People are known to have bias in favor of people that resemble themselves, so even if you tell them to just 'hire the best' they won't necessarily control for their biases and more so if that data isn't being collected and reviewed.
So if one wants to actually hire the best, they should measure these things, otherwise you can't correct for things like 'our Cal alum hiring manager keeps passing up a bunch of good Stanford grads' or more serious versions of that.
Also, why NOT measure "diversity"? You don't have to act on it, but it sure would be nice to have the metrics for any later analysis. It's not hard and isn't something that takes many resources to compile.
It's not arbitrary. The gender and ethnic makeup of the tech industry are significantly divergent from the general population, and there is no job-performance reason for it.
It's not like it's the NBA, where you have to be a tall strong man to make the team. Anyone can program, market, manage, etc.
Implicit in your statement is a belief on your part that women, black people, etc. are inherently less qualified than white and Asian men. Would you like to back that implication up with evidence?
Would you further like to consider that historical evidence indicates that when an industry claims it has meritocratic processes which naturally produced lopsided hiring with respect to gender and race, that claim has been debunked, often spectacularly, by simple tactics such as blind interviews in which the interviewer cannot discover the gender/race of the candidate?
Perhaps after considering this, you should find some other use for that reflex arc which produces "hire the best" statements whenever you're exposed to the word "diversity".
It's sounds like it's implicit in your belief that women and minorities are inferior and need to be selected for.
Any situation where you have two candidates and hire one because of irrelevant characteristics you won't be hiring the best.
You want sexist and racist hiring policies, but only when it suits the genders and race's you want. Do you here outrage from white people about being under represented compared to Asians?
I want meritocratic hiring processes. I just happen to suggest -- and I believe evidence is strongly on my side -- that a process which produces lopsided gender/race ratios is extremely unlikely to be meritocratic. I also do not believe many of the other sound bytes people inevitably throw out in this argument (example: "that group just isn't interested in our field" / "they just prefer other things", another classic which will, I'm sure, rear its head in this thread sooner or later).
> a process which produces lopsided gender/race ratios is extremely unlikely to be meritocratic.
What is the strong evidence for this statement that you claim exists?
It is very unintuitive to me, as I have always thought the problem existed earlier in the pipeline than the actual interview/hiring process. I am always open to my view being changed, so I would like to see this evidence that is strongly on your side.
Here's an example: look up the historical gender breakdown of computer programmers. Hint: once upon a time there wasn't a "pipeline problem" to use as an excuse, and for-profit free-market competitors were doing just fine with a very different gender ratio than they have now.
It is unlikely that, in the (short -- less than a generation to transform the industry) period in question, women magically became unqualified to work as programmers. Especially since many of them already had worked as programmers and nobody had found reason to complain about their work.
Or, more bluntly: the sudden disappearance from a field of an entire category of people, whose only distinguishing characteristic is being of a gender historically denied economic opportunities, is unlikely to be explained by a similarly sudden catastrophic decline in their qualifications. The continuing absence of people of that category is equally unlikely to be explained by their lack of qualifications.
Here's another example: take a look at what happens to other fields when they introduce effective blind interviewing (in which interviewers cannot discover the gender/race of the candidate prior to making a hire/no-hire decision). It once again does not suggest that previous processes were anything approaching meritocratic.
One notable example I've heard of, is symphony orchestras (and if Malcolm Gladwell covers it, it is surely a mainstream concept and dumbed down quite a lot, accessible to a non-tech audience!)
Specifically, it's the Julie Landsman story, where a fantastic French Horn performance of great power and force turned out to be produced by a small woman, and the people hiring lost their shit when they found out.
A funny side note is that, to be a CEO, typically you have to be a TALL white man. Now, in tech, I could see there being a bias for large HEADED people of whatever gender or color, on the (unjustified) assumption that raw brain material volume would help. But surely, the distance from the ground said head is, can't have much to do with it! And yet, it's another 'meritocracy' fail.
AI will be a lot better at this sort of thing (while perhaps inventing its own meritocracy fails). But we're not there yet.
Woah, where did I ever claim women are unqualified to work as programmers? You're putting words in my mouth. There are many other possible explanations.
As for your example, it is not "strong evidence". Historical breakdown doesn't mean much since very few people were computer programmers at all back then. You can't extrapolate from that to the distribution of people who would become programmers decades later once it would be a super mainstream profession that a huge fraction of the population would enter.
