Gender differencies + effect size

#1
Hi,

I have just received some results from various tests. I would like to check the gender differencies (boys vs girls) in test scores.

I am using wilcox test and cohen function for effect size. Could you please name other tools/test that might help my in the researches?

Thank you :)
 
#2
Hi, as far as I understand you have only a factor variable with two levels as predictor, but multiple outcomes. Thus, you could perform multiple T-tests, or, if parametric assumptions are violated, robust versions of the T-test such as the U-Test. However, an alternative amd probably better approach to multiple T-tests would be a Manova, since here you can test for differences in group combinations as well. Finally, you could use discriminant analysis to look which variables discriminate best between boys/girls.
 
#3
Thank you for your hints - much appreciated.
And what about more specific problem.: I have just one test? Test1 scores for about 300 Boys and 250 Girls. I use U-test -> there is very small p-value (so girls results are diff then boys). However, when I compare means, std etc there are not so much diffrent (it is not so obvious that the results depend on gender). I am trying to use additional measure to prove effect size (i am thinking about cohen). Or should I analyze deeply influential observation/outliers etc?
Thank you once again.