Slashdot Mirror
Gender and Tenure Diversity In GitHub Teams Relate To Higher Productivity
New submitter Bogdan Vasilescu writes: Diversity in teams is a double-edged sword. Increased team diversity results in more varied backgrounds and ideas, providing the team with access to broader information, enhanced creativity, adaptability, and problem solving skills. However, due to greater perceived differences in values, norms, and communication styles in more diverse teams, members become more likely to engage in stereotyping, cliquishness, and conflict.
In a recent study, researchers from University of California, Davis and Eindhoven University of Technology, The Netherlands have analyzed the effects of gender and tenure diversity on productivity and turnover for more than 23,000 open-source projects on GitHub. Using regression modeling, they showed that after controlling for team size and other confounds (such as a project's age, development model, or amount of social activity), both gender and tenure diversity are positive and significant predictors of productivity, together explaining a small but significant fraction of the data variability. On an economic and societal scale, these findings suggest that added investments in educational and professional training efforts and outreach for female programmers will likely result in added overall value.
The paper describing the results (preprint PDF here) will be presented at the prestigious ACM CHI Conference on Human Factors in Computing Systems, in Seoul, South Korea, in April 2015.
In a recent study, researchers from University of California, Davis and Eindhoven University of Technology, The Netherlands have analyzed the effects of gender and tenure diversity on productivity and turnover for more than 23,000 open-source projects on GitHub. Using regression modeling, they showed that after controlling for team size and other confounds (such as a project's age, development model, or amount of social activity), both gender and tenure diversity are positive and significant predictors of productivity, together explaining a small but significant fraction of the data variability. On an economic and societal scale, these findings suggest that added investments in educational and professional training efforts and outreach for female programmers will likely result in added overall value.
The paper describing the results (preprint PDF here) will be presented at the prestigious ACM CHI Conference on Human Factors in Computing Systems, in Seoul, South Korea, in April 2015.
I suspect this may be a matter of the disposition of the team causing both diversity and efficiency, rather than diversity directly causing efficiency. Teams that aren't full of a-holes tend to accept more varied members and end up more diverse, and also happen to work well together...
Quoting from the paper
So, basically if you use a completely BS measure of "productivity" (# of commits) teams that are more diverse and with longer tenure tend to be slightly (1% to 2.5%) more "productive".
Where do I sign up for grant money to produce crap research like this?
... actual diversity.
Having a guy and a girl does not mean you have diversity of thought or ideas or talent or ability. It means you have one person that can pee standing up and another that can gestate babies if you give them 9 months and all the ice cream.
If you have a white girl and a black guy that does not mean you have diversity. You have one person that sits down to pee and one person that has to make sure he wears bright clothing when he walks at night to avoid getting hit by cars.
REAL diversity is diversity of MIND not diversity of sex organ or color. You can have more diversity with 10 black women then some random mix of genders and races because those 10 black women might actually be very different people while your random mix of genders and races might only be different on the surface.
This obessession with statistical diversity figures is tedious, stupid, and increasingly depressing.
I am not my gender.
I am not my skin color.
I am not my nationality.
How many people do you think that share my phenotypes are likely to agree with me or think the same way as I do about everything? I assure you... not a lot of them do. I am an odd duck. I am as likely to find someone with a vagina or a different shade of skin that holds my views or thinks the way I do as one that superficially resembles me.
Diversity is not gender or skin color. Diversity is mind.
And if there is anything ironic about the people crying for diversity it is how intolerant they frequently are of anyone that holds any view that is even a little different from their own.
What ever their intentions, they're not creating diversity. They're bullying everyone around them to fall into exact lockstep according to arbitary statistical data that doesn't actually mean anything.
Imagine if we were a bunch of cats. Does it really f'ing matter how many cats of some random number of spots are doing one thing or another?
This is the 21st century. Get over it. You're fighting last century's civil rights issues instead of opening your eyes to the civil rights issues of the 21st century.
What you should be concerned with are things like privacy, corruption, freedom, economic mobility, free flow of information, etc.
That is something that will actually matter. This gender/race crap is relevant only so far as you're dealing with actual racists etc. And since most of these claims boil down to subconscious bias the reality is that we know we're not dealing with the old bigotry anymore. There are bigger issues out there. Focusing on this if anything simply makes it worse because you're making it harder for people to be gender/race blind.
All this crap boils down to at some point is that people want UNofficial race/gender quotas. I think Channel 4 in the UK recently told their managers they'd only get bonuses if their staffs reflected specific race and gender quotas. That is what this stuff creates.
And assuming you got race and gender quotas ubiquitously imposed everywhere, how would that be good for traditionally discriminated populations? After all, everyone would know that they HAVE to be hired even if they're not competent because there are quotas. Which means everyone will assume right or wrong that they're incompetent because they can be hired without being competent.
Just think it through, kids. You're like that simple kid that shows up to track and just runs through all the hurdles without understanding he was supposed to try to jump over them.
I've decided to stop wasting my time responding to AC trolls/sockpuppets... so if you want a response from me... login.
Doubtless they had some statistics, mostly bad, and went with one that looks better than the others. The advantage is that few people think of commit count as productivity, so they don't game it. I'd like to see the study redone over a longer time, so there could be productivity statistics based on the progress of projects rather than a simple activity measure.
There's also possible biases here. It could be that women are more conservative in what they want to do than men, so the projects that get started and really don't go anywhere (because they're impossible, infeasible, or overly ambitious) tend to be man-heavy. It could be that male chauvinist assholes don't generally work well with anybody, so they first drive women away and then make the project suffer, while men women like to work with tend to do better as a team. Since this is open source development, where everybody works on what they personally decide to do, this will be hard to separate out.
"When you have eliminated the unacceptable, whatever is left, however improbable, must be the truthiness" - Holmes