I think we're at the point where we respect people enough to just say, "Yup, I don't care what sort of junk my boss has, or what color it is, or what other kinds of junk s/he likes to bump it with."
The intent of the article is to say we should have some sort of mandated demographics among leadership or in the board room?
Because trying to pick a c-team to match up with census data seems like it would mean not picking a c-team based on merit.
The color and kind and intent of junk should probably just not be something that we give a fuck about.
I haven't seen one yet, save maybe doing so publicly in order to properly virtue-signal and keep the torches-and-pitchfork brigade off your back.
It gets very wearisome being berated all the time about diversity , when you just want to go build things. The compiler doesn't care if you're black, white, brown, or purple, and, honestly, I'm very suspicious of anyone who is more interested in the color of ones skin or, as you put it, the configuration of their junk, over getting the job done, done right, and done quickly.
Yes. Studies show diverse teams perform better... (and incidentally they also show that people who believe themselves to be unbiased are actually more likely to make biased decisions).
Going to call BS on this. "Studies show..." -- without a citation seems like an opinion.
Lab experiments using mock juries is all I'm finding. Doesn't quite seem relevant to technology teams.
Most teams are distribute these days... so literally, other than Google Hangouts and conference calls, I don't get any daily indicators about my team's composition.
I don't think it would matter if we did, but I don't see any studies that pertain to tech about diversity improving productivity.
Ellison/Mullin (2014) used 8 years of data from professional services firms in 60 locations. Is that close enough to tech?
You assert that distributed teams change the relationship between diversity and productivity, but I assume you're doing that without evidence, right? What's that opinion based on?
We're hiring devs, designers, PM's, digital producers etc. working with recruitment companies, advertising in various places - meet ups, job boards etc etc.
Recieved hundreds of CV's, run dozens of interviews.
I can count on one hand the number of women that sent their CV through, and we interviewed two, hiring one.
This isn't a tech thing - this goes back to children growing up in an environment where gender should make no difference to their desire or interest in STEM subjects, which should be encouraged regardless.
Now the conundrum:
Given two candidates (m/f) of equal footing (talent, skill, comms etc) - would you offer work to the female or male?
Given a stronger male candidate over a female candidate - should I offer the job to the female instead just because she's a she?
so essentially you want to be known as company which discriminate white people and men in particular and prefer gender or race over several skills, good luck with such reputation, be sure I will always choose company not known for having these practices when shopping around
Of course we don't - and it's rather disingenuous to say we do.
We want to promote diversity where we can, and if you walked into our office today you'd be surprised at the balance we've got.
My statement was highlighting what the current state of tech is, and to make a more gender diverse working space i would have to sacrifice quality / skills / talent, which I will not do.
But if I am not willing to sacrifice, how am I supposed to have a more diverse space?
Because the first few generations of people will always lack the skill or confidence where they feel like a minority, ideally you want to create an environment where they all can flourish together; but not everyone se's things this way - and businesses more than anything expect immediate returns on their investments.
As OP has stated, and I concur - diversity always encourages one to be more empathic, and generally leads to better teams.
Side note - your definition / standards of beauty or customer service will be wildly different to mine, which will be wildly different to someone on the other end of the planet. You can have diverse teams, with varying strengths and weaknesses in individuals but still be a net positive.
> to make a more gender diverse working space i would have to sacrifice quality / skills / talent, which I will not do.
The main point here I've already addressed in my reply to your parent comment, but there's an additional point worth making.
There are now two (to my knowledge) studies that show diverse teams outperform heterogenious teams. The diverse teams feel less comfortable because your team members might fail the "would I enjoy working with you on a sunday" test, or the "do I want to go for a beer with you" test, but the additional perspectives provide meaningful insight
Edit: The point being that assessing team members on the basis of individual skill misses an important point about the way teams behave as a whole
You say you want to promote diversity, but do you? Do you have smoke and non-smoke offices, meat and non-meat, nude and non-nude? Do you promote equally people who drive to work, walk, and those that take a bike? Is there promotion in diversity of music taste, where each different taste is given proportional place in the office? Do you have proportional use of programming languages, promoting the diversity in preference?
Have you ever used team building in order to make the group more homogeneous and working as a singular mind? Such effort goes directly against promoting diversity. Group cohesion is general opposite to diversity in behavior, goals, values, culture and opinions. The less disagreement and infighting there is, the less diversity you got.
Example off the top of my head; a study into mixed-race juries (Sommers 2006) showed that not only did the predictive power of the group increase in the presence of black jurors, but the prodictive power of the white jurors increased without the black juror needing to actually intervene. Their presence caused a behaviour change.
If that’s the case, we should not aim for equality, but instead aim for at least a token minority member presence in all groups, since this produces the intended effect.
