Having run two different factorial anova's i get this output, but i am confused as to why the effect of informal language is different between the different tests run.
can someone explain this to me? thanks!
Hi,
You have different sets of predictors in the two models, with Informal as the only common predictor. You also are using Type III sums of squares, which treats each variable as if it is entered last in the model. Put another way, the sums of squares under Type III computation are adjusted for every other predictor. In your case, the effect of Informal has been adjusted for two different sets of predictors.
To show this, you can change the sums of squares computation method to Type I and as long as you leave Informal as the first variable in the model, it should give you the same SS/MS and F no matter what the remaining predictors are.
Hope this helps,
Scot
Hi, thanks this does help. Is this current analysis okay though, or do I need to change it to the type 1 as you noted?
- Badges
- Science topic
- Similar topics
- Philosophy
- Philosophy of Mathematics
- Informality