9 thoughts on “Against Kanazawa

  1. Kaufman & Wicherts’ response to Kanazawa is kind of a mixed bag. It leads off with what I think is a weak critique, which relates to Kanazawa’s use of the term “objective”. Here I think they fall into the same problem we’ve been discussing in another thread – which is that, by “objective” attractiveness, Kanazawa was (as I understand it) referring to something other than the authors surmise.

    The authors present their next point – that they can’t tell whether Kanazawa took into account a significant divergence in ratings across interviewers – as being related to the first. Once we correct for the apparent misunderstanding over the “objective” point, though, I’m not sure how significant the consequences of this variation are.

    The next critique offers both a rather strong point and several weak points. The strong point the authors raise is that Kanazawa’s article did not address the fourth and, so far, last phase (Wave IV) of Add Health, which covered the participants at the oldest age range (data collected when the participants were 25 to 34). They note that by Wave IV, the differences in attractiveness ratings were no longer significant across ethnicities. You would never know this was the case from reading Kanazawa’s article. (They also argue that the perceptible differences in Wave III (ages 18-26) – are nonetheless insignificant after taking into account “random variation due to the raters”, but they don’t show their work here.)

    However, the authors turn this into what I perceive as another dubious semantic critique – that, in light of the foregoing, Kanazawa shouldn’t have used the term “women” (or “men”) because Waves I/II covered the participants with a total range of age 12-22. The average for both groups’ participants was somewhere between their 16th and 17th birthdays. The authors conclude that only in Waves III and IV were the participants old enough to be called “women” and “men” instead of “girls” and “boys”. It’s hard to believe the authors would squander the very palpable hit they score with the Wave IV revelation on such a critique. Even in social science literature, and even in an era arguably marked by the progressive infantilisation of adolescence, it’s not exactly rare to include individuals in the 16+ and 15+ range – sometimes even lower! – within the rubric of “women” and “men”. The authors seem to get hung up on the juridical artifice of the currently prevailing age of majority (notwithstanding that one certainly comes across notions like “a woman who is a minor” in the law).

    In another puzzling moment, the authors say that the ratings for Waves I/II “couldn’t possibly (we hope!)” reflect a reaction to *sexual* attractiveness of the participants. Because, you know, some level of sexual attraction to the appearance of individuals age 15-22 that made up the majority of Waves I/II would be abnormal! Trained psychologists writing this stuff, folks.

    Anyway, the authors have done a big favour by opening up the data a bit more to readers, where Kanazawa didn’t.

  2. Nemo, I think it would be helpful if you explained here your view about the import of ‘objective’.

  3. JJ, I take your point. I should really learn how to do hyperlinks to other posts. But just to recap:

    When Kanazawa refers to objective physical attractiveness, I understand him to be referring to the observed tendency of the person’s appearance to trigger an actual attraction response in other individuals, as measured by the test, regardless of the reasons. I think he was also using the term to distinguish it from what he calls “subjective” (i.e. self-rated) attractiveness.

    There are certainly other commonly used senses of “objective” that are consistent with the points raised by Kaufman & Wicherts and many others, and they pretty clearly one or more of those other senses applied here. Probably Kanazawa should have anticipated this and better articulated his qualified use of “objective attractiveness” – assuming for the sake of argument here that I’m not mistaken about what he really intended.

    One thing that made this confusion perhaps all the more inevitable is the conjuction of (i) the apparent (I’d say regrettable) interchangeability of “attractiveness” and “beauty” in many people’s vocabularies; with (ii) the fact that “objective beauty” has acquired a particular meaning in aesthetics. Thus, many people hear “objectively attractive” as “objectively beautiful” – a term of art – and think Kanazawa is concluding something along the lines of “black women’s appearances are less intrinsically worthy of appreciation” (or insert some similar formulation here from your preferred philosopher of the objectivity of beauty).

  4. Where I said “they pretty clearly one or more” in the post above, I meant to write “they pretty clearly thought one or more”. Sorry.

  5. Nemo, I think that what you have in mind is close to the idea of intersubjective. The trouble is such ideas do not validate Kanazawa’s language.

    For example, if I do a study and announce that asparagus is objectively disgusting, I can be faulted becaused of my sample. E.g., it shouldn’t be just 3 and 4 year olds, unless I explicitly say so. In addition, it should not be people who live an an anti-asparagus community. The problem with each group is that, for different reasons, we can’t generalize from the group reaction to reactions in a less restricted population.

    We don’t know what the population that K drew on is like. That’s one huge fault. In any case, the chances are that they are racist. Among other things, like many, many people in the US, they’d probably do poorly on IAT tests for racism. Worse still, they might be unaware of how they’d do.

  6. JJ, good post. I think the concept of intersubjectivity is relevant here, and but I’m not sure it would do to, for example, simply replace “objective” with “intersubjective” in Kanazawa’s article. For example, I alluded in the other thread to the idea that Kanazawa might be using “objective” in a way similar to a pollster might conclude that Smith is objectively more popular than Jones. It would not seem equivalent to say that that it is intersubjectively true that Smith is more popular than Jones.

    Anyway, I pretty much agree with your comments about sample (though the relative merits of IAT are a separate matter). Absolutely, someone can be faulted on the sample. However, from what I can see, most of even the sample-focused critiques of Kanazawa haven’t really taken the tack that the sample was non-representative, rather they’ve come back round somehow to the use of “objective”. For example, you seemed to imply in the other thread that the problem with the respondents was not that the sample was non-representative but that the people in it were likely biased (they weren’t bias-free robots, etc.). In fact, you’ve said above that the individuals in the sample were probably racist, like many, many people in the US – so I take your basic concern to be not that K’s sample was unrepresentative (like the 3 and 4 year olds in your asparagus example), but that it was representative. Of course, the sample population might not be representative of some larger population, (and we could have an interesting discussion about what the sample should really be representative of to back up K’s assessment).

    Actually, the problems with the sample appear different depending on how one interprets K’s use of “objective”. For the sense I think he intended, it’s important to the validity of his conclusion that the sample be representative but not that it be bias-free. If you take him in another sense, it becomes important to the validity of his conclusion that the sample be bias-free but not that it be representative. Most people seem to be in the latter camp, and this is reflected in the tenor of their criticisms of the sample.

Comments are closed.