s*****e 发帖数: 157 | 1 I want to study the association between gender and one ordinal variable. For
CMH test, the general association is not significant, but mean scores
across gender are significantly different. Which result should I use? Thanks
! | o****o 发帖数: 8077 | 2 CMH tests for linear relationship
you could have more complex relationship beyond being linear trend | s*****e 发帖数: 157 | 3 If the general association is significant, but mean scores
across gender are not significant. Should I still use the second result?
Thanks
! |
|