PG851: GLM--ANOVA

This blog is for the use of students enrolled in either section of PG851. It is intended to be a forum for asking, discussing, clarifying, and helping.

Saturday, March 14, 2009

Post Hocs

If one is running a factoial ANOVA and has unequal cell sizes and homogeniety of variance is it appropriate to use a bonnferoni procedure since in this case Sheffe can only be done by hand and tukey can't be used? Also, Is it okay to use Type 2 sums of squares if you are most interested in the effects of one level of a predictor on an outcome but still see other levels as important?

posted by psychvmj @ 2:48 PM  

1 Comments:

  • At 12:02 AM, Blogger Justin said…

    I believe Bonferonni should be OK in that case.

    Type II should be used if all main effects are of equal importance. Type I is used if your most important cells have the biggest ns and all of your effects are of unequal importance. Type I and II are also typically used with nonxperimental designs.

     

Post a Comment

<< Home