Hi,
I have an experiment with 2 groups and 3 conditions. I am interested
in looking at the difference between groups for the 3 conditions after
controlling for the effect of 2 variables.
I had two options-
1. Flexible factorial - I can use a subject factor and look at the
group X condition interaction effects but i will not be able to
control for effect of covariates
2. Full factorial - I use 2X3 full factorial design -and use
covariates (padded with zeros for all other levels of all other
factors) and look at group differences.
Which is a better approach?
Also, if I use the latter approach, is it correct to assign covariates
for only one level of the group factor and pad the rest with zeros?
For example, my design matrix has 6 colums for 2X3 factors so I add
four more columns2 for each level of the group factor.
I'll be grateful for any ideas/suggestions!
Thanks!
Megha
--
Megha Sharda
Graduate Student
National Brain Research Centre
Manesar, Gurgaon -122050
India
Ph: +919810943202
|