Does publicity in the science press drive citations? A vindication of peer review

Abstract: We study how publicity, in the form of highlighting in the science press, affects the citations of research papers. After a brief review of prior work, we analyze papers published in Physical Review Letters (PRL) that are highlighted across eight different platforms. Using multiple linear regression we identify how each platform contributes to citations. We also analyze how frequently the highlighted papers end up in the top 1% cited papers in their field. We find that the strongest predictors of medium-term citation impact -- up to 7 years post-publication -- are Viewpoints in Physics, followed by Research Highlights in Nature, Editors' Suggestions in PRL, and Research Highlights in Nature Physics. Our key conclusions are that (a) highlighting for importance identifies a citation advantage, which (b) is stratified according to the degree of vetting during peer review (internal and external to the journal). This implies that we can view highlighting platforms as predictors of citation accrual, with varying degrees of strength that mirror each platform's vetting level.