Altruism (Still) Does Not Predict Mating Success in the Judd et al. (2022) Data: A Response to Arnocky et al. (2023)

Arnocky et al. (2023) have used new methods to further explore the research of Judd et al. (2022). By transforming the data, recoding the data, and/or deleting some of the data, the authors were able to find some significant effects (at p < .05) among the many effects they are using to consider the costly signaling hypothesis supported. A more detailed exploration of the data indicates that the conclusions reached by Judd et al. (2022) are a much more accurate representation of findings. Across the four main effects in Arnocky et al. (2017), there was an average effect size beta =.21. A reanalysis of Judd et al. (2022) shows an average effect size beta =.06. There were no significant sex moderation effects across any method of data transformation (failure to replicate). Whether statistical analyses directly align with those of Arnocky et al. is unknown as the authors declined to share their data. Our data are available on the Open Science Framework and I encourage researchers to explore the data further and reach their own conclusions about whether findings have indeed been replicated. Further direct replications and alternative assessment methods are needed to help establish whether altruism has a role in mating success in humans. I conclude by providing a more general discussion regarding the importance of considering effect sizes in replication research.