Science of Science

This paper studies the role of information on the evolution of beliefs and smoking in the United States in the 20th and early 21st centuries. We develop a dynamic and dynastic model of smoking, mortality and beliefs. The information about the harmfulness of smoking comes from three different sources: (i) medical information or public health messages, including obfuscation from the tobacco industry, (ii) learning from individual health shocks,  and (iii) social learning, understood as the diffusion of information and beliefs within and across social groups over time. We estimate the model using data on smoking behavior, health information and data on beliefs on the effect of smoking on health that cover several decades and different social groups. The estimated model shows that each of these mechanisms played an important role in the formation of beliefs about the harmfulness of smoking and that social learning was particularly important for low-educated individuals.

The paper develops a model of non-market allocation of resources such as the awarding of grants to meritorious projects, honors to outstanding students, or journal slots to quality publications. On the supply side, the available budget of grants is awarded to applicants who are evaluated most favorably according to the noisy information available to reviewers. On the demand side, stronger candidates are more likely to obtain grants and thus self-select into applying, given that applications are costly. We establish that if evaluation is perfect, grading on a curve inefficiently discourages even the very best candidates from applying. More generally, when the budget is insufficient to award grants to all applicants, the equilibrium unravels if information is symmetric enough—the paradox of relative evaluation. Leveraging a technique based on the quantile function pioneered by Lehmann, we characterize a broad set of non-market allocation rules under which an increase in evaluation noise in a field (or course) raises participation in that field, and reduces participation in all other fields. We illustrate the practical relevance of the model by exploiting a change in the rule for apportioning the total budget to applications in different fields at the European Research Council, showing that a one standard deviation increase in own evaluation noise leads to a 0.4 standard deviation increase in the number of applications and budget share. Moreover, we derive insights for the design of evaluation institutions, particularly regarding the endogenous choice of noise by fields or courses and the optimal aggregation of fields into panels.

Clinical research should conform to high standards of ethical and scientific integrity, given that human lives are at stake. However, economic incentives can generate conflicts of interest for investigators, who may be inclined to withhold unfavorable results or even tamper with data in order to achieve desired outcomes. To shed light on the integrity of clinical trial results, this paper systematically analyzes the distribution of p-values of primary outcomes for phase II and phase III drug trials reported to the ClinicalTrials.gov registry. First, we detect no bunching of results just above the classical 5% threshold for statistical significance. Second, a density discontinuity test reveals an upward jump at the 5% threshold for phase III results by small industry sponsors. Third, we document a larger fraction of significant results in phase III compared to phase II. Linking trials across phases, we find that early favorable results increase the likelihood of continuing into the next phase. Once we take into account this selective continuation, we can explain almost completely the excess of significant results in phase III for trials conducted by large industry sponsors. For small industry sponsors, instead, part of the excess remains unexplained.