Andrew Gelman has provided plenty of evidence, though. I'd certainly say that PNAS, though it's a high quality journal, does sometimes go for "marketable" social science content over less shiny, but more substantive work.