My question concerns how SPSS does Bonferroni corrections for chi-squared tests on contingency tables > 2x2. For example when doing the post-hoc pair-wise comparisons between the 4 groups (in columns), are the adjustments based on just the row or the whole matrix? For the former, the adjusted alpha level (threshold) would be 0.0083, (.05/6); for the latter it would seem to be .00167 (.05/30). Which are they using to declare that a p-value is significant, while preserving the experiment-wise alpha of 0.05? Their documentation is unclear on this issue.
David, note that Emet asked about "contingency tables > 2x2"--i.e., k x 2 tables where k > 2.
EDIT: I suppose "> 2x2" actually means a j x k table where MIN(j,k) >= 2 and MAX(j,k) > 2. Putting that another way, it is a contingency table with more than 4 cells.
Emet, the SPSS mailing list (see link below) is a much better place to post questions about SPSS. It would also help if you posted your command syntax so that readers know which procedure you're using. E.g., CROSSTABS is the most common procedure for computing Pearson's Chi-square. But IIRC, it does not have any multiple comparison procedures built in. So I suspect you are using Custom Tables instead. The point is, don't make readers guess about all this--be clear about what you are doing by posting your syntax. (Even if you are mostly a point-and-click user, you can see the command syntax in the output window. And you can generate it by exiting the GUI dialog via the PASTE button rather than Okay.)
HTH.
http://spssx-discussion.1045642.n5.nabble.com/
If you want to test independent for contingency table to 2x2 use alpha in crosstabe more than it use chi squer .
If you want to know a difference between 4 groups must use ANOVA table for know which group cause this difference use post hoc in SPSS bonfferone test or LSD or Turky or any one of them by properties
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS