Bo-Christer Björk and Paul Catani have published "Peer Review in Megajournals Compared with Traditional Scholarly Journals: Does It Make a Difference?" in Learned Publishing.
Here's an excerpt:
We report on a small pilot study in which we looked at the citation distributions for articles in megajournals compared with journals with traditional peer review, which also evaluate articles for contribution and novelty. We found that elite journals with very low acceptance rates have far fewer articles with no or few citations, but that the long tail of articles with two citations or less was actually bigger in a sample of selective traditional journals in comparison with megajournals.