bonferroni correction to a mann whitney u test?

[Jennifer L A Shaw]

Hi,
I was wondering if anybody can help me.
My data failed the levenes test (ANOVA) so I applied the Kruskal Wallis non-parametric test. There were significant differences, so now I want to know where those differences lie by doing the Mann Whitney U test. I have 24 groups to compare with 3 reps in each group. Therefore, I would ideally like to apply a bonferroni correction to my Mann whitney test to avoid a type 1 error.

I have written my syntax like this:

NPAR TESTS
/MANN-WHITNEY = TotalCu BY byPlot (1, 5)
/POSTHOC={BONFERRONI, ALPHA ([0.0020833])}

but I get an error message like this:

.17-23: error: Syntax error at 'POSTHOC': expecting end of command.
/POSTHOC

Does anybody know what it is that I am doing wrong?

Can you apply a Bonferroni correction to Mann whitney in PSPP? If yes, how? If not, is there an alternative option?

I have tried doing mann whitney for each and every pair of groups without the bonferroni correction and everything comes out significant. When I know that it shouldn't

Jen