MPs more likely to respond to women's requests, driven by female legislators bias against men

[Metaphor]

In particular, I don't know why the first author (Gabriele Magni) is surprised, considering his own research shows that American voters prefer black women and Asian women as political candidates, compared to white men.

It seems that Republicans, Conservatives and the religious prefer the white man over the black or asian woman, but those of other political allegiance prefer the reverse.

Yes. But overall, the numbers favour black and Asian women.

 

[ruby sparks]

In particular, I don't know why the first author (Gabriele Magni) is surprised, considering his own research shows that American voters prefer black women and Asian women as political candidates, compared to white men.

It seems that Republicans, Conservatives and the religious prefer the white man over the black or asian woman, but those of other political allegiance prefer the reverse.

Yes. But overall, the numbers favour black and Asian women.
View attachment 28934

I know. That's where I got it.

And...

"In a survey, however, respondents sometimes provide untruthful answers. For instance, they may hide their dislike for Black candidates for fear of revealing their racism. To address this concern, we embedded in the survey a conjoint experiment that allowed us to elicit true voters’ preferences. We presented respondents with pairs of candidates who varied in their demographic traits. We looked at gender, race, sexual orientation, gender identity, health, religion, education, age, and political experience. And then we asked them to choose which candidate they would be more likely to support. Because we randomized these traits, we can estimate the effect of each candidate characteristic and the interaction of candidate characteristics (e.g. gender and race) on vote choice."

I don't understand how this gets around the tendency for some not to give truthful answers.

 

[Metaphor]

In both the US House and Senate, page 1 and 2 of the report, women get better results than men in getting elected. That's what the 'US Congress' part was about.

Your first quoted statement, about state legislative elections, does not somehow invalidate that women are more likely to win overall.

It seems to show that in the cases you selected, Democrat women do better, and in the ones I selected, they don't.

But the point is that overall, women do better or the same as men, when they stand for election.

 

[ruby sparks]

But the point is that overall, women do better or the same as men, when they stand for election.

Overall in what type of election?

 

[Metaphor]

But the point is that overall, women do better or the same as men, when they stand for election.

Overall in what type of election?

In US House and Senate elections.

In state races, it seems there is no difference in incumbent winning rates between sexes. But women are more likely to win as non-incumbents than men as non-incumbents.

In state elected executive positions, Republican women were more likely to win than Republican men, but Democrat women were less likely to win than Democrat men.

 

[ruby sparks]

I don't know enough about the cultural forces in Latin America that might make this happen...

If I had to guess, I'd say that in 'western', more developed countries there's now more of a push to assist and enable women (and possibly minorities). This general trend may be showing up in more than one of the studies you have cited.

 

[ruby sparks]

But the point is that overall, women do better or the same as men, when they stand for election.

Overall in what type of election?

In US House and Senate elections.

In state races, it seems there is no difference in incumbent winning rates between sexes. But women are more likely to win as non-incumbents than men as non-incumbents.

In state elected executive positions, Republican women were more likely to win than Republican men, but Democrat women were less likely to win than Democrat men.

Yes. I was replying to your specific point about Democrat women, which seemed to be the case in some types of election but not others.

 

[Metaphor]

Yes. But overall, the numbers favour black and Asian women.
View attachment 28934

I know. That's where I got it.

And...

"In a survey, however, respondents sometimes provide untruthful answers. For instance, they may hide their dislike for Black candidates for fear of revealing their racism. To address this concern, we embedded in the survey a conjoint experiment that allowed us to elicit true voters’ preferences. We presented respondents with pairs of candidates who varied in their demographic traits. We looked at gender, race, sexual orientation, gender identity, health, religion, education, age, and political experience. And then we asked them to choose which candidate they would be more likely to support. Because we randomized these traits, we can estimate the effect of each candidate characteristic and the interaction of candidate characteristics (e.g. gender and race) on vote choice."

I don't understand how this gets around the tendency for some not to give truthful answers.

I suspect (and again, I haven't had a chance to read the study), that hypothetical candidates were 'built', and different people were asked to rate these candidates.

So, for example, candidate A was built to have all the same traits (let's say, straight, cis, no disability, Christian, university graduate, 44 years old, 10 years in politics) as candidate B, except candidate A was male and white, and candidate B was a black woman.

Overall, people preferred candidate B, even though no single person was asked about both candidate A and B.

