Ways Out of the Replication Crisis

The talk addresses different aspects of the replication crisis in the social sciences and suggests directions to overcome the crisis. Even if direct or exact replications of research findings are hard to achieve in practice, we should aim at designing replication studies that allow us to systematically learn from both replication success and failure. The talk addresses issues related to the definition of replication, the metrics and analyses used to assess replication success, the problem of insufficient power, and researchers’ insufficient knowledge about the underlying ‘true’ data-generating mechanisms. Supported by NSF grant #2015‐0285‐00 and IES grant #R305D190043