I'm not sure I disagree with your thesis, but a piece missing from your presentation is the tremendous growth of the industry. We don't have fewer female programmers, we have more than before. Just not nearly so many more as men.
> that a process which produces lopsided gender/race ratios is extremely unlikely to be meritocratic
Then you have the wrong process in mind. AFAIK the breakdown of employees is very similar to the breakdown of graduates (and hence applicants). So there's no bias in hiring, all the bias is coming before (socio-economic status? personal choices?).
> Implicit in your statement is a belief on your part that women, black people, etc. are inherently less qualified than white and Asian men. Would you like to back that implication up with evidence?
Many people, whenever it is suggested that something is wrong with hiring processes if they produce lopsided gender/race ratios, have a reflex arc which causes them to say something about how changing processes would be bad because you have to be sure to "hire the best". Or, about as commonly, that any change would "lower the bar".
There is no charitable interpretation of this reflex.
The only possible way to interpret it is as a naked statement that groups currently disproportionately underrepresented in the industry, or at specific companies, are made up of inherently inferior people. Otherwise, a process which "hires the best" would hire more of them, and hiring more of them would not "lower the bar".
With all due respect, that's not the only possible way to interpret it.
The intuition being expressed is as follows: Say you work at a factory, filling boxes of widgets. Widgets of varying quality stream past you, coming out of the widget stamper, such that you only have time to grab, on average, one widget out of every ten. Your job is to grab the best widgets you can spot, and put them into your bin. that's how the factory sorts widgets into different grades of quality. Further down the line, other failed startup founders toil away at other bins, snagging the ones you miss, from best to worst.
Most of the widgets are sort of a greyish mucky colour, but about one out of twelve will display a splash of original colour from the plastic that was recycled into widgets. You think nothing of it, but one day your pointy-haired manager comes and tells you that the "colourful" widgets are extra-popular and people are trading them. The company can sell boxes with at least 1/3 coloured widgets for ten times the price of regular widget boxes. Your job is now to ensure that at least one out of every three widgets that you grab is "colourful."
As a consequence, the average quality of widget in the boxes that you pack declines.
There are a lot of wrong and simplistic assumptions behind this understanding of "changing hiring processes." The knee-jerk assumption that doing that means implementing a naive quota system is not the least of these, but there are many others.
Furthermore, it is much more likely to be put forth by not-very-tactful people who have nothing to gain (and perhaps something to lose) extraneously, from greater workplace diversity; these are the people who are the most likely to both be sympathetic to, and express, views that seem at best borderline-racist.
This does not, however, mean that there is no charitable interpretation to the 'reflex' you are describing. It is not a cunning excuse to perpetuate *-ism in the workplace. It is what they believe to be a serious and valid objection to "changing hiring processes," as they understand it.
So to try to provide an interpretation in which the previously-excluded group is not perceived as inferior, you... use an analogy in which the previously-excluded group is inferior (because the only way the "colorful" widgets wouldn't be making it in earlier is if they weren't "the best widgets", since the previous process was purely based on widget quality).
I do not think you are going to convince me that way.
It comes down to proportions. Let's say 10% of widgets are awesome and the remaining 90% are not, regardless of colour. Per parent's post, let's assume 1/12 of the widgets are colourful. So now (rounded roughly):
If you're choosing freely, you'll (hopefully) end up with the awesome top 10% (1/12 of them being colourful). However, if you're forced to make 1/3 of your widgets colourful (and there's only 1/12 of them overall), you end up with:
The previously "excluded" (or rather, not promoted) group is exactly just as good as the rest, it's just not numerous enough -- if you want 1/3 of colourful widgets naturally (instead of the expected 1/12), their averages have to compete against the rest's best, and that's where the perceived inferiority comes from.
I don't have the motives that you seem to be attributing to me, so perhaps you should re-read my comment a bit more carefully if you want to guess at what I might or might not try to convince you of.
I think that it's important to note, however, that the analogy functions without relying on any notion that the "colourful" widgets are inherently inferior. It was this point of your own that I was responding to.
Notion of diversity being a better metric than actual performance is makes sense only if you see companies as a vehicle for hiring, not a tool of delivering value.
So if one wants to actually hire the best, they should measure these things, otherwise you can't correct for things like 'our Cal alum hiring manager keeps passing up a bunch of good Stanford grads' or more serious versions of that.