How EIRD Is Sex Research?: A Commentary and Reanalysis of Klein et al. (2021)

ABSTRACT

Klein, Savaș, and Conley (2021) argued that sexual science is overdependent on WEIRD (Western, Educated, Industrialized, Rich, and Democratic) samples. Though we agree that sexual science needs to increase its generalizability and inclusivity, we describe concerns with their measurement strategy of categorizing samples as WEIRD or Not WEIRD based on the country from which a sample was drawn. Reanalyzing their data with publicly available global metrics of Education, Industrialization, Richness, and Democratic Values (what we refer to as EIRDness), we find (1) EIRDness metrics were not particularly correlated; (2) countries coded as WEIRD by Klein et al. do not appear reliably EIRDer than those that were not; and (3) and categorical measurement models of EIRDness did not support profiles of EIRD and Not EIRD countries. With these limitations in mind, we then express further concerns about the application utility of Klein et al.’s WEIRDness critique, and unintended political implications embedded in its methodology. We conclude by harkening back to critiques of the WEIRD framework, and suggest that the pursuit of a more equitable and just sexual science – which we applaud Klein et al. for pushing our field to consider – may be better served to alternative frameworks for critiquing its sampling practices.

This article refers to:

How WEIRD and Androcentric Is Sex Research? Global Inequities in Study Populations

Acknowledgement

We wish to thank Carm De Santis and Dr. Alex Williams for their helpful feedback on this manuscript.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

¹ One of us (Sakaluk) was a reviewer for V. Klein et al. (2021) and raised this concern on two occasions.

² Supporting analyses that are only made possible by Klein et al. sharing their data file–exemplary transparent research practices by Klein et al. for which we are grateful.

³ A mere coincidence: the Klein of ManyLabs 2 is a different Klein than the V. Klein et al. (2021) under discussion here.

⁴ Sakaluk had recommended this methodological strategy and these open data sources to Klein et al. (either to replace their dichotomized approach, or at least supplement it), but this suggestion was declined.

⁵ Standardization of variables occurred in the larger dataset of all EIRDness values for all nations that we could find, of which the values for the nations in V. Klein et al.’s (2021) sample were just a subset. The sample mean and standard deviation of these values in the Klein et al. subset should therefore not be expected by readers to be 0 and 1, respectively.

⁶ Adjudicating this claim and the impact Klein et al.’s dichotomization strategy has on it becomes quite complicated if one takes our aggregation critique seriously. If a reader feels their analysis ought to be a tabulation of what percentage of samples are WEIRD in all/almost all respects (e.g., W, but also E, I, R, and D), then this is surely an overestimate of WEIRDness. Or rather, if a reader feels their analysis ought to be a tabulation of what percentage of samples are WEIRD in at least one respect (i.e., W and/or E, and/or I, and/or R, and/or D), then this is surely an underestimate of WEIRDness.