4 ms·

I think that the term "diverse" has many definitions. For example, if a gender split at a tech company was 50/50, but everyone was white, would we still conside

I think that the term "diverse" has many definitions. For example, if a gender split at a tech company was 50/50, but everyone was white, would we still consider that company diverse? What is the right definition of "diversity" and which groups (gender vs race vs socio-economic vs ...) should we spend our time working on? You can see how identity politics become problematic very quickly.

noobermin 10y ago
When it comes to a single tech company, just by the reason of statistics, I can understand how it's hard to worry about diversity because you're basically taking a small N sampling of the distribution. When you're talking about all tech companies as a whole it makes more sense...and similarly, all colleges, or even individual colleges, who tend to enroll a larger sample of the population. Of course, when we're talking about Google or Facebook, we at big N territory, so if you don't expect biases, the make up should reflect the general population.
jasonwatkinspdx 10y ago
You're overcomplicating things. In the ideal if we erased biases in education and staffing, we'd expect the various distributions we measure across tech companies to match a random sampling of the population at large. When we see large deviations from the population at large, it's a clear signal that something is going on. This isn't that obscure, and we certainly don't need to engage in a false, hand wringing, semantic debate derailment along the lines of "but what IS diversity REALLY anyway?" Please. You know exactly what we're talking about here. Your post is an exercise in identity politics as well. Welcome to same bowl the rest of us are in.
- spaceflunky 10y ago
  My query is more along the lines of how to invest resources to correct large deviations from the population at large and which deviations are more important than others. For example, HMC is attempting to address diversity, but only along gender lines. There is no attempt to address racial diversity in any means. Further in very progressive companies like Google and Facebook, Hispanic and Blacks in leadership positions are less than 1%, compared to a US population of 17% and 12% nationwide (respectively). In these companies, 24% of leadership roles are held by women. That means that with regards to women, they are hitting 48% of their target (if the target is to get a 50/50 gender split). Whereas, with regards to race, they are hitting 5% of their target with Hispanics and 8% with blacks (if the goal is to get close to the population sample at large). So why is that that most "diversity" program only focus on gender? I know that the two aren't mutual exclusive, but it's no secret that far more resources are spent on gender diversity than anything else.
  - jasonwatkinspdx 10y ago
    > For example, HMC is attempting to address diversity, but only along gender lines. There is no attempt to address racial diversity in any means. This particular article is about gender imbalance, but I'm sure HMC has policies and controls that address race as well. I'd be curious about the numbers. I visited there 2 decades ago and it was definitely a very white student population.
- drspacemonkey 10y ago
  >In the ideal if we erased biases in education and staffing, we'd expect the various distributions we measure across tech companies to match a random sampling of the population at large Has anyone actually proven that, or is that just an assumption?
Torgo 10y ago
Depending on the profession, "diversity" means black people or women. If you are diverse but don't have many women, you aren't actually diverse for the purposes of media hectoring. If you are diverse but you don't have enough black people, you aren't actually diverse for the purposes of media hectoring. Which sports field in the graph below is the most representative of US demographics? Which one would be recognized for achievements in diversity? https://i.imgur.com/hbmGgQD.png https://i.imgur.com/hbmGgQD.png It would be the NBA: http://www.espn.com/nba/news/story?id=5267604 http://www.espn.com/nba/news/story?id=5267604 because "diversity" means black people, not "representative of America's diversity." In tech, if your ethnic demographics match the population, you are not diverse. To be diverse you need to hire less white people. Pinterest was not diverse because women were only 27% of tech hires. Ethnically the hires are 60% Asian, white males 35%, but the conclusion was too many white males. https://www.fastcompany.com/3051659/inside-pinterests-plans-to-fix-its-diversity-problem https://www.fastcompany.com/3051659/inside-pinterests-plans-... "Diversity" is not about accurate representation.
- toomanybeersies 10y ago
  That FC article skims past, but doesn't quite touch on a very important point that's very central to the article: > Silicon Valley companies are going after the same small pool of nonwhite, nonmale candidates There's a limited pool of female software engineers and of non-white-or-asian software engineers. Universities are only graduating so many women, somewhere around the rate that Pinterest is hiring (~20%), if you hire more than this, you're just making sure that other companies can only have less than 20% women. The change needs to start at the bottom, there need to be more women doing software engineering at university. You can't just hire all of the existing pool of female software engineers, or black software engineers, and then call yourself diverse, and shame on all the companies that now have only 5% female engineers.
- FT_intern 10y ago
  Let's deep dive into diversity further. Why do we want diversity? Because apparently, we want people with different perspectives and experiences. Now with "diversity", we have to specify how we split up the population into groups. We can use gender, race, country of origin, income, etc. The media and general "diversity officers" use race and gender. However, this type of group division is not efficient for the goal of different perspectives and experiences. Female Americans have very similar experiences and perspectives compared to male Americans, as do black Americans and white Americans. Differences in country of origin provide much higher differences in perspectives and experiences. A black American would have a more similar viewpoint with a white American than a black African immigrant. Dividing the groups by country of origin is MUCH more effective at satisfying the goal. When viewed through the lense of country of origin, the tech field is very diverse in comparison to every other field out there. The field has a high immigrant %, and even the Americans are generally first generation Americans with immigrant parents, providing an extra boost in diversity. Ironically, using country of origin to divide up groups shows that tech is much more diverse than fields like journalism/HR/etc. that vehemently advocate for diversity in tech.