 

[ruby sparks]

So, for example, candidate A was built to have all the same traits (let's say, straight, cis, no disability, Christian, university graduate, 44 years old, 10 years in politics) as candidate B, except candidate A was male and white, and candidate B was a black woman.

Overall, people preferred candidate B, even though no single person was asked about both candidate A and B.

Yes, but then, if someone is inclined to say they prefer candidate B, because of their gender or skin colour, because it makes them seem more liberal, then they would do that. So I don't see how it gets around that.

 

[Metaphor]

So, for example, candidate A was built to have all the same traits (let's say, straight, cis, no disability, Christian, university graduate, 44 years old, 10 years in politics) as candidate B, except candidate A was male and white, and candidate B was a black woman.

Overall, people preferred candidate B, even though no single person was asked about both candidate A and B.

Yes, but then, if someone is inclined to say they prefer candidate B, because of their gender or skin colour, because it makes them seem more liberal, then they would do that. So I don't see how it gets around that.

So, it's a 'liberal' thing to do to prefer black and Asian women, solely on the basis of sex and race?

What percentage of respondents do you think are lying to themselves and others about their actual preferences? And, how do you interpret this lying in a particular direction?

I mean, let's say that most people are not actually biased by race and sex and don't prefer Asian and black women to white men. But they are lying about and saying they do prefer it. What has happened to American culture that preferring black and Asian women to white men is seen as socially desirable and yet simultaneously is a position nobody really holds?

 

[ruby sparks]

So, it's a 'liberal' thing to do to prefer black and Asian women, solely on the basis of sex and race?

If I'm using the word liberal the right way (as in progressive, non-conservative) then yes, I think so.

What percentage of respondents do you think are lying to themselves and others about their actual preferences? And, how do you interpret this lying in a particular direction?

I'm referring to something the researcher suggested, and said they controlled for, but I don't see how they did. In the current political climate, I'd guess that when there's fibbing, it's fibbing to seem more right on. Which is what the researcher suggested.

I mean, let's say that most people are not actually biased by race and sex and don't prefer Asian and black women to white men. But they are lying about and saying they do prefer it. What has happened to American culture that preferring black and Asian women to white men is seen as socially desirable and yet simultaneously is a position nobody really holds?

I don't understand you there. Who said it was a position nobody holds? It just means that some of those that don't espouse it might want to seem as if they do, in the current political climate (heightened awareness, wokeness etc). It would in those cases be lip service to that.

 

[Emily Lake]

I know. That's where I got it.

And...

"In a survey, however, respondents sometimes provide untruthful answers. For instance, they may hide their dislike for Black candidates for fear of revealing their racism. To address this concern, we embedded in the survey a conjoint experiment that allowed us to elicit true voters’ preferences. We presented respondents with pairs of candidates who varied in their demographic traits. We looked at gender, race, sexual orientation, gender identity, health, religion, education, age, and political experience. And then we asked them to choose which candidate they would be more likely to support. Because we randomized these traits, we can estimate the effect of each candidate characteristic and the interaction of candidate characteristics (e.g. gender and race) on vote choice."

I don't understand how this gets around the tendency for some not to give truthful answers.

I can't speak to the methodology in this specific survey. But in general, a well-designed conjoint survey can weed out untruths and some biases. As a very simplified example, we might ask if people prefer salad or french fries as a side dish. Let's assume that we think there's a reasonable chance that people will say they prefer salad, but that they might actually prefer french fries - it might be that they're trying to appear more health-conscious than they actually are... but it might also believe that they really do prefer salad and are unaware of their own patterns and biases. So you give them several questions that pair salads, french fries, green beans, mashed potatoes, and yams with a variety of main dishes. Then you go through the responses in aggregate and find the patterns. It might turn out that even though some people claimed they prefer salads, they consistently selected meal combinations that included french fries over those that included salads.

The difficulty is that it can be difficult to design a conjoint survey that controls for all of the influencing factors. Good conjoint studies tend to be very long surveys... and long surveys are expensive - people don't take those kinds of surveys for free, and the longer it is the more you've got to pay people to spend time taking a survey, and the smaller the pool of people interested in taking it. So there's also a lot of risk of selection bias. It can be very difficult to ensure that you've got a valid sample that is large enough that it can be extrapolated to a population.

