The passions and the p-values: overwhelming openness and the affective labor of radically public scholarship

• In Event: Critical Data Studies: Ethics and Human Contexts of Data Science I

Abstract

This paper integrates autoethnography, qualitative interviews, and fieldwork conducted in academic and industry data science environments to explore the emotion work and affective labor (Hochschild, 1979) in today’s radically public conduct of scientific research and publishing. The open science and reproducibility movements are calling for significant changes to the practice of science. Researchers from across fields are increasingly obligated to share data, analysis, code, and other materials for replicating published studies. Many fields are having intense debates over what replication entails and the relevance of p-values -- particularly psychology and clinical trial research (Taichman et al., 2016; Benjamin et al, 2017; Lakens et al., 2017). More broadly, such movements are taking place in the context of a rise of computationally- and data-intensive research, which is shifting conditions under which knowledge is produced.

Previous work has examined competing values in these movements (Fecher and Friesike 2014) and examined the work of openness (Levin and Leonelli 2017). In this paper, we take an affective approach to these changing contexts of the conduct of science, focusing on the experiences and emotions of early career researchers. We have observed how this institutionally unsettled ground has consequences for researchers, especially in relation to career paths. Researchers must make uncertain, potentially high-risk decisions about whether to release their data/code publicly, to post pre-prints of draft papers prior to peer review, to take public pledges to uphold principles of open science, to attempt to replicate a high-profile published study, to publicly critique a published study, and more.

Authors

• R. Stuart Geiger, UC-Berkeley Institute for Data Science

• Laura Noren, Obsidian Security