Cross-National Replication of Prosocial Simulation Effect Using Cumulative Link Mixed Modelling

Previous work has found mentally simulating events of helping others can enhance prosocial intentions. However, to date, this "prosocial simulation effect" (PSE) has only been demonstrated in North America. We provide the first preregistered replication of PSE outside of North America in a New Zealand sample, following existing protocols (Experiment 1: N = 40) and with modifications to rule out an additional confound (Experiment 2: N = 40). Moreover, given evidence that metric models are problematic for assessing ordinal data, we conducted cumulative link model (CLM)-based analyses. Both experiments provide statistically robust support for the PSE outside of North America, lending greater credence to this effect. We also show that, relative to CLM-based analyses, metric models can underestimate effects in ordinal data, yielding inconsistent results across near-identical experimental designs. We consider this issue against the backdrop of the replication crisis and recommend the use of CLM-based analyses in all research reliant on ordinal scales.