noobermin 10y ago

When it comes to a single tech company, just by the reason of statistics, I can understand how it's hard to worry about diversity because you're basically taking a small N sampling of the distribution. When you're talking about all tech companies as a whole it makes more sense...and similarly, all colleges, or even individual colleges, who tend to enroll a larger sample of the population. Of course, when we're talking about Google or Facebook, we at big N territory, so if you don't expect biases, the make up should reflect the general population.

jasonwatkinspdx 10y ago

You're overcomplicating things. In the ideal if we erased biases in education and staffing, we'd expect the various distributions we measure across tech companies to match a random sampling of the population at large. When we see large deviations from the population at large, it's a clear signal that something is going on. This isn't that obscure, and we certainly don't need to engage in a false, hand wringing, semantic debate derailment along the lines of "but what IS diversity REALLY anyway?" Please. You know exactly what we're talking about here. Your post is an exercise in identity politics as well. Welcome to same bowl the rest of us are in.

spaceflunky 10y ago
My query is more along the lines of how to invest resources to correct large deviations from the population at large and which deviations are more important than others. For example, HMC is attempting to address diversity, but only along gender lines. There is no attempt to address racial diversity in any means. Further in very progressive companies like Google and Facebook, Hispanic and Blacks in leadership positions are less than 1%, compared to a US population of 17% and 12% nationwide (respectively). In these companies, 24% of leadership roles are held by women. That means that with regards to women, they are hitting 48% of their target (if the target is to get a 50/50 gender split). Whereas, with regards to race, they are hitting 5% of their target with Hispanics and 8% with blacks (if the goal is to get close to the population sample at large). So why is that that most "diversity" program only focus on gender? I know that the two aren't mutual exclusive, but it's no secret that far more resources are spent on gender diversity than anything else.
- jasonwatkinspdx 10y ago
  > For example, HMC is attempting to address diversity, but only along gender lines. There is no attempt to address racial diversity in any means. This particular article is about gender imbalance, but I'm sure HMC has policies and controls that address race as well. I'd be curious about the numbers. I visited there 2 decades ago and it was definitely a very white student population.
drspacemonkey 10y ago
>In the ideal if we erased biases in education and staffing, we'd expect the various distributions we measure across tech companies to match a random sampling of the population at large Has anyone actually proven that, or is that just an assumption?

Torgo 10y ago

Depending on the profession, "diversity" means black people or women. If you are diverse but don't have many women, you aren't actually diverse for the purposes of media hectoring. If you are diverse but you don't have enough black people, you aren't actually diverse for the purposes of media hectoring. Which sports field in the graph below is the most representative of US demographics? Which one would be recognized for achievements in diversity? https://i.imgur.com/hbmGgQD.png https://i.imgur.com/hbmGgQD.png It would be the NBA: http://www.espn.com/nba/news/story?id=5267604 http://www.espn.com/nba/news/story?id=5267604 because "diversity" means black people, not "representative of America's diversity." In tech, if your ethnic demographics match the population, you are not diverse. To be diverse you need to hire less white people. Pinterest was not diverse because women were only 27% of tech hires. Ethnically the hires are 60% Asian, white males 35%, but the conclusion was too many white males. https://www.fastcompany.com/3051659/inside-pinterests-plans-to-fix-its-diversity-problem https://www.fastcompany.com/3051659/inside-pinterests-plans-... "Diversity" is not about accurate representation.

toomanybeersies 10y ago
That FC article skims past, but doesn't quite touch on a very important point that's very central to the article: > Silicon Valley companies are going after the same small pool of nonwhite, nonmale candidates There's a limited pool of female software engineers and of non-white-or-asian software engineers. Universities are only graduating so many women, somewhere around the rate that Pinterest is hiring (~20%), if you hire more than this, you're just making sure that other companies can only have less than 20% women. The change needs to start at the bottom, there need to be more women doing software engineering at university. You can't just hire all of the existing pool of female software engineers, or black software engineers, and then call yourself diverse, and shame on all the companies that now have only 5% female engineers.
FT_intern 10y ago
Let's deep dive into diversity further. Why do we want diversity? Because apparently, we want people with different perspectives and experiences. Now with "diversity", we have to specify how we split up the population into groups. We can use gender, race, country of origin, income, etc. The media and general "diversity officers" use race and gender. However, this type of group division is not efficient for the goal of different perspectives and experiences. Female Americans have very similar experiences and perspectives compared to male Americans, as do black Americans and white Americans. Differences in country of origin provide much higher differences in perspectives and experiences. A black American would have a more similar viewpoint with a white American than a black African immigrant. Dividing the groups by country of origin is MUCH more effective at satisfying the goal. When viewed through the lense of country of origin, the tech field is very diverse in comparison to every other field out there. The field has a high immigrant %, and even the Americans are generally first generation Americans with immigrant parents, providing an extra boost in diversity. Ironically, using country of origin to divide up groups shows that tech is much more diverse than fields like journalism/HR/etc. that vehemently advocate for diversity in tech.