Individual researcher's peer review balance.

<p>Perfect balance would be if researchers reside on the 1∶1 line, since then they provide the same number of reviews as their submissions require. Red data points indicate individuals whose over- or under-contribution is unlikely to have resulted from chance (p>0.05, binomial test). Blue data points indicate p>0.05, though note that tests for individuals close to the origin have low statistical power. Panel (a) corresponds with the assumption that reviews are distributed equally among co-authors, (b) with the assumption that the submitting author is responsible for all reviews. Inset: dynamics of proportion of researchers over-contributing by more than double (upper line) or under-contributing by less than half (lower line). Axes of the main plot are square root scaled to better illustrate variation close to the origin. Y-values in both panels, and x-values in panel b are slightly jittered to assist visualization of otherwise overlaying points.