5 ms·

I'm just gonna copy another comment I made elsewhere: The problem is that the gender differences cited cannot, in aggregate, explain the 70/30 bias toward men

I'm just gonna copy another comment I made elsewhere: The problem is that the gender differences cited cannot, in aggregate, explain the 70/30 bias toward men in the industry. There's a real similarity to the global warming debate, here. There are many factors that affect global mean temperature. For example, over the very long term, the sun will get brighter. But that can only account for an astonishingly tiny fraction of the warming observed on the planet. Someone who wishes to debate honestly would admit that, while there are many factors that affect global mean temperature, the rise on CO2 concentration in the atmosphere is significant, and has a demonstrable effect on temperature, making it by far the strongest influencer. Someone who wishes to debate dishonestly will use these other explanations as a method to muddy the debate and try and turn attention toward lesser factors that cannot explain the whole effect. Now, if you look at this current "debate", yes, there are undoubtedly biological gender differences that can explain some male/female bias in engineering (among many other disciplines). But no serious psychologist or sociologist would claim that those differences can account for the dramatic real world bias in the industry. It simply doesn't follow. So to claim those gender differences lie at the root of bias in the industry is a distraction. At best it's a gross misunderstanding of the science and statistics. At worst, it's a deliberately dishonest attempt to muddy the waters and confuse the debate. Personally, my bet is it's more the former than the latter.

