Women in Science. Part 1. The Productivity Puzzle - J. Scott Long on Why Women Biochemists Publish Less than Men

Satisfactory and robust explanations of sex differences in scientific productivity remain elusive.This article provides a multidimensional, longitudinal description of the productivity of male and female biochemists. Several jindings have implications for explaining differences in productivity.Sex differences in the numbers ofpublicarions and citarions increase during the first decade of the career but are reversed later in the career. The lower productivity of females results from theiroverrepresentation among nonpublishers and their underrepresentation among the extremelyproductive. Among biochemists who publish, differences cannot be explained by patterns of collaboration, which are nearly identical for males and females, with one exception: females aremuch more likely to collaborate with a spouse. The smaller number of citations received byfemalesresults from their fewer publications, not from the quality of their publications. Papers by femaleson average receive more citations than those by males. These and other findings suggest futuredirections for research to understand sex differences in scientific productivity, Explaining the immense variation in theproductivity of scientists is a major researchobjective in the sociology of science. Thesex of a scientist is an important source ofvariation in scientific productivity. Thelesser productivity of females has been es-tablished in dozens of studies covering di-verse fields, spanning decades, and using amyriad of measures (see Cole& Zuckerman1984; Fox 1983; Hornig 1987; Long 1987;and Zuckerman 1987 for reviews of thisliterature). Many explanations for sex dif-ferences in productivity have been pro-posed. Differences may exist in personalcharacteristics such as ability, motivation,and dedication, or in educational back-ground. Obligations of family and childrenmay differentially affect the careers ofmales and females. Discrimination maymake resources more difficult for femalesto obtain, which in turn can limit their abil-ity to publish. A fifth explanation magni-fies the effects of other processes. The im-portance of even small differences in levelsof causal variables and subtle forms of dis-crimination may be enhanced by processesof cumulative advantage and disadvantage(Merton [1968] 1973). Unfortunately, noneof these explanations has been very suc-cessful in accounting for sex differences inproductivity. Indeed, Cole and Zuckerman(1984) aptly label these sex differences “theproductivity puzzle.”The premise of this article is that thefailure to explain sex differences in pro-ductivity is partially the result of incom-plete descriptions of what it is that needs tobe explained. Too often studies of sex dif-ferences in productivity are based on thedifferences in means of a few indicators ofproductivity measured at a single time.] Theresulting evidence is limited in three im-portant ways. First, comparisons of meansmask differences in the distributions of pro-ductivity. For example, knowing that theaverage male publishes twice as much asthe average female provides no insights onwhether these differences are due to a dis-proportionate number of females havinglow productivity, a disproportionate num-ber of males having extremely high pro-ductivity, or some combination of these. Agiven difference in means can be gener-ated by a variety of distributions generatedby substantively different processes. Sec-ond, there is no single measure of produc* 1 woufd like to thank Rachel A. Rosen feld, .Lowell L, Hargens, Thomas F. Gieryn, and an anonymousreviewer for their comments on this article. Funding was provided by grant SES-8304029 from theNational Science Founoktion. Direct correspondence to the author at Indiana University, Departmentof Sociology, 744 Ballantine Hall, Bloomington, IN 47405.