Cover Effects on Citations Uncovered: Evidence from Nature

Despite the prominent role played by bibliometric indicators for evaluating research, progress in pinning down the determinants of citation ows has so far been hindered by endogeneity issues. Based on 30 years of bibliometric data, we exploit a Regression Discontinuity Design to causally identify the effects that an article featured on the cover of the journal Nature has on citations to all articles by its authors. We confirm that, over time, cover articles are cited significantly more than non-cover articles, with this difference being long-lasting. However, when considering all articles by Nature authors, we find evidence of a crowding-out effect: the publication of a cover article causes citations to previous articles by its authors to decline sharply relative to citations to articles by non-cover authors.