interesting - one way to test the efficacy of something like this might be to compare median outcomes of adopted kids (whose parents went through some parenting education) with that of the non adopted general population (after controlling for other factors).
Now which PhD candidate will bite :)