We wish to thank Ploderl (2013) for the commentary written in response to our article (Lyons, Lynch, Brewer, & Bruno, 2013) and especially for raising the issue of ecological validity. Moreover,wetoowish towarnreadersagainstmisinterpretingormisunderstanding the results of our study. Unfortunately, we fear that Ploderl’s Letter may actually contribute to the confusion, rather thandiminishit.Specifically,whileweagreethat thereaders shouldbecarefulnot to‘‘apply thefindingsdirectly to the real world,’’at leastuncritically,wefeelPloderlmakesasimilarerror by directly transposing our laboratory-based results into a real, albeit hypothetical, world. In addition, Ploderl appears to misinterpret how gaydar effectiveness was measured in our study. The main issue is clear. In our study, we adopted a 50:50 target-to-lureratio,whichis typicalofdiscriminationtasks,whereas, in the real world, the homosexual-to-heterosexual ratio is closer to 5:95. The commentary contests that the difference in ratio is critical and that we should be careful in drawing conclusions from the data for this reason. Although we agree that changing the base rate to reflect the real world is an interesting empirical question,wedisagreewithPloderl’sconclusionsthat,oncetransposed into the realworld,our resultswouldsuggestgaydareffectiveness to be on average very low and nearing 16 %. Our study investigated whether people were able to correctly categorize individuals (as homosexual or heterosexual) at an above-chancelevel,whenviewingblackandwhitephotographs— thus partly simulating the conditions experienced on dating websites. In other words, we asked whether it was possible to claim that a gaydar exists at all, at least under our experimentally designed condition, or whether correct categorization is mainly the product of lucky (or unlucky, in some cases) guessing. Our results indicated that people could, in fact, categorize sexual orientationcorrectlyatalevel thatwasabovemereguessing.Therefore,ourstudysuggestedthatgaydarcaneffectivelyexist,at least under certain circumstances. In addition, we were interested in determining whether gaydar was more or less effective in raters who were themselves homosexual as compared to heterosexual raters, but we did not find evidence that either group performed better than the other. As stated, our experiment was based on a 50:50 target-to-lure ratio and participants were not made aware of the ratio prior to engaging in the task. This type of procedure is quite common in discrimination research (e.g., Bruno, Higham, & Perfect, 2009), sinceitallowsafairamountofbothpositiveandnegativetrials to judge performance. The likely result of an experimental design with a ratio set to 5:95 is, to us, an empirical rather than a theoretical question (the number of overall items used would need significant increasetoallowforsufficient targettrialsandpower). It is also a question that does not necessarily address real-world relevance. Before addressing the issue of real-world relevance, we raise an important caveat. In our study, we calculated both hit (HR) and false alarm rates (FAR), which Ploderl discusses, and the signal detection statistics d0 and c, which measure discrimination accuracy and bias, respectively. Importantly, d0 and c are independent from each other, meaning that the discrimination abilityofan individualcanbemeasuredindependently fromthat person’s inclination to over/under-use the ‘‘homosexual’’ label (oranyother label, for thatmatter).Therefore,d0 is amuchbetter indexofgaydarability thansimplylookingatHRandFAR—the D. Bruno (&) M. Lyons Department of Psychology, Liverpool Hope University, Hope Park, Liverpool L16 9JD, UK e-mail: brunod@hope.ac.uk
Read full abstract