Skip to main content
. 2022 Feb 18;11:e65419. doi: 10.7554/eLife.65419

Figure 1. A review of the literature on the non-visual effects of light reveals a sex bias.

Figure 1.

We analyzed a sample of the existing literature on the non-visual effects of light as a starting point for understanding the sex bias in the field. The sample included a total of 180 articles, and the breakdown of participant sex was then obtained in 166 articles. Binomial tests were conducted to evaluate the possibility that deviations from an even 50:50 sex distribution were attributable to chance alone. We implemented the Benjamini-Hochberg correction for multiple comparisons to control false-discovery rate (FDR). The proportion of female volunteers in each paper (represented by a dot) was plotted against the year of publication. Samples for which the proportion of female patients deviated significantly from 0.5 (P ≤ 0.05) were determined to be biased and colour-coded as orange. The marginal histograms show the numbers of papers irrespective of publication year (histogram on the right y axis), or irrespective of proportion (histogram on top x axis). Methods for paper selection are included in Methods.

Figure 1—source code 1. R code to produce Figure 1.
Figure 1—source data 1. Excel spreadsheet containing the data underlying Figure 1.