Hi K, looks like an intersting study, you'll have to splain it to me some time.
I'm curious how you will analyze your results. It looks like pretest scores vary widely between groups and regression to the mean may be an issue. Will you use pretest scores as a covariate in the analyses of gains? Also how well does PA predict RC in the various groups.
I think i've mentioned before that I'm working on a preschool intervention project looking at training attention and we have some similar design issues.
Yes, it has to be said there are probably going to be issues, although I must emphasise also that those results are just a small sample of the full set of participants (there are only 4 per group there; the overall group size will be be closer to 15 per group). Hopefully this will even out some of the variance in the pretest scores.
I've done all allocation to groups randomly to avoid having to use multi level modelling for my data, since it involves 5 classes across two schools. I'm hoping I don't get significant differences between schools that will necessitate MLM...
It will be interesting to see if PA predicts RC especially at this age, since PA is supposed to pretty much hit ceiling at age 8.
I'm also rubbish with statistics so it's going to take me a while to get through it all! But hopefully should have much more substantial data by the end of this term, july-aug. That's when I get to start writing up properly!
it would be hard to work with MLM because of the small N's. Random assignment is your friend.
It's very pretty.
*runs and hides from complicated stuff*
Commenting by HaloScan