rwcarlsen 9y ago
I don't think he has a gross misunderstanding of the science. If you look at the example cartoon distributions he draws for gender differences - there is huge overlap. From my reading, he is advocating for non-discriminatory methods of trying to fix that gap. He's basically saying we shouldn't try to fix discrimination (that we think is bad) with discrimination (that we excuse as good).
- zzalpha 9y ago
  He acknowledges that in practice the gender differences are real, but small. Then uses the existence of those small differences to justify fundamentally changing their hiring policies. The latter does not follow from the former. Again, as with the global warming debate, it makes a lot more sense to tackle the larger influencers first, before addressing these smaller factors.
  - rwcarlsen 9y ago
    I think his thesis is we shouldn't fix discrimination by actually perpetrating the same thing - just against a different group.
    - manicdee 9y ago
      If a see-saw is tipped to one side, you need to apply more weight to the other to balance it out. Nobody is suggesting that you can continue applying weight to the other end forever and still maintain balance, but you can't maintain balance unless you have some to start with. Even achieving a 30/70 split in a male-dominated industry will help defuse some of the rampant masculinity which passes as social interaction (aka primate chest-beating). Having people around who think differently and don't want to conform to the established practices of pissing contests, dick-measuring and chest thumping will help establish better modes of operation and contribute to better decisions being made (no more buying product X because the saleswoman had huge breasts or the salesman displayed dominant primate behaviour).
      - rwcarlsen 9y ago
        I'm not arguing we shouldn't try to get things more balanced. Maybe a better analogy is one side of the see-saw is winning because bullys (gender discrimination) are shoving people off one side. So to fix the problem, we decide to start shoving people off of the other side too. So now it is in balance, but everybody is shoving each other and generally being jerks.
    - nodamage 9y ago
      If you have two equally qualified candidates but 90% of the time the male candidate is hired over the female candidate, and then you change your hiring practices to reduce that ratio to 50%, are you now discriminating against the male candidate? What if you literally reduce it to a coinflip?
      - rwcarlsen 9y ago
        That seems like a bit of a straw-man. The problem is that "equally qualified" is poorly defined. We all have biases (conscious and unconscious) - and we should be working to keep them in check and reduce bad discrimination (i.e. not based on "good" things like merit). I don't think anybody is arguing that they wouldn't want 50/50 hiring results given the repeated game of hiring between two hypothetically equally qualified candidates of each gender. But there are many unknowns. The author is saying that there are differences between the distributions of characteristics between genders (even if small). Some are biologically driven, some cultural. Since we don't know where any particular individual fits in their gender's distribution (also knowing the distributions overlap a ton), we shouldn't wholesale give them different treatment based on their membership in that group. Given representative candidates from each group we should expect hiring to be proportional to the difference in means of the distribution(s) of relevant characteristics (to the job) between the groups (genders) of the applicants. Consider a business where 90% of applicants are female and 10% male. If the business chooses to hire 50/50, they are probably getting some sub-par males assuming the distribution of qualifications between male and female for the two sets of applicants are similar (which we don't have much reason to believe they differ much). In this case the females might be annoyed with the "special" treatment given to male hires that perform worse. And male hires might wonder if they got the job because they were well qualified or maybe just because they were male. In my opinion, this kind of "reverse" discrimination perpetuates bitterness and discriminatory thinking on all sides. I don't think of myself as a strongly religious person, but I think Christ had something when he was teaching that an "eye for an eye" was the old way, but the better way to live is to treat others how you want to be treated.
computerex 9y ago
> But no serious psychologist or sociologist would claim that those differences can account for the dramatic real world bias in the industry. It simply doesn't follow. What's your evidence for this? Debrah Soh, the author supposedly has a PhD in sexual neuroscience so one would think she is qualified to make comments about this. And although I can't confirm this myself, the article does say that societies with greater gender equity also have greater gender gaps due to the differences in what each gender values.
- zzalpha 9y ago
  Well, if we're going to just appeal to authority, I guess I'll do the same: https://www.psychologytoday.com/blog/sexual-personalities/201708/google-memo-about-sex-differences https://www.psychologytoday.com/blog/sexual-personalities/20... The article points out that for most of the biases cited in the memo, the differences are on the order of a few percentage points, which cannot account for the real-world outcomes we see. To quote: Still, I think it's important to keep in mind that most psychological sex differences are only small to moderate in size, and rather than grouping men and women into dichotomous groups, I think sex and sex differences are best thought of scientifically as multidimensional dials, anyway (see here). To be honest, I'm not sure how Ms. Soh can seriously argue the merits of the memo, here. Yes, technically the memo is not itself inaccurate, in that there are gender biases that result in practical population-level effects. But how one can seriously use those facts to conclude that the 70/30 population split in the valley is a product of those differences, I really don't know, unless confirmation bias is at play. It just doesn't make any sense. Heck, just looking historically it doesn't follow... it's not like the biases we see are inevitable. Our industry used to have a different composition. And other STEM fields have seen gender disparities even out. So neither facts nor history support the conclusion.
  - depressedpanda 9y ago
    I'm not sure anybody who has read studies about biological preferences claims that they explain the 70/30 population split. That would be ludicrous. What nobody seems to consider is that an initial small biological difference may very well be amplified into a large cultural preference over time, due to network effects. A self fulfilling prophesy, if you will. I'm just speculating here, but it seems plausible to me, at least.

rwcarlsen 9y ago

I don't think he has a gross misunderstanding of the science. If you look at the example cartoon distributions he draws for gender differences - there is huge overlap. From my reading, he is advocating for non-discriminatory methods of trying to fix that gap. He's basically saying we shouldn't try to fix discrimination (that we think is bad) with discrimination (that we excuse as good).

