[p-e-w]

The peer review process isn't actually blind in practice. The number of experts working in a given problem space is small enough that they know most of their peers, and usually can identify at a glance who the author of a paper is, from the topic, writing style, and experimental design, which they can recognize from previous publications and correspondence.

[UncleMeat]

This is true to an extent, but also overblown. There have been a few experiments testing this and people wildly overestimate their ability to unblind paper authors correctly.