and then do your discriminant function on grp2(1,2). DF *will* require a group-variable with only 2 values. If there are "2500 cases with some missing values", then there are "unfortunately, fewer than 2500 useful cases" until you get rid of those MISSINGs (drop those variables, or fill in values). -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
