Improving Transparency in Observational Social Science Research: A Pre-Analysis Plan Approach

Fiona Burlig
DOI: https://doi.org/10.2139/ssrn.2731920
2016-01-01
SSRN Electronic Journal
Abstract:Social science research has undergone a credibility revolution, but these gains are at risk due to problematic research practices. Existing research on transparency has centered around randomized controlled trials, which constitute only a small fraction of research in economics. In this paper, I discuss three scenarios in which study preregistration can be credibly applied in non-experimental settings: cases where researchers collect their own data; prospective studies; and research using restricted-access data. Finally, I outline suggested contents for observational pre-analysis plans, and highlight where these plans should deviate from those designed for experimental research.
What problem does this paper attempt to address?