Conjoint surveys can be extremely good methods of collecting data that is often hard to quantify. But the methodology, sample structure, and design ends up being extremely important. It can be difficult to judge whether or not a conjoint study is well designed, or if it has inherent skew or bias that is unaccounted for.

A sample size of 2000 is a decent size, depending on how many conjoint pairs they evaluated. If it was on the order of 10 to 15 pairs, that's probably a good size, provided the randomization of demographic and policy elements presented was sufficiently random (harder than you might think). Without access to the demographic profile of the respondents, it's difficult to say for sure. It at least passes first muster for sample size, so it's probably not complete garbage

 

[J842P]

Yes, cultural norms are different in Western societies than they are in Latin America.

Maybe just a nitpick, but Latin American societies are generally considered Western Societies.

This should be obvious from the name... they follow Western religions, speak Western languages, etc...

 

[J842P]

In particular, I don't know why the first author (Gabriele Magni) is surprised, considering his own research shows that American voters prefer black women and Asian women as political candidates, compared to white men.

It seems that Republicans, Conservatives and the religious prefer the white man over the black or asian woman, but those of other political allegiance prefer the reverse.

Yes. But overall, the numbers favour black and Asian women.
View attachment 28934

I wouldn't have expected the disparity to be greater for Asian women for the groups that are less likely to favor either group of women... I mean, take a look at the "Religious". You have like a -7% compared to -1% Asian vs Black. Maybe, because Black women tend to be visibly religious?

 

[Metaphor]

I'm referring to something the researcher suggested, and said they controlled for, but I don't see how they did. In the current political climate, I'd guess that when there's fibbing, it's fibbing to seem more right on. Which is what the researcher suggested.

I mean, let's say that most people are not actually biased by race and sex and don't prefer Asian and black women to white men. But they are lying about and saying they do prefer it. What has happened to American culture that preferring black and Asian women to white men is seen as socially desirable and yet simultaneously is a position nobody really holds?

I don't understand you there.

It's about, as you say, 'fibbing'. If people don't really prefer black and Asian women, why are they lying and saying they do? Why are they concerned about being "right on"? Why is it regarded as more socially desirable to be biased towards black and Asian women versus white men? And further, why is it regarded as more socially desirable if in fact the majority of people do not hold such a position?

I would say that, if most people do not hold the position but it is regarded as socially desirable to say you do, then there is heavy cultural influence from establishment media pushing the viewpoint that it's better to prefer black and Asian women to white men--an establishment media that does not, then, reflect the mainstream in American society.

 

[ruby sparks]

I know. That's where I got it.

And...

"In a survey, however, respondents sometimes provide untruthful answers. For instance, they may hide their dislike for Black candidates for fear of revealing their racism. To address this concern, we embedded in the survey a conjoint experiment that allowed us to elicit true voters’ preferences. We presented respondents with pairs of candidates who varied in their demographic traits. We looked at gender, race, sexual orientation, gender identity, health, religion, education, age, and political experience. And then we asked them to choose which candidate they would be more likely to support. Because we randomized these traits, we can estimate the effect of each candidate characteristic and the interaction of candidate characteristics (e.g. gender and race) on vote choice."

I don't understand how this gets around the tendency for some not to give truthful answers.

I can't speak to the methodology in this specific survey. But in general, a well-designed conjoint survey can weed out untruths and some biases. As a very simplified example, we might ask if people prefer salad or french fries as a side dish. Let's assume that we think there's a reasonable chance that people will say they prefer salad, but that they might actually prefer french fries - it might be that they're trying to appear more health-conscious than they actually are... but it might also believe that they really do prefer salad and are unaware of their own patterns and biases. So you give them several questions that pair salads, french fries, green beans, mashed potatoes, and yams with a variety of main dishes. Then you go through the responses in aggregate and find the patterns. It might turn out that even though some people claimed they prefer salads, they consistently selected meal combinations that included french fries over those that included salads.

The difficulty is that it can be difficult to design a conjoint survey that controls for all of the influencing factors. Good conjoint studies tend to be very long surveys... and long surveys are expensive - people don't take those kinds of surveys for free, and the longer it is the more you've got to pay people to spend time taking a survey, and the smaller the pool of people interested in taking it. So there's also a lot of risk of selection bias. It can be very difficult to ensure that you've got a valid sample that is large enough that it can be extrapolated to a population.

Conjoint surveys can be extremely good methods of collecting data that is often hard to quantify. But the methodology, sample structure, and design ends up being extremely important. It can be difficult to judge whether or not a conjoint study is well designed, or if it has inherent skew or bias that is unaccounted for.

