Replication is one key component towards a robust cumulative knowledge base. It plays a critical function in assessing the stability of the scientific literature. Replication involves closely repeating the procedure of a study and determining if the results are similar to the original. For decades, behavioral scientists were reluctant to publish replications. Reasons were epistemic and pragmatic. First of all, original studies were viewed as conclusive in most cases, and failures to replicate were often attributed to mistakes by the replicating researcher. In addition, failures to replicate may be caused by numerous factors. This inherent ambiguity made replications less desirable to journals. On the other hand, replication successes were expected and considered to contribute little beyond what was already known. Finally, editorial policies did not encourage the publication of replications, leaving the robustness of scientific findings largely unreported. A series of events ultimately led the research community to reconsider replication and research practices at large: the discovery of several cases of large-scale scientific misconduct (i.e., fraud), the invention and/or application of new statistical tools to assess strength of evidence, high-profile publications suggesting that some common practices may be less robust than previously assumed, failure to replicate some major findings of the field, and the creation of new, online tools aimed to promote transparency in the field. To deal with what is often regarded as the crisis of confidence, initiatives have been developed to increase the transparency of research practices, including (but not limited to) pre-registration of studies, effect size predictions and sample size/power estimation, and, of course, replications. Replication projects themselves evolved in quality: From replications that were originally as small in sample as problematically small original studies to large-scale “Many Labs” collaborative projects. Ultimately, the development of higher quality replication projects and open science tools has led (and will continue to lead) to a clearer understanding of human behavior and cognition and have contributed to a clearer distinction between exploratory and confirmatory behavioral science. The current bibliography gives an overview of the history of replications, of the development of tools and guidelines, and of review papers discussing theoretical implications of replications.
ECONOMICS OF RANDOMNESS: BEHAVIORAL ASPECTS OF DECISION MAKING UNDER RISK