zzalpha 9y ago
He acknowledges that in practice the gender differences are real, but small. Then uses the existence of those small differences to justify fundamentally changing their hiring policies. The latter does not follow from the former. Again, as with the global warming debate, it makes a lot more sense to tackle the larger influencers first, before addressing these smaller factors.
- rwcarlsen 9y ago
  I think his thesis is we shouldn't fix discrimination by actually perpetrating the same thing - just against a different group.
  - manicdee 9y ago
    If a see-saw is tipped to one side, you need to apply more weight to the other to balance it out. Nobody is suggesting that you can continue applying weight to the other end forever and still maintain balance, but you can't maintain balance unless you have some to start with. Even achieving a 30/70 split in a male-dominated industry will help defuse some of the rampant masculinity which passes as social interaction (aka primate chest-beating). Having people around who think differently and don't want to conform to the established practices of pissing contests, dick-measuring and chest thumping will help establish better modes of operation and contribute to better decisions being made (no more buying product X because the saleswoman had huge breasts or the salesman displayed dominant primate behaviour).
    - rwcarlsen 9y ago
      I'm not arguing we shouldn't try to get things more balanced. Maybe a better analogy is one side of the see-saw is winning because bullys (gender discrimination) are shoving people off one side. So to fix the problem, we decide to start shoving people off of the other side too. So now it is in balance, but everybody is shoving each other and generally being jerks.
  - nodamage 9y ago
    If you have two equally qualified candidates but 90% of the time the male candidate is hired over the female candidate, and then you change your hiring practices to reduce that ratio to 50%, are you now discriminating against the male candidate? What if you literally reduce it to a coinflip?
    - rwcarlsen 9y ago
      That seems like a bit of a straw-man. The problem is that "equally qualified" is poorly defined. We all have biases (conscious and unconscious) - and we should be working to keep them in check and reduce bad discrimination (i.e. not based on "good" things like merit). I don't think anybody is arguing that they wouldn't want 50/50 hiring results given the repeated game of hiring between two hypothetically equally qualified candidates of each gender. But there are many unknowns. The author is saying that there are differences between the distributions of characteristics between genders (even if small). Some are biologically driven, some cultural. Since we don't know where any particular individual fits in their gender's distribution (also knowing the distributions overlap a ton), we shouldn't wholesale give them different treatment based on their membership in that group. Given representative candidates from each group we should expect hiring to be proportional to the difference in means of the distribution(s) of relevant characteristics (to the job) between the groups (genders) of the applicants. Consider a business where 90% of applicants are female and 10% male. If the business chooses to hire 50/50, they are probably getting some sub-par males assuming the distribution of qualifications between male and female for the two sets of applicants are similar (which we don't have much reason to believe they differ much). In this case the females might be annoyed with the "special" treatment given to male hires that perform worse. And male hires might wonder if they got the job because they were well qualified or maybe just because they were male. In my opinion, this kind of "reverse" discrimination perpetuates bitterness and discriminatory thinking on all sides. I don't think of myself as a strongly religious person, but I think Christ had something when he was teaching that an "eye for an eye" was the old way, but the better way to live is to treat others how you want to be treated.

computerex 9y ago

> But no serious psychologist or sociologist would claim that those differences can account for the dramatic real world bias in the industry. It simply doesn't follow. What's your evidence for this? Debrah Soh, the author supposedly has a PhD in sexual neuroscience so one would think she is qualified to make comments about this. And although I can't confirm this myself, the article does say that societies with greater gender equity also have greater gender gaps due to the differences in what each gender values.

zzalpha 9y ago
Well, if we're going to just appeal to authority, I guess I'll do the same: https://www.psychologytoday.com/blog/sexual-personalities/201708/google-memo-about-sex-differences https://www.psychologytoday.com/blog/sexual-personalities/20... The article points out that for most of the biases cited in the memo, the differences are on the order of a few percentage points, which cannot account for the real-world outcomes we see. To quote: Still, I think it's important to keep in mind that most psychological sex differences are only small to moderate in size, and rather than grouping men and women into dichotomous groups, I think sex and sex differences are best thought of scientifically as multidimensional dials, anyway (see here). To be honest, I'm not sure how Ms. Soh can seriously argue the merits of the memo, here. Yes, technically the memo is not itself inaccurate, in that there are gender biases that result in practical population-level effects. But how one can seriously use those facts to conclude that the 70/30 population split in the valley is a product of those differences, I really don't know, unless confirmation bias is at play. It just doesn't make any sense. Heck, just looking historically it doesn't follow... it's not like the biases we see are inevitable. Our industry used to have a different composition. And other STEM fields have seen gender disparities even out. So neither facts nor history support the conclusion.
- depressedpanda 9y ago
  I'm not sure anybody who has read studies about biological preferences claims that they explain the 70/30 population split. That would be ludicrous. What nobody seems to consider is that an initial small biological difference may very well be amplified into a large cultural preference over time, due to network effects. A self fulfilling prophesy, if you will. I'm just speculating here, but it seems plausible to me, at least.