A sample size of 2000 is a decent size, depending on how many conjoint pairs they evaluated. If it was on the order of 10 to 15 pairs, that's probably a good size, provided the randomization of demographic and policy elements presented was sufficiently random (harder than you might think). Without access to the demographic profile of the respondents, it's difficult to say for sure. It at least passes first muster for sample size, so it's probably not complete garbage

I think I understand some of that. In a nutshell, you confuse them, to weed out what they really think?

Anyhows, let's say it at least partially worked in this case. Would that mean that many people (in this study, most, overall) do actually prefer black/female candidates?

I suppose the next questions are (a) if so, why, and (b) whether it is a good or a bad thing.

If I took this survey, I myself might (I'm not certain) come out favouring a female candidate (or a black or other minority female candidate over a white man let's also say). Or, maybe that's just me virtue-signalling, and the conjoin test would weed me out.

But let me indulge in the vanity that I did favour females. If asked why, I'd say that I think more women in positions of power is a good thing, because (a) I think they might be better in the job and (b) they are and have been underrepresented. I wouldn't pick a woman just because she's a woman, obviously, but if all other things seemed fairly equal, I might. And to be honest, it's my impression that things aren't necessarily often equal, for me, because it's my impression that women say more things that I like to hear. As to blackness, I think it might be more (b) that might sway me.

I think a matriarchy is at least something we should try out.

 

[Metaphor]

Yes. But overall, the numbers favour black and Asian women.
View attachment 28934

I wouldn't have expected the disparity to be greater for Asian women for the groups that are less likely to favor either group of women... I mean, take a look at the "Religious". You have like a -7% compared to -1% Asian vs Black. Maybe, because Black women tend to be visibly religious?

It'd be difficult to say. "Religious" is not a very good group descriptor though. People of different religions can all describe themselves as 'religious' but the difference between a religious Christian evangelical, a religious Muslim, and a religious Wiccan are likely to be vast.

 

[J842P]

Yes. But overall, the numbers favour black and Asian women.
View attachment 28934

I wouldn't have expected the disparity to be greater for Asian women for the groups that are less likely to favor either group of women... I mean, take a look at the "Religious". You have like a -7% compared to -1% Asian vs Black. Maybe, because Black women tend to be visibly religious?

It'd be difficult to say. "Religious" is not a very good group descriptor though. People of different religions can all describe themselves as 'religious' but the difference between a religious Christian evangelical, a religious Muslim, and a religious Wiccan are likely to be vast.

No doubt, but in the aggregate if you describe yourself as "religious" in America, you are very likely a Protestant Christian.

Although, I do find religious Christians and religious Muslims pretty much the same, along with religious Jews.


Note also, when people speak of "Asian" they usually do mean people from Southern Asia.

 

[Metaphor]

Yes, cultural norms are different in Western societies than they are in Latin America.

Maybe just a nitpick, but Latin American societies are generally considered Western Societies.

This should be obvious from the name... they follow Western religions, speak Western languages, etc...

Perhaps, though it's my opinion that Chile, Colombia and Uruguay are more culturally similar to each other than say Germany, Ireland, and Italy.

 

[J842P]

Yes, cultural norms are different in Western societies than they are in Latin America.

Maybe just a nitpick, but Latin American societies are generally considered Western Societies.

This should be obvious from the name... they follow Western religions, speak Western languages, etc...

Perhaps, though it's my opinion that Chile, Colombia and Uruguay are more culturally similar to each other than say Germany, Ireland, and Italy.

I'm not so sure about that. I would consider Uruguayans, Argentinians, Italians, and Spaniards to be more similar to each other than, say, Germans, Irish, and English.

I would say Latin America is like an extension of Southern Europe, culturally. This is particularly evident in South America.

Central America / Mexico and the Caribbean are definitely a different cluster, but obviously there is a lot of similarity with Spanish culture in particular.

In any case, they would all be clearly Western if Western means anything other than literally from continental Europe. Of course, Central America/ Mexico / the Caribbean have some heavily syncretic elements from their indigenous populations as well as from the descendants of African slaves in the Caribbean, but the over-arching culture is basically European.


There are notable exceptions, e.g. the Bolivians maintain a lot of Andean identity, Guatemala outside of Guatemala City is Mayan.