While quotas appear to sometimes be necessary to break some environments open, they're polarising & produce resentment. I wouldn't expect that to be a necessary part of a high performance team if you can address the bias that keeps under-represented groups out in the first place.
The evidence that hiring processes are both biased and poorly predictive of productivity is pretty strong. I would start by fixing that (and am doing).
Do those studies explain why it relates to productivity?
Correlation studies are useful but limited. If the increase in productivity is correlated to employees general acceptance of different views, then yes, diversity of music taste should show a small positive effect but diversity in politics and religion should have a much bigger effect, and I would guess that nudism would have a major effect given peoples strong views on nudity. This could be rather easy tested as causation.
If correlation is with genetics, studies on people born of different gender but raise/living as an other should provide proof of that. Same goes to people born of one ethnic group but raise under an other. Fairly easy to test, but I doubt productivity has anything to do with genetics, as genetics generally don't have much causation with cultural phenomena.
As with other correlation studies, such studies tend to raise question around the studies themselves. What work are the test subjects doing, and is the result same for all kind? As an example, work demanding multiple spoken languages would benefit of having different people that talk fluently different languages, and people from different places and different cultures often speak different languages fluently. Such aspects need to be considered and removed from studies in order to not create unintended correlations.
> Do those studies explain why it relates to productivity? Correlation studies...
Sommers showed causation. The presence of black subjects in the heterogenous group caused the white subjects in that group to behave more conscientiously/objectively, increasing the performance of that group compared to the control.
Others like Ellison/Mullin discuss performance in relation to social capital. Homogenous groups tend to find it easier to build more social capital, so feel more comfortable, but more easily succumb to groupthink.
> I can count on one hand the number of women that sent their CV through, and we interviewed two, hiring one.
Studies show that:
(a) You can create a biased process inadvertently before the CV stage based on the language you use and approach to sourcing.
(b) CVs themselves are also a biasing influence.
(c) Interviews are extremely biasing unless done in a very structured way, and even then they're not great.
So it's very likely that you're (without meaning to) introducing additional bias on top of the constraints on the talent pool. This isn't finger-pointing, it's very hard to run a bias-free process. Even the Googles of the world fail at this.
> ...an environment where gender should make no difference to their desire or interest in STEM subjects
Are you saying that the environment is neutral, or that it should be neutral? First would be wrong, second would be correct.
> Given two candidates...
Your hypotheticals are premised on the idea that your hiring process and judgement are objective, and that you need to positively discriminate in order to redress the balance... that's simply not a view supported by the literature on this.
Should be neutral of course; and I concur with all of your points. We've gone back and had our female staff give us plenty of input in our outreach / hiring / on-boarding practices, but it's certainly something we need to keep working at; I think a bit of herd mentality also kicks in when you're already on the inside and can no longer see things from an outsiders perspective.
The population of women in these fields is very small, so the number of exceptional ones is small as well.
Forcing a match to nationwide demographics is inappropriate. That's not the population that should be considered.
Hiring people who are poor coders and collaborators on account of their gender, race, religion, or sexual orientation is equally inappropriate whether it's a white male or minority female.
My female partner in tech has encountered this so regularly she is now suspicious of female hires. Behaviors & performance that would get a male terminated are tolerated and instead show up on a CV as experience at multinational companies. I'm talking basics: unwilling & unable to use source control. This drives resume bias as richmarr highlights.
This is the result of inappropriate affirmative action. It harms the very people you are seeking to help.
"Less than 5% of investors surveyed by LinkedIn rated diversity as their 'top concern' as compared to macroeconomic conditions, raising capital, hitting revenue targets, or hiring the right people. More than half of investors ranked founder commitment to a diverse team as the least of their concerns when considering to invest."
I'd be surprised if even an investment firm made entirely of "women and minorities in tech" had drastically different priorities than this 5%.
Diversity for diversity's sake is not a real solution.
Studies also show that women are paid less than men for similar work.
Where are all the tech companies created by women, with mostly women as employees capitalizing on the market inefficiency giving them competitive advantage?
> Studies also show that women are paid less than men for similar work.
Yes they do, but don't get hung up on women. Studies also show a beneficial effect from racial diversity.
As for your other question, is your point that studies are sometimes wrong, or that women are worth less than men, or that there's a gap in the market?
Companies taking advantage by mothers who want to the workforce once their kids are in school by offering flexible working.
Companies dropping their long-standing policies of hiring from specific schools or attainment levels. These policies are historically quite common in professional services firms, but heavily reduce the available talent pool (and thus the quality of hires, because schools and attainment levels are not strongly predictive of productivity).
When we bring parenthood into the picture I think this is confounding the results that the studies should account for, but in too many cases, they do not.
Then the mainstream media and people championing social justice causes often take these results the wrong way (even if they do so for noble reasons) and run with it too far.
