Durham RIOTs Club Seminar - ‘Context-dependency as an explanation for effect size heterogeneity and non-replicability in large-scale replication projects' Prof. Mario Gollwitzer (Ludwig-Maximilians-Universität München)

The next Durham RIOTs Club Seminar is by Dr Mario Gollwitzer, (Ludwig-Maximilians-Universität München) entitled ‘Context-dependency as an explanation for effect size heterogeneity and non-replicability in large-scale replication projects'.

Context dependency – the extent to which empirical findings in multi-site direct replications vary across contexts – can explain low replication rates and has fundamental implications for the design of replication studies. Here, I present a strategy to empirically estimate the extent to which observed context dependency is due to conceptually relevant vs. irrelevant conditions, applicable even when the exact nature of the conceptually relevant conditions is unknown. According to this strategy, conceptually relevant context dependency – the extent to which context characteristics moderate a hypothesized effect substantively and, thus, qualify a theory – can be estimated by eliminating irrelevant variation from the overall estimate of effect size heterogeneity.

In this talk, I will exemplify this approach and discuss its usefulness for (a) interpreting the results of existing replication studies and (b) designing multi-site replication projects.he project is embedded in a larger research consortium called “META-REP”. I will also give an overview of some of the individual projects funded within the META-REP consortium.

Held in conjunction with DRMC, Durham Library and the Durham RIOTs Club.