have come to the final bit of the meta-analysis, where one looks for moderators, or other factors that explain the heterogeneity in the main effect. none of the sensible ones i've tested seem to be significant, and looking down the list of papers at the ones that are far away from the mean, i've come to the conclusion that the main variable that explains the spread is that the outlying studies were crap. unfortunately, i'm guessing this will not fly when it comes to peer review.