2018.04.23 03:10 SubsaharanAmerican How to Blackpill an unsuspecting Data Scientist: Have them analyze Columbia University's Speed Dating Experiment Dataset (Fisman, Iyengar, Kamenica, & Simonson, 2006)
Title:Gender Differences in Mate Selection: Evidence from a Speed Dating ExperimentNote the complete absence of any mention of the weights of the covariates in absolute terms (rather than the gender-relative terms as they've done here), forcing the reader to guess whether physical attractiveness is important to women at all. I wish I could say this was something limited to the abstract, but the entire paper reads like this (more on this below).
Author(s):Fisman, Raymond J. Iyengar, Sheena Sethi Kamenica, Emir Simonson, Itamar
Journal Title:Quarterly Journal of Economics
Abstract:We study dating behavior using data from a Speed Dating experiment where we generate random matching of subjects and create random variation in the number of potential partners. Our design allows us to directly observe individual decisions rather than just final matches. Women put greater weight on the intelligence and the race of partner, while men respond more to physical attractiveness. Moreover, men do not value women's intelligence or ambition when it exceeds their own. Also, we find that women exhibit a preference for men who grew up in affluent neighborhoods. Finally, male selectivity is invariant to group size, while female selectivity is strongly increasing in group size.
I remember being slightly shocked upon first viewing the graphs below:https://jonahsinick.com/72/ :
If we average over all participants, we find that participants of above average attractiveness had twice as many suitors as participants of below average attractiveness.
The correlation matrixes (1, 2) give the impression of contradicting a claim in the original study:Similarly, after the dataset was uploaded to Kaggle, young aspiring data scientists were shocked to find the blackpills waiting for them at the end of their number crunching:
Women put greater weight on the intelligence […] while men respond more to physical attractiveness.The apparent contradiction is explained by the fact that the subsets of events that I used were different from the subset of events that the authors reported on in their paper. On one hand, I omitted the events with fewer than 14 people. On the other hand, the authors omitted others:
Seven have been omitted…four because they involved an experimental intervention where participants were asked to bring their favorite book. These four sessions were run specifically to study how decision weights and selectivity would be affected by an intervention designed to shift subjects’ attention away from superficial physical attributes.The intervention of asking participants to bring their favorite book seems to have had the intended effect. One could argue that the sample that I used is unrepresentative on account of the intervention. But to my mind, the intervention falls within the range of heterogeneity that one might expect across real world events, and it’s unclear to me that the events without the intervention give a better sense for gender differences in mate selection across contexts than the events with the intervention do.
A priori one might still be concerned that my choice of sample would lead to me developing a model that gives too much weight to intelligence when the rater is a man. But I chose the features that I did specifically with the intent of creating a model that would work well across heterogeneous speed dating events, and made no use of intelligence ratings to predict men’s decisions.
https://i.imgur.com/QYHAcV1.pngThe problem is, and to Fisman et al credit, a careful reading of the original paper -- and, more importantly, the tables -- show these blackpills were there, hiding in plain sight. The authors just, for whatever reason, decided not to draw attention to them.
The basic results, by gender, are shown in Table III, columns (1) and (2). There is a clear difference in the attribute weights on attractiveness and intelligence: males put more weight on physical attractiveness than females do, while females put more weight on intelligence. This is consistent with the predictions of both the evolutionary and social structure theories of mate selection described in the introduction.I hope the math here is clear. I hope it's also clear why this might be seen as disingenuous. They basically subtracted the female OwnRatings coefficient (0.119) from the males' (0.140), = 0.021, then divided by the female coefficient = "18 percent higher" effect of physical attractiveness. Yes, while technically true, I would argue the more notable finding is the fact that female coefficient is 0.119 (vs the male 0.14) in the first place. Clearly, of the measured covariates, physical attractiveness is the strongest predictor for both sexes. The second major notable finding, IMO, is that the bulk of the explanatory power of attractiveness on the female rater's decisions remains even when only using the average of ratings that OTHER women ("Consensus") gave the male target (column 4).
The magnitudes of these differences are large. Each additional attractiveness point (on a 10-point scale) increases male likelihood of saying Yes by 2.1 percentage points more than it increases the female likelihood of saying Yes. This implies that the effect of physical attractiveness is 18 percent higher for males. The implied effect of intelligence on the probability of Yes is 4.6 percentage points for women compared with 2.3 percentage points for men. We look at the statistical significance of these differences in column (3), where we pool all subjects and include an interaction term RatingMale for each attribute; for both attractiveness and intelligence, the interaction term is significant at the 5 percent level. We do not observe any difference across genders in the importance of ambition. When we repeat the same exercise using the average of all subjects other than i, i.e., Rating-ijc, as the measure of partner attributes, we obtain qualitatively similar results (reported in columns (4)–(6) of Table III).10 Hence, the results are not driven by idiosyncratic assessments of the attributes.
The results are reported in Table IV, columns (1) and (2). For attractiveness, the interaction term is insignificant for both men and women. For ambition, however, the interaction term is insignificant for females but is significantly negative ( p < 0.01) for males. Furthermore, the effect of an increase in ambition above a man’s own level, given by the sum of the direct effect and the interaction term, is negative. In other words, men strictly prefer women with their own level of ambition to women more ambitious than they are. A two-tailed test on the significance of the sum of the coefficient reveals that this effect is statistically significant ( p 0.05). The results on intelligence are qualitatively similar to those on ambition: no slope change for females while for males the slope change at the self-rated level is significant; additionally, the implied effect of increased intelligence above a man’s selfrated level (given by the sum of the two coefficients) is negative, though insignificantly so. When we use Otheric (i.e., the average rating of subject i by his partners on characteristic c) in place of Selfic in columns (3) and (4), we obtain similar results.12 Hence, we demonstrate that on average men do not value women’s intelligence or ambition when it exceeds their own; moreover, a man is less likely to select a woman whom he perceives to be more ambitious than he is.Footnote 12:
One exception is the increased attention to attractiveness that women exhibit toward more attractive men.It's interesting the coefficient I highlighted in the pic, which happens to be one of the highest coefficients among all of the interaction terms, is relegated to brief mention in the footnotes. Its omission as a finding is conspicuous. Following the examples of how the other interaction terms' coefficients have been interpreted, the statement regarding this coefficient should have read: On average, a woman is more likely to select a man who she rates higher than how she, herself, is rated by other men on average (column 3), irrespective of her own self-rating (column 1) This finding may be secondary to the fact women generally rate men harsher than men rate women in this study (and others).
2015.09.13 16:16 WHU_IdeaLab We are the team of IdeaLab!, the biggest student-led founders' conference in Europe. AUA.