The Consequences of Predicting Scientific Impact in Psychology Using Journal Impact Factors

Abstract

An academic journal’s impact factor (hereafter JIF) is an average measure of the citation count of individual articles published in that journal. JIF is used to assess merit, predict impact, and allocate resources, but the actual number of citations to individual articles is only modestly correlated with the JIFs of the journals in which they are published. We counted PsycInfo citations to 1,133 articles published in nine leading psychology journals (1996–2005). Both article length ( r =.31) and reference list length ( r = .41) predicted log-transformed citation counts better than JIF ( r = .27). Articles with fewer graphs and more structural equation models were more frequently cited. Citation count was better predicted by a model based on article length and citation count rather than JIF. When JIF was used to predict citation count, the impact of women authors and social science research was underestimated. These findings distinguish impact in science, as measured by JIF, from actual impact in psychology, and they show the unintended consequences of using a measure of the former to predict the latter.