Loading…

The Evidence for Good Genes Ovulatory Shifts in Arslan et al. (2018) Is Mixed and Uncertain

In Arslan et al. (2018), we reported ovulatory increases in extra-pair sexual desire, in-pair sexual desire, and self-perceived desirability, as well as several moderator analyses related to the good genes ovulatory shift hypothesis, which predicts attenuated ovulatory increases in extra-pair desire...

Full description

Saved in:
Bibliographic Details
Published in:Journal of personality and social psychology 2021-08, Vol.121 (2), p.441-446
Main Authors: Arslan, Ruben C., Driebe, Julie C., Stern, Julia, Gerlach, Tanja M., Penke, Lars
Format: Article
Language:English
Subjects:
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In Arslan et al. (2018), we reported ovulatory increases in extra-pair sexual desire, in-pair sexual desire, and self-perceived desirability, as well as several moderator analyses related to the good genes ovulatory shift hypothesis, which predicts attenuated ovulatory increases in extra-pair desire for women with attractive partners. Gangestad and Dinh (2021) identified errors in how we aggregated two of the four main moderator variables. We are grateful that their scrutiny uncovered these errors. After corrections, our moderation results are more mixed than we previously reported and depend on the moderator specification. However, we disagree that the evidence for moderation is robust and compelling, as Gangestad and Dinh (2021) claim. Our data are consistent with some previously reported effect sizes, but also with negligible moderator effects. We also show that what Gangestad and Dinh (2021) call an "a priori[...]more comprehensive and valid composite" is poorly justifiable on a priori grounds, and follow-up analyses they report are not robust to a composite specification that we consider at least as reasonable. Psychologists have to become acquainted with techniques such as cross-validation or training and test sets to manage the risks of data-dependent analyses. In doing so, we might learn that we need new data more often than we intuit and should remain uncertain far more often.
ISSN:0022-3514
1939-1315
DOI:10.1037/pspp0000390