The piece has a structural problem that undermines its stated purpose. "News that shifts your beliefs" promises rigorous belief-updating, but what's actually on offer is vibe-shifting: replace a scary statistic with a friendlier one, let the juxtaposition do the work, and call it debunking.
The OkCupid section is the clearest case. The Rudder data is revealed-preference: women on a dating platform rating men they might actually message. The counter-studies are stated-preference: survey respondents performing neutrality under social desirability pressure. These are not the same behavioral unit, and conflating them isn't a debunk. "It's not clear why the OkCupid chart looks so different" is not an explanation. It's a shrug dressed as epistemic humility.
The misattributed paternity section commits the same move in a different register. Studies reporting high numbers are dismissed as "biased samples" without any demonstration that the Swedish blood-group study is more representative. The reader is just asked to trust the credibility differential Schubert asserts.
What the piece is actually modeling isn't careful reasoning—it's the *aesthetic* of careful reasoning. Charts and citations do visual-methodological labor without doing analytical labor. The arguments are barely made, the counter assertions are still expecting a lot of unearned credulity from the audience, just like the original assertions supposedly being debunked. The whole thing gives off paltering vibes.
Yeah this is a pretty basic error in understanding what the OK cupid chart is showing. It's a shame too, because it could have been an interesting thing to look in to
Also, it's not just the revealed vs stated preference, but also the fact that OkCupid is a dating site, and thus only shows datingsite-specific ratings. Part of the claim isn't just "women find men less attractive", but "on dating apps women are much more selective than men". It is now a well-known argument that having access to very attractive people (whether through dating apps or Instagram pictures) is depressing the dating market because everyone gets unreasonable standards. The OkCupid chart could also be read as part of that trend, and just shows that the effect is stronger for women than men.
So I agree, the coverage in this article of the chart was rather lazy, which is a shame!
I think the distinction between home production and leisure is not clear cut enough to support this analysis. Nobody likes doing laundry, but cooking, gardening, shopping, home renovation, child care, and more are all activities that are both home production and things people intentionally want to do more of.
Home production takes more labor than paid production for similar output. An hour of shopping, prep, cooking and serving at home produces a meal for one small family. An hour of work in a family restaurant may contribute 10% of the labor involved in providing meals to 100 families.
The piece has a structural problem that undermines its stated purpose. "News that shifts your beliefs" promises rigorous belief-updating, but what's actually on offer is vibe-shifting: replace a scary statistic with a friendlier one, let the juxtaposition do the work, and call it debunking.
The OkCupid section is the clearest case. The Rudder data is revealed-preference: women on a dating platform rating men they might actually message. The counter-studies are stated-preference: survey respondents performing neutrality under social desirability pressure. These are not the same behavioral unit, and conflating them isn't a debunk. "It's not clear why the OkCupid chart looks so different" is not an explanation. It's a shrug dressed as epistemic humility.
The misattributed paternity section commits the same move in a different register. Studies reporting high numbers are dismissed as "biased samples" without any demonstration that the Swedish blood-group study is more representative. The reader is just asked to trust the credibility differential Schubert asserts.
What the piece is actually modeling isn't careful reasoning—it's the *aesthetic* of careful reasoning. Charts and citations do visual-methodological labor without doing analytical labor. The arguments are barely made, the counter assertions are still expecting a lot of unearned credulity from the audience, just like the original assertions supposedly being debunked. The whole thing gives off paltering vibes.
Yeah this is a pretty basic error in understanding what the OK cupid chart is showing. It's a shame too, because it could have been an interesting thing to look in to
Also, it's not just the revealed vs stated preference, but also the fact that OkCupid is a dating site, and thus only shows datingsite-specific ratings. Part of the claim isn't just "women find men less attractive", but "on dating apps women are much more selective than men". It is now a well-known argument that having access to very attractive people (whether through dating apps or Instagram pictures) is depressing the dating market because everyone gets unreasonable standards. The OkCupid chart could also be read as part of that trend, and just shows that the effect is stronger for women than men.
So I agree, the coverage in this article of the chart was rather lazy, which is a shame!
I think the distinction between home production and leisure is not clear cut enough to support this analysis. Nobody likes doing laundry, but cooking, gardening, shopping, home renovation, child care, and more are all activities that are both home production and things people intentionally want to do more of.
Home production takes more labor than paid production for similar output. An hour of shopping, prep, cooking and serving at home produces a meal for one small family. An hour of work in a family restaurant may contribute 10% of the labor involved in providing meals to 100 families.
This is true but irrelevant for the issue of time use comparison.
People think the OK Cupid data and General Social Survey data are incongruent.
They are not, it is essentially the difference between the median of a distribution and the edge that produces both effects.
0.5-1.0 standard deviation in attractiveness is sufficient for both to occur.