If you are using two role models, why would you not use one male and one female?
What do we learn besides "advertising (sort of) works" (p>5%) from this paper?
This wouldv'e been a much better paper lol
The Gender Problem in Economics is SOLVED!
-
"As shown in Table 2, in 2015, the proportion of women who enrolled
in Intermediate Microeconomics within a year of Principles is around 13 percent and is the same across treatment and control groups."
Two things are weird about this. One, why do the authors paint over this? One class has 11% and the other 15%, they are not "the same".
Two, and this is the confusing part, why would you have two pre-treatment cohorts in 2015 that are separate from the post-treatment and control cohorts in 2016?
Think of it this way: there are two professors, Peter and Paul. The treatment happens in 2016, in Paul's class.
So the 2015 class difference is going to help boost the post treatment difference in a DiD (this "control group" is just that you were in class A in professor John's class versus class B in professor Paul's class, in the prior cohort before the treatment). A potentially spurious difference for 2015 means that even if the 2016 classes had approximately the same enrolment rate between Peter and Paul, you may find a significant difference in the DID. Not obvious this is much better than not doing a DiD.
Also note that the balance test fails, proving how tenuous this result is. You could have also written the paper to show even stronger results that
a) the speech treatment causes students in the treatment class to be born internationally rather than in America, twenty years earlier (p value 0.00 rather than 0.05)
b) the speech treatment causes students to go back one year in the and enrol in principles as freshman rather than juniors (p value 0.00 rather than 0.05)
Or maybe international students are more likely to major in econ, and students who choose to do econ as freshmen rather than juniors are more likely to major in econ.
But who knows.No good. Gender bias unless the teachers are Petra and Paulie (who is gender fluid)
-
If you are using two role models, why would you not use one male and one female?
What do we learn besides "advertising (sort of) works" (p>5%) from this paper?It has a significant impact on women but not on men, hence the paper
But why does it have a significant impact on women? You are obviously assuming that it is because the role models are female. Maybe their priors on financial or professional success conditional on being an econ major were lower, and women value that more than men? There is no clear identification if you do not assume the result to begin with.
-
If you are using two role models, why would you not use one male and one female?
What do we learn besides "advertising (sort of) works" (p>5%) from this paper?It has a significant impact on women but not on men, hence the paper
But why does it have a significant impact on women? You are obviously assuming that it is because the role models are female. Maybe their priors on financial or professional success conditional on being an econ major were lower, and women value that more than men? There is no clear identification if you do not assume the result to begin with.
Precisely. If the role model was an attractive male would that have had the same effect?
-
It has a significant impact on women but not on men, hence the paper
What they don't state in the paper is that while the estimated effect is +8 pp for females, it is -6 pp for males.
The standard errors of both estimates are roughly the same size. Yet they claim a large effect for females and no effect for males? Who refereed this?
-
It has a significant impact on women but not on men, hence the paper
What they don't state in the paper is that while the estimated effect is +8 pp for females, it is -6 pp for males.
The standard errors of both estimates are roughly the same size. Yet they claim a large effect for females and no effect for males? Who refereed this?
it's a joke
-
This effect size reminds me of LaCour and Green, who demonstrated that simply a 20-minute conversation with a gay canvasser produced a large and sustained shift in attitudes toward same-sex marriage for Los Angeles County residents.
That basic result turned out to be true. See Broockman and Kalla.
This result doesn't actually strike me as very surprising. It's also not a serious contribution to the study of economics, but I don't doubt the result.A 10 minute conversation is apparently sufficient to reduce prejudice:
https://science.sciencemag.org/content/352/6282/220I find it EXTREMELY unlikely that it is this easy to change fundamental beliefs, which prejudice is supposed to be based on. If you can change something that easily you measure something less fundamental and therefore less interesting. How can there be any prejudice if it is this easy, given that so many institutions in society promote anti-prejudice values? These results make no sense to me.
-
This effect size reminds me of LaCour and Green, who demonstrated that simply a 20-minute conversation with a gay canvasser produced a large and sustained shift in attitudes toward same-sex marriage for Los Angeles County residents.
That basic result turned out to be true. See Broockman and Kalla.
This result doesn't actually strike me as very surprising. It's also not a serious contribution to the study of economics, but I don't doubt the result.A 10 minute conversation is apparently sufficient to reduce prejudice:
https://science.sciencemag.org/content/352/6282/220
I find it EXTREMELY unlikely that it is this easy to change fundamental beliefs, which prejudice is supposed to be based on. If you can change something that easily you measure something less fundamental and therefore less interesting. How can there be any prejudice if it is this easy, given that so many institutions in society promote anti-prejudice values? These results make no sense to me.Prejudice is actually based on preferences.
If preferences can be changed in 10 minutes, wtf are we all doing.
-
I guess it will be like male and female tennis. All departments will have real economists and clowns like that, different leagues but same titles. And in the seminar we pretend to be at the same level. Not sure how long such a pooling equilibrium will last though.
-
I guess it will be like male and female tennis. All departments will have real economists and clowns like that, different leagues but same titles. And in the seminar we pretend to be at the same level. Not sure how long such a pooling equilibrium will last though.
that's actually separating and to some extent it's already there
-
1. Perform several, severely underpowered field experiments on fad topics
2. Search for a variable that turns out barely significant (due to pure chance)
3. Spin a (politically correct) story around the result
4. ProfitI don't think that ex-post choice of the treatment variable was likely to be a problem. While they did not register their RCT until after the experiment was completed, the acknowledgements show that the project was funded by the Undergraduate Women in Economics Challenge, which suggests that the hypothesis was chosen before the project was done.
But low power is a big problem either way. Given how large the SEs are, they never could have published a null result in a good journal. So the only way we hear about the paper is if the estimated effect is comically large.