Replicability in the Social Sciences: Extent, Reasons and Consequences
The replicability crisis in the social sciences threatens the credibility of research results and the basis of our work. However, more studies are needed to determine the exact share of empirical results that cannot be replicated, and whether this problem differs on the basis of topical areas (e.g., educational-research), research designs (e.g., experimental-research) and data-collection methods (e.g., online-surveys).
There are many reasons why previous results might not be replicated. Firs, cases of scientific misconduct have been identified, where results have been intentionally fabricated. Second, publication bias causes confirming studies to be published, whereas the majority of null results remain invisible. Third, different, but equally plausible methods of data collection and analysis often lead to different conclusions.
Replication as the most important research validation institution may take different forms: reanalyzing (a) the same or (b) alternative data with (c) the same or (d) different methods. One particular interesting form of replication methodology represents the so called ‘crowd sourcing data analyses’, where different teams conduct parallel analyses and later compare the methodology and results.
The aim of this session is (1) to stimulate the debate about the need of more replication in the social sciences, (2) to illustrate the ways to efficiently organize this institution and (3) to gain knowledge about the most important reason for the failure to replicate results. We invite researchers to submit papers discussing all these topics. In particular we welcome examples of replications of existing results and studies comparing substantive conclusions when using different methods and data.