Replication and Political Methodology

Abstract

'Replication crisis' describes the fact that a wide range of published results fails to be replicable. Is it true that we cannot trust our conclusions if we cannot replicate our findings? Is this true for any research, qualitative and quantitative? If so, what are more rigid policies that result in replicability and reproducibility of empirical findings? In social science, we face complex research designs. Sometimes, we need to re-code data, revise our estimation strategy, and update our findings. Occasionally, we lose observations amid the research process (always remembering that we study humans and their social interaction). How can we ensure replicability in such complex settings? What are tools and strategies that help us to conduct more trustworthy research? This House Lecture addresses these, and other questions related to replicability and reproducibility. On the one hand, it aims to offer strategies and suggestions on how to produce replicable findings in quantitative and qualitative settings. On the other hand, it brings together different perspectives on the topic from our panellists as a result of their experiences as editors, authors, supervisors, or open science activists.