Including statistical controls for occupation is effectively assuming that women do not have agency in the choices that they make; that they do not choose different careers in response to discrimination.
-
-
Surely that depends on the question you're trying to answer? If you're interested in whether an earnings gap is driven by wage discrimination or something else, controlling for occupation is essential.
2 replies 0 retweets 1 like -
Male-female comparisons within occupation are not apples-to-apples comparisons.
1 reply 0 retweets 1 like -
Replying to @erinhengel @joshualord2700 and
Conditional on occupation, women may be paid more despite experiencing direct discrimination (getting paid less for doing the same job) simply because they’re that much better than their male peers.
1 reply 0 retweets 0 likes -
Replying to @erinhengel @joshualord2700 and
One reason women may be better than their male peers conditional on occupation could be due to a selection effect caused by discrimination. E.g., only exceptional women choose to pursue particularly lucrative fields with few women and in which discrimination is more prevalent.
2 replies 0 retweets 0 likes -
Certainly plausible, but very speculative. Do you know of any evidence that this is happening?
2 replies 0 retweets 0 likes -
Yeah, my own work! (Check out my website.) In general, however, this is indeed an under-researched area in economics.
1 reply 0 retweets 2 likes -
OK, thanks, I'll take a look. Also very much agree that more research is needed, because at present there's ample ambiguity for different sides to cherry-pick evidence.
3 replies 0 retweets 1 like -
Yes, I think if statistics are worthless, the best thing we can do is keep pushing to make sure experiments into discrimination are rigorous and ones into gender differences are allowed to be talked about (and also rigorous, obviously)
2 replies 0 retweets 0 likes -
Statistical evidence frequently cited to minimise the likelihood of direct discrimination is weaker than it ostensibly seems to the statistically less-informed. Basically, we can all agree that more research is needed, and that no candidate hypotheses should be off-limits.
1 reply 0 retweets 0 likes
Yes. If stats are no good here, then they are not. But then the conversation about stats are over. Outside of it, choices made are still relevant.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.