I think we should compare never married men to never married women, otherwise I believe it to be disingenuous to make that wage gap argument based on gender.
Hoover Institution Senior Fellow Thomas Sowell argues that economic differences between working men and women are not generally due to employer discrimination, as is widely alleged: https://www.youtube.com/watch?v=8EK6Y1X_xa4
>... confounding the results that the studies should account for...
You're absolutely right that studies should make sure to control for potentially important variables. I don't know which studies you referring to. Not a topic I've looked at.
That said, criticising a second group of studies can't reasonably cast shade over the first group, if that was your implication.
31 comments
[ 5.1 ms ] story [ 89.1 ms ] threadI think we're at the point where we respect people enough to just say, "Yup, I don't care what sort of junk my boss has, or what color it is, or what other kinds of junk s/he likes to bump it with."
The intent of the article is to say we should have some sort of mandated demographics among leadership or in the board room?
Because trying to pick a c-team to match up with census data seems like it would mean not picking a c-team based on merit.
The color and kind and intent of junk should probably just not be something that we give a fuck about.
I haven't seen one yet, save maybe doing so publicly in order to properly virtue-signal and keep the torches-and-pitchfork brigade off your back.
It gets very wearisome being berated all the time about diversity , when you just want to go build things. The compiler doesn't care if you're black, white, brown, or purple, and, honestly, I'm very suspicious of anyone who is more interested in the color of ones skin or, as you put it, the configuration of their junk, over getting the job done, done right, and done quickly.
Yes. Studies show diverse teams perform better... (and incidentally they also show that people who believe themselves to be unbiased are actually more likely to make biased decisions).
Lab experiments using mock juries is all I'm finding. Doesn't quite seem relevant to technology teams.
Most teams are distribute these days... so literally, other than Google Hangouts and conference calls, I don't get any daily indicators about my team's composition.
I don't think it would matter if we did, but I don't see any studies that pertain to tech about diversity improving productivity.
You assert that distributed teams change the relationship between diversity and productivity, but I assume you're doing that without evidence, right? What's that opinion based on?
Recieved hundreds of CV's, run dozens of interviews.
I can count on one hand the number of women that sent their CV through, and we interviewed two, hiring one.
This isn't a tech thing - this goes back to children growing up in an environment where gender should make no difference to their desire or interest in STEM subjects, which should be encouraged regardless.
Now the conundrum:
Given two candidates (m/f) of equal footing (talent, skill, comms etc) - would you offer work to the female or male?
Given a stronger male candidate over a female candidate - should I offer the job to the female instead just because she's a she?
We want to promote diversity where we can, and if you walked into our office today you'd be surprised at the balance we've got.
My statement was highlighting what the current state of tech is, and to make a more gender diverse working space i would have to sacrifice quality / skills / talent, which I will not do.
But if I am not willing to sacrifice, how am I supposed to have a more diverse space?
that's like expecting that photo models should be ugly women as well or sellers should be rude people to have diversity just for the sake of diversity
As OP has stated, and I concur - diversity always encourages one to be more empathic, and generally leads to better teams.
Side note - your definition / standards of beauty or customer service will be wildly different to mine, which will be wildly different to someone on the other end of the planet. You can have diverse teams, with varying strengths and weaknesses in individuals but still be a net positive.
The main point here I've already addressed in my reply to your parent comment, but there's an additional point worth making.
There are now two (to my knowledge) studies that show diverse teams outperform heterogenious teams. The diverse teams feel less comfortable because your team members might fail the "would I enjoy working with you on a sunday" test, or the "do I want to go for a beer with you" test, but the additional perspectives provide meaningful insight
Edit: The point being that assessing team members on the basis of individual skill misses an important point about the way teams behave as a whole
Have you ever used team building in order to make the group more homogeneous and working as a singular mind? Such effort goes directly against promoting diversity. Group cohesion is general opposite to diversity in behavior, goals, values, culture and opinions. The less disagreement and infighting there is, the less diversity you got.
Studies show gender & racially diverse teams perform better.
Not aware of any studies into whether diversity of music taste relates to productivity.
Example off the top of my head; a study into mixed-race juries (Sommers 2006) showed that not only did the predictive power of the group increase in the presence of black jurors, but the prodictive power of the white jurors increased without the black juror needing to actually intervene. Their presence caused a behaviour change.
While quotas appear to sometimes be necessary to break some environments open, they're polarising & produce resentment. I wouldn't expect that to be a necessary part of a high performance team if you can address the bias that keeps under-represented groups out in the first place.
The evidence that hiring processes are both biased and poorly predictive of productivity is pretty strong. I would start by fixing that (and am doing).
