Jonathan, I am so sorry
http://aisel.aisnet.org/amcis2015/e-Biz/GeneralPresentations/11/
is behind a paywall. It wasn't when I first found it, and that version is
miles away from me at the moment. It describes a truly fascinating
empirical simulation laboratory participation experiment, which shows that
anonymous review is more accurate than review with identity disclosure,
which is actually very easy to find literally centuries of replication, but
it also found that the costs were more similar than conventional wisdom.
I want everyone to see it because of what the specific
experiment says about ways to detect bias at the lowest possible cost. I
have a feeling that you will quickly think of ways to extend it to study
projects' editing.
Can someone who has access to that paper please share the method
and results as fair use?