What isn’t clear is whether these analyses are well controlled. E.g. is it over indexing on companies flourishing in wealthy coastal cities. I doubt it’s a causal effect personally. But I don’t have numbers either way.
Promoting inclusivity for sexual orientation and gender is different from trying to improve diversity. I think it’s pretty rare to see the latter relative to racial diversity efforts.
These diversity "studies", as the grandparent points out, are light on data and heavy on marketing. Several years back when I looked up the McKinsey 2012 (or 2013, whichever) study that started all this, it was a marketing flyer with zero data.
It's really convenient you can make such a hefty claim and provide essentially no evidence to back it up.
The McKinsey stuff and other stuff is generally publically available data. It is real. I think you can find it for some resources at least. The question is whether it’s compelling.
As requested, the quote:
> Companies tend not to have data on the sexual orientation of their employees. Such studies are not possible to do.
The studies are absolutely possible to do. Throwing up our hands because companies "tend not to have data on sexual orientation" reads like a cop-out when the comment you responded to was attempting to get at some data around this very issue.
> The McKinsey stuff and other stuff is generally publically available data.
Is it? I found a report [1], which isn't the one which I previously remember seeing but it'll have to do since that older one evades me now.
Look at the footnotes. Several pages say their source is the "McKinsey Diversity Database". I searched for it on Google and DuckDuckGo and it didn't show up for me. Do you have a link to the underlying data for this database? I tried and failed to find one.
Lastly, review the authors at the end of the doc. If you put the authors' bios at the top of the report, the conclusion would be surprising to exactly zero people. That doesn't mean their conclusion is invalid, but given that much of their data is tied up in a "diversity database" that I can't find is VERY suspect.
[1]: https://www.mckinsey.com/~/media/mckinsey/business%20functio...
> data isn't important as long as I agree with the conclusion
Your quote does not contain a claim that pretends data is not important so long as I agree with the conclusion. It in fact contains a claim that the data is not possible to acquire. And I think it’s kind of dumb of you to pretend otherwise. You cannot practically do a good study or even surface stats on company performance vs sexual orientation diversity when companies do not collect sexual orientation data. That is not a controversial statement. Observe how I am not making a claim that pretends what that data would say in either direction.
I’m not really interested in trying to find this thing. But I don’t feel like you’re reading what I’m saying. It is very likely a real dataset. It’s just not very deep. It’s public info like stock prices, leadership profiles, and published diversity stats. You can find similar datasets if not this one precisely. And if you would again look at what I said originally, it should be expected that there is a relationship. Why? Because the fastest growing companies of the recent past are generally growth stocks based on the coast in naturally more diverse areas. As I already stated it’s probably not causal.
I think it’s unnecessary to suggest anything here is “VERY suspect”. That’s unlikely. It’s just not very interesting. Like, they’re showing p values of 0.1 on some pages. Shrug. You’re reframing this discussion, which could have been mild agreement off the top of insufficient evidence to identify a causal effect, to an annoying about whether the exercise is a lying conspiracy or not. I’m just not interested in engaging with that.