Correlation studies are useful but limited. If the increase in productivity is correlated to employees general acceptance of different views, then yes, diversity of music taste should show a small positive effect but diversity in politics and religion should have a much bigger effect, and I would guess that nudism would have a major effect given peoples strong views on nudity. This could be rather easy tested as causation.
If correlation is with genetics, studies on people born of different gender but raise/living as an other should provide proof of that. Same goes to people born of one ethnic group but raise under an other. Fairly easy to test, but I doubt productivity has anything to do with genetics, as genetics generally don't have much causation with cultural phenomena.
As with other correlation studies, such studies tend to raise question around the studies themselves. What work are the test subjects doing, and is the result same for all kind? As an example, work demanding multiple spoken languages would benefit of having different people that talk fluently different languages, and people from different places and different cultures often speak different languages fluently. Such aspects need to be considered and removed from studies in order to not create unintended correlations.
Sommers showed causation. The presence of black subjects in the heterogenous group caused the white subjects in that group to behave more conscientiously/objectively, increasing the performance of that group compared to the control.
Others like Ellison/Mullin discuss performance in relation to social capital. Homogenous groups tend to find it easier to build more social capital, so feel more comfortable, but more easily succumb to groupthink.
If they do then shouldn't we do exactly nothing and let the market sort it out?
If the companies with diverse teams are the most efficient then surely they will lead by example.
> I can count on one hand the number of women that sent their CV through, and we interviewed two, hiring one.
Studies show that:
(a) You can create a biased process inadvertently before the CV stage based on the language you use and approach to sourcing.
(b) CVs themselves are also a biasing influence.
(c) Interviews are extremely biasing unless done in a very structured way, and even then they're not great.
So it's very likely that you're (without meaning to) introducing additional bias on top of the constraints on the talent pool. This isn't finger-pointing, it's very hard to run a bias-free process. Even the Googles of the world fail at this.
> ...an environment where gender should make no difference to their desire or interest in STEM subjects
Are you saying that the environment is neutral, or that it should be neutral? First would be wrong, second would be correct.
> Given two candidates...
Your hypotheticals are premised on the idea that your hiring process and judgement are objective, and that you need to positively discriminate in order to redress the balance... that's simply not a view supported by the literature on this.
The population of women in these fields is very small, so the number of exceptional ones is small as well.
Forcing a match to nationwide demographics is inappropriate. That's not the population that should be considered.
Hiring people who are poor coders and collaborators on account of their gender, race, religion, or sexual orientation is equally inappropriate whether it's a white male or minority female.
My female partner in tech has encountered this so regularly she is now suspicious of female hires. Behaviors & performance that would get a male terminated are tolerated and instead show up on a CV as experience at multinational companies. I'm talking basics: unwilling & unable to use source control. This drives resume bias as richmarr highlights.
This is the result of inappropriate affirmative action. It harms the very people you are seeking to help.
"Less than 5% of investors surveyed by LinkedIn rated diversity as their 'top concern' as compared to macroeconomic conditions, raising capital, hitting revenue targets, or hiring the right people. More than half of investors ranked founder commitment to a diverse team as the least of their concerns when considering to invest."
I'd be surprised if even an investment firm made entirely of "women and minorities in tech" had drastically different priorities than this 5%.
Diversity for diversity's sake is not a real solution.
As I've commented above but will summarise here, there are two main points to make in reply to this position
(i) You're assuming the status quo is based on merit, which studies unambiguously show isn't the case
(ii) Studies also show that diverse teams perform better. So this is also about competitive advantage.
Where are all the tech companies created by women, with mostly women as employees capitalizing on the market inefficiency giving them competitive advantage?
Yes they do, but don't get hung up on women. Studies also show a beneficial effect from racial diversity.
As for your other question, is your point that studies are sometimes wrong, or that women are worth less than men, or that there's a gap in the market?
A couple of examples of change in that area;
Companies taking advantage by mothers who want to the workforce once their kids are in school by offering flexible working.
Companies dropping their long-standing policies of hiring from specific schools or attainment levels. These policies are historically quite common in professional services firms, but heavily reduce the available talent pool (and thus the quality of hires, because schools and attainment levels are not strongly predictive of productivity).
Then the mainstream media and people championing social justice causes often take these results the wrong way (even if they do so for noble reasons) and run with it too far.
I think we should compare never married men to never married women, otherwise I believe it to be disingenuous to make that wage gap argument based on gender.
Hoover Institution Senior Fellow Thomas Sowell argues that economic differences between working men and women are not generally due to employer discrimination, as is widely alleged: https://www.youtube.com/watch?v=8EK6Y1X_xa4
You're absolutely right that studies should make sure to control for potentially important variables. I don't know which studies you referring to. Not a topic I've looked at.
That said, criticising a second group of studies can't reasonably cast shade over the first group, if that was your implication.