It's not even just early career. The problem is that many fields in academia are small enough that researchers end up knowing most of the others. So, if your report is going to be public and the other guy, with whom you had a nice beer last year at the Annual Congress of Researchology, has produced a slightly disappointing paper (not even fraud, just mediocre science), it's going to be psychologically harder to go full-on steam on the guy and point out all the lacking points of the paper, knowing how much work it's going to take to just redo everything, just for the paper to be published in a less renowned journal.

Of course, you could say "it's just part of the job, you'll assume next time you see the guy to have a beer". But we're all humans, and, some exceptions apart, we want to be somewhat nice to people we like...

As you can see, I'm very torn by Open Review. I think it could easily yield a very mild, nicey-nicey reviewing among established peers of the field. Maybe the solution would be for reviews to be open, but manage a way for anonymity to be respected. Note that the Editors know it all already, and they should be the warrant of the quality of the process. The problem is, they struggle to get reviewers, so they can't be too picky about the quality of their reviews...