> Until then I'm going with that studies I've seen that all seem to say roughly the same thing (despite widely varying sample sizes and quality of methodology)

There are none with acceptable sizes. You're just talking.