Jeez, how long has it been since we had a good chat about comely girls? Via Dienekes, this article shows that in a sample of early 20-something girls at a Scottish university, daughters of happily married parents (HM) had more attractive and feminine faces than those whose parents either had separated pre-puberty (S) or had remained married but unhappily (UM). Color pictures in PDF here (L-to-R: S, UM, HM). As for body shape, the HM daughters had lower BMI (i.e., were slimmer) and lower waist-to-hip ratio (WHR, i.e., a more hourglass figure) than the other two groups. The authors make it clear that theirs is a correlational study that remains agnostic on the causal mechanisms, though Dienekes’ post & comments discuss possibilities.
First, though, there’s an interesting wrinkle: the authors and some of their references argue that the independent variable is presence of father (or parental cohesion), which yields the order S -> UM -> HM. The dependent variables are the markers of attractiveness, etc. However, the functions do not always turn out to be monotonic — some curves decrease and switch to increasing like a U. Table 2 in the article summarizes the rank-ordering on the 3 facial markers. For “attractiveness,” the order is UM -> S -> HM. For “health,” it’s UM -> S = HM. And for “masculinity,” it’s UM = S -> HM. Thus, if our graph has “parental cohesion” on the x-axis, the curve would be U-shaped for “attractiveness” and “health,” though J-shaped for “masculinity.” For body shape, Figure 2 shows that one variable, “impedence” (a measure of % body fat), wasn’t significantly different among the groups. “Waist-to-chest ratio” (WCR) is a measure of the inverted-triangle shape of the upper body, something that men don’t pay much attention to. The two important variables are BMI and WHR. On the latter, HM had significantly more hourglass figures than UM, though the S didn’t differ significantly from either. On BMI, the HM were significantly slimmer than S, and apparently UM are in between.
So, aside from perhaps BMI, the facial and bodily markers suggest that the underlying cause increases thusly: UM -> S -> HM. Call it “home harmony.” The biological correlate of this in the literature the authors cite is response to stress (cortisol), so perhaps in the sample the daughters of UM parents experienced greater stress from the arguing, bickering, and so on, compared to the daughters of S parents, who at least weren’t frequently fighting in the daughter’s presence. Remember, the sample was of university students, so it likely didn’t include those from the underclass or the lower end of the working class. That suggests that, above a certain threshold of SES, having antagonistic parents stay together produces more dissonance in the home than if they separated, at least from the daughter’s p-o-v.
Now on to the possible causes. Well, the first is what I just mentioned: the prevailing view that differences among father absent or father present homes reflect differences in stress during childhood. Humans have evolved a common set of responses, and those who happen to grow up without fathers turn out a different way from those who grow up with fathers. This is an environment / chance explanation. But as mentioned at the lead author’s webpage and in the comments at Dienekes’ post, there is also a (not mutually exclusive) genetic explanation: fathers who are apt to easily leave their wife & children, or who are too unruly for the wife to bother staying with, could be this way in part due to genes (perhaps for response to testosterone) which they pass on to their daughters.
Moreover, since it’s usually not impossible to read warning signs about who’s less reliable & dependable than who else, we could also look at the mothers who mate with the more flight-prone males. Such females would likely show a greater propensity for risk-taking or thrill-seeking, presumably heritable, so assortative mating could be exacerbating the genetic influence of father. It would be interesting to take a large sample of males from S, UM, and HM parents and ask them who they were most attracted to among the S, UM, and HM female composite faces. That would settle whether there was an assortative mating effect. It would also be interesting to genotype those from S vs HM parents to see if the former were more likely than expectation to have the 7R allele at the DRD4 locus — if so, that would suggest involvement of heritable personality traits like novelty-seeking, impulse control, and so on, in both the dissolution of the marriage as well as the suite of behavioral outcomes of the daughter. It might also suggest why the daughters of S parents were judged more attractive than those from UM parents — presumably the S parents were more “wild child”-like than the UM parents who had to “wimp out” to some degree in suppressing their impulses to split up. Perhaps greater “wild child”-ness increases one’s sexiness score (since “sexy” usually connotes something more exciting or thrilling than just “attractive” or “beautiful”). The prototype here would be Angelina Jolie, who looks more than a bit masculine, who’s well known to be possessed of a thrill-seeking disposition, and whose parents divorced when she was a baby.
To close, why would a tinge (though not an excess) of masculinity and rebelliousness make a female sexier, when these usually serve to make males sexier? These traits mix a “danger” component with the “beauty” component, which creates the thrill. My hypothesis is that, assuming the “cheesecake theory” of aesthetic pleasure popularized in — that art, cuisine, etc. are human devices to directly stimulate our evolved pleasure centers — we enjoy stimulating not just the “relaxation” (or parasympathetic) division of our nervous system, but occasionally the “danger” (or sympathetic) division as well. Things that highly stimulate these divisions are, respectively, the beautiful and the sublime in an older terminology. Those whose aesthetic preferences lead them to want more than others to stimulate their “danger” system would appreciate a greater dosage of dangerous, masculine traits in females. Rawrrr.