From the U.S. to Denmark and Beyond: It’s Time to Reimagine the Use of Control Variables in Observational Research

ABSTRACT

Methodologists from a variety of social and behavioral sciences have called attention to misunderstandings, misclassifications, and misapplications of third variable “controls” in observational research. We are unaware, however, of a comprehensive discussion of these issues in the sexological literature. In this article, we attempt to detail several of the more important potential pitfalls within the context of a case study commentary on a recent Journal of Sex Research inquiry into Danish persons’ pornography use, sexual satisfaction, and sexual behavior. We emphasize that our own (still limited) edification came only through personal error and chart an optimistic path forward wherein the current state of practice can be transformed into theoretical progress and innovation.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Supplementary Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/00224499.2024.2379559.

Notes

1 It should be noted that when the Kendall’s Tau-b correlation between more frequent pornography consumption and lower sexual satisfaction among men (−.147) reported by Chapman et al. (2023) is converted to r (see Walker, 2003, for the conversion formula), its magnitude (−.229) is actually larger than the upper bound of our meta-analytic CI (r = −.16). Thus, rather than the Chapman et al. (2023) correlation between more frequent pornography consumption and lower sexual satisfaction among men being weaker than previous findings (as Chapman et al.’s Abstract might be read to suggest), it is stronger.

2 Please see the same (Wright, 2021) for the first author’s confession that they have been guilty of the same, many times over. Hopefully their peers will accept this and their other efforts to right these scientific wrongs as acceptable penance (e.g., Wright & Tokunaga, 2023; Wright et al., 2022).

3 As in most areas of the social and behavioral sciences involving controversial topics, there are important debates and deliberations beyond those raised in the present article in the pornography effects literature (e.g., regarding measurement, study design and inference, and replicability). Interested scholars, for example, might consider readings such as the following: Bridges (2023), Fisher and Kohut (2017), Kohut et al. (2020), Malamuth (2018) and Wright (2022b).