 |
 |
Dear all, We have a block design with condition A and B. With contrast A > B, we are able to look at the activation A > B using one-sample t test. To look at the sex differences, we built a two-sample t test of men vs. women in group analysis. There is no difference between men and women. However, one reviewer asked us to use ANOVA model (condition by sex) to look at the sex effect. To do ANOVA, we extracted the individual contrast images of A>0 and B>0, and then built ANOVA with condition (A vs. B) as between-subject variable and sex (men vs. women) as within-subject variable. The results showed that we obtained very similar results of main effect of condition (A vs. B) in ANOVA as of the one sample t test of A vs. B. However, we observed huge sex activation (men vs. women) in ANVOA model, while nothing with two-sample t test. Does anyone know what happened? Is it possible that the ANOVA was not valid at all? Any suggestion is highly appreciated! Thanks, Sheng ANOVA design: [image: ANOVA.png]