The Pew survey I mentioned in my last post also had questions about whether certain characteristics were more true of men or women (people could also say that there was no difference). The characteristics were intellingent, decisive, hard-working, compassionate, emotional, ambitious, arrogant, creative, manipulative, outgoing, and stubborn. Here are the correlations between opinions on these topics and preferences for a girl vs. a boy. A positive sign means that people who thought the characteristic was more true of women were more likely to prefer a girl; a negative sign means that people who think it was more true men were more likely to prefer a girl. You could say that positive signs mean the characteristic is valued--people prefer a child of the sex they think has that characteristic.
Compassionate -.008Stubborn -.015
Manipulative -.059Emotional -.094
For example, people who thought men were more intelligent than women preferred a boy by 54%-21%; people who thought women were more intelligent than men preferred a boy by a much smaller margin 39%-30%. In general the correlations seems reasonable, but I'm surprised that hardworking is the strongest positive association and compassionate make no difference. I'd think it would be much better to have a lazy and compassionate child than a hardworking and hardhearted one.
Of course, this analysis combines men and women. It seems obvious that there will be differences between the sexes--that women will value qualities like compassion and creativity more than men do. It seems obvious, but there's not much evidence that it's true. For example, the correlation of beliefs about compassion and preference is -.052 among men and .015 among women. Neither correlation is statistically significant, and the difference between them isn't either. The general pattern of correlations is similar for men and women.