Scientific Productivity and the Reward Structure of Science

Regression analyses of longitudinal data for a probability sample of chemists provide estimates of the causal links between chemists' predoctoral training, early productivity, recognition and organizational context and productivity at the end of the first postdoctoral decade. After identifying certain methodological problems involved in using Science Citation Index citation counts to assess the effects of collegial recognition on later productivity, I report tests of several hypotheses proposed to account for scientists' conformity to productivity norms. Apart from the calibre of the Ph.D. department, measures of socialization have no direct effect on decade productivity. Early productivity and collegial recognition do contribute to decade productivity, but the strength of their effects varies by the research orientation of the first employer: early productivity is more important for those employed in universities, whereas collegial recognition is particularly important for chemists in contexts that do not stress scholarly publication.