In March 2013, the IMGAME project went to the Department of Economic Psychology in the Polish University of Social Sciences and Humanities Faculty in Wroclaw to conduct Imitation Games on the topic of sexuality; specifically, to what extent can straight men pass as gay. The topic is interesting because of the changing attitudes to homosexuality in Poland. As the tourist guide to Wroclaw puts it:
Do not be afraid of visiting Poland because you are gat. Just bear in mind when you are here that you are not in The Netherlands and the situation in Polish cities cannot be compared to that found in London’s Soho or Madrid’s Chueca. Not yet. (Source: The Visitor Dolny Slask, Vol. 55, p. 14)
In 2013 we found that, somewhat to our surprise, that the Pass Rate for straight men pretending to be gay was 0.58; roughly on a par with the ability of non-Christians to pretend to be Christian’s in Scandinavian countries. There several possible explanations for this: first, our expectations were wrong and 0.58 is the correct answer; second, that the gay community in Wroclaw is relatively amorphous so that gay men have only a limited range of shared experiences to draw on, which in turn makes it easier for Pretenders to provide plausible answers; and third, that there some as yet unknown features of the sample or procedure that influenced the result.
To distinguish between these possible explanations, we returned to Wroclaw in March 2014 and attempted to replicate the study. As they were so efficient in 2013, we recruited the same local organising team – Jakub Traczyk, Agata Sobkow and Kasia Kulwicka-Durmowicz — and also kept the instructions and recruitment procedures as identical as possible. Thanks to their efforts, we were able to run one Step One session with 40 players (20 gay participants and 20 straight participants) on Friday 21 March. This is a minor change from 2013, when we only had 34 participants.
Over the weekend (one change from 2013, when Step Two ran on a Tuesday and Wednesday) we then recruited just over 190 straight Pretenders to take part in Step Two. Final judging will be done over the next few weeks when, in another change from last year, the ever-developing software will be able to host Step Four directly, rather than requiring us to email a bespoke Word document to each of the 40 Judges.
Given this we do not have any large sample results to report as yet. We can, however, summarise the results of Step One and compare them to last year. Superficially, there appears to be a marked difference in the Pass Rate for straight Pretenders when compared to 2013, as shown in the Table below.
|2013 Pass Rate||2014 Pass Rate|
|Gay Judge / Straight Pretenders||0.71||0.40|
|Straight Judge / Gay Pretenders||0.88||0.85|
In contrast, the ability of gay pretenders to pass as straight is almost completely unchanged. Several important caveats need to be stated, however, before any conclusions are drawn. First, as shown by the fact the final pass rate in 2013 was 0.58, compared to a Step One Pass Rate of 0.71, shows that Step One is a poor predictor of Step Four.
This is directly related to the second caveat: sample sizes in Step One are very small (n = 34 in 2013 and 40 in 2014). As such, it is possible for small changes in frequencies to have a relatively large impact on the summary statistic. This is particularly apparent in this case as the distribution of answers for Gay Judges from 2013 and 2014 seems very similar: a low number of wrong guess, no Don’t Knows and a majority of correct identifications (see Table below):
Given this, the most reliable conclusion that can be drawn from the data available so far is that there is no evidence of any change from last year. We therefore await Step Four results with interest!