Severine Toussaert, Oxford University

"Predicting Social Science Results: A Review and Meta-analysis"

Abstract

Researchers rely on their beliefs about future outcomes when making important decisions, such as whether to pursue a risky project or to accept a paper for publication. In recognition of the scientific value of these beliefs, increasingly many researchers are collecting predictions of research results, although the practice is not yet systematic. Before it becomes so, it could be beneficial to obtain a better understanding of the returns to collecting predictions and the best practices to follow when doing so. We contribute to this by conducting a narrative review and a quantitative meta-analysis of the literature of participating papers in the social sciences (N=108). This paper presents preliminary findings from our systematic evaluation and conceptual discussions that address four key themes: (i) why predictions are being collected, (ii) who is collecting and providing them, (iii) how this is being done, and (iv) whether/when predictions tend to be informative. We leverage a pooled forecast dataset to tackle the final question. Our discussion concludes with evidence-based suggestions on best practices and possible next steps for the literature.

Severine Toussaert is an Associate Professor in Economics at St John's College, Oxford University. She holds two Masters degrees from Sciences Po and Paris School of Economics as well as a PhD in Economics from New York University.

Severine uses a combination of theory and experiments to study topics at the intersection of economics and psychology. One of her main research interests pertains to the study of self-control in different contexts such as weight loss, smoking cessation or work productivity.

You can read more about Severine Toussaert here

CEBI contact: Felix Chopra