> Why wouldn't we want that?

Because it's possible to have multiple goals at the same time. Reality is a co-optimization problem.

The points that you have made above are why there are still many more female nurses -- it's not like they will accept every single man, there's simply a lower threshold than there is for women.

I think a lot of the arguments about whether AA makes sense at all should really be reframed as discussions about how much AA makes sense. How do goals like diversity, inclusion, and minimizing inequality weigh against other objectives like rewarding merit and maximizing productivity outcomes? Surely the balance-point is not zero, but it's conceivably much closer to zero than today's policies would imply.

> Because it's possible to have multiple goals at the same time. Reality is a co-optimization problem.

And one thing to consider is that people don't choose colleges just based on academics, they also choose them based on social environments, and as such universities have an incentive to try to create amicable social environments to attract the best students.

Case in point: I decided not to go to a good engineering school because it was more than 80% male and in a small town. The social environment was desperately bro-ish. I backpedaled and went to a liberal arts college primarily because of vastly superior social environment.

> I decided not to go to a good engineering school because it was more than 80% male and in a small town. The social environment was desperately bro-ish. I backpedaled and went to a liberal arts college primarily because of vastly superior social environment.

I dont think many people have such a privilege

I assume you're referring to the fact that liberal arts colleges are usually private and more expensive, which is true, but I paid for it with loans, so to the extent that one would be advised to go massively in debt for college, it's possible.

My parents had only saved enough for me to go to a state school, which got me through one semester at a private college, and the rest were student loans, and working often multiple part-time jobs (and full-time in the summer).

I suppose one argument is that men in general are stronger, and it's probably a good idea to have strong nurses around, especially with the current obesity rates. I'm absolutely making things up though, not sure if any studies on that kind of thing have been done recently.

True but if female nurses are having issues lifting patients that would then be an opportunity for private enterprise to invent machines/tools that assist with this

>an opportunity for private enterprise to invent machines/tools that assist with this

I think those machines/tools are called "orderlies."[0]

[0] https://www.indeed.com/career-advice/finding-a-job/what-is-a...

Yes male nurses already get drafted into lifting the heavy patients, and dealing with the ornery patients, just because they're male, so it's no surprise many men nope out of nursing.

Any metric that is not the true metric that is optimized on will lead to worse results if overfit. You need to make sure there's at least a reasonable amount of variance.

> If we assume that:

> Whatever merit based metrics the school uses to determine admission are predictive of eventual performance

This seems a questionable assumption.

The best individual components don't necessarily result in the best system.

There may well be benefits to diversity that are not easily quantified at the individual level.

The classic example of the development team that created a computer vision system that didn't work well at all with certain skin tones but they were unaware of this until feeling it live at a conference comes to mind. They didn't have anyone on their dev teams or QA teams with that skin tone.

There are numerous studies showing that diverse teams come up with better solutions. Diversity pays dividends at the group level, and assessments that only consider things at the individual level may miss that

> There are numerous studies showing that diverse teams come up with better solutions.

There are also numerous studies that show diversity has either no effect or even negative effects on various measures of group performance - the academic literature on this issue is a total mixed bag. Here's how the OECD's [1] summary of the literature on diversity and firm performance:

> "Contrary to the often assumed, direct positive impact of diversity on business performance, research shows that at the firm level, the business case for diversity is not particularly strong. However, while the impact of diversity might be small, there is a strong economic argument against discrimination and non-inclusion based on the sizeable cost associated with it. Finally, the chapter notes ethical reasons for fostering a just and equitable labour market alongside the economic argument for diversity."

The literature also suggests the effects of diversity on group performance, whether positive, neutral or negative, are situation specific, depend on your measure of 'performance', and your definition of 'diversity'. Again, from the OECD:

> "A number of studies make a difference between ‘highly job-related diversity’, such as educational background, job position or function in the company, and diversity aspects that are ‘less job-related’, e.g. gender, ethnicity or age. Measurement of team performance includes multiple indicators, such as efficiency, creativity, innovation and productivity. Findings are somewhat mixed, but impacts of gender, ethnicity or age diversity are found to be either very small or insignificant. Some meta-analyses find no significant impact of gender composition, ethnicity or age on team performance (Horwitz and Horwitz, 2007[19]; Schneid et al., 2016[20]), while others find negative, but very small impacts (Bell et al., 2011[21]; Joshi and Roh, 2009[22])3. Most studies do, however, find a positive relationship between team performance and having teams with different professional backgrounds and other task-related characteristics (Bell et al., 2011[21]; Horwitz and Horwitz, 2007[23]; Joshi and Roh, 2009[22]). Overall, these findings from meta-analyses seem to suggest that team diversity in terms of gender, ethnicity or age do not matter much for team performance. However, there is some evidence that demonstrates the importance of situational settings by examining under what specific conditions diversity dynamics may unfold and how."

[1] https://www.oecd-ilibrary.org/sites/5f8c1531-en/index.html?i...

Some patients might prefer a doctor/nurse of the same gender.