3 ms·

The "Do you consider yourself a member of an underrepresented demographic in technology" result is interesting - 92.06% No, but a total summing to 156.3%. Assum

by mrec 8y ago

The "Do you consider yourself a member of an underrepresented demographic in technology" result is interesting - 92.06% No, but a total summing to 156.3%. Assuming this question was phrased as "either No, or check all that apply", this suggests that many of those who do feel underrepresented feel that way on multiple counts. Is the "The survey also highlights some challenges, as the number of women is still lower than the industry average of women in programming fields" editorial comment based solely on this result? It seems to exclude the possibility of women who don't identify as belonging to an underrepresented group.

dep_b 8y ago
It was a bit weird to me as well. There's like 7.x% of gay people which would leave 0.4% for non-gay people that feel that they belong to some kind of minority. It makes gay people look like having a ton of issues at once and I can't imagine that's true. Or is it: "Yes I'm a woman, but I don't feel underrepresented"?
- lern_too_spel 8y ago
  Women are definitely underrepresented in technology relative to their representation in the human population at large. LGBT are not underrepresented in my experience. If anything, a higher percentage of technologists are LGBT vs. the general population.
Vinnl 8y ago
I think being a woman means (and certainly is intended to mean) you're automatically part of an under-represented group - namely, that of women. There are (far) fewer women among Rust users than in the general population, which you could say is the (a?) definition of under-represented. The percentages can add up to >100% because you can be part of multiple groups, e.g. a gay woman.
- mrec 8y ago
  1. Yes, but if the "women" stats only come from a question about self-identifying as underrepresented, your "automatically" doesn't kick in. If women don't check that box, you don't know they're there. 2. I get that, I'm just surprised it's so pronounced. With most statistical distributions, when 0 (flavours of underrepresentation) covers 92% of responses, you'd expect 1 to cover a major chunk of the remainder, and that doesn't appear to be the case here.
swfsql 8y ago
> as the number of women is still lower than the industry average of women in programming fields I think averages shouldn't be goals by themselves. And there's no good reason on why should Rust approach this Hypothetical Average, since Rust Lang != Hypothetical Average Lang. I mean that I guess being closer to this hyp. avg. doesn't make it "better", whatever that means..
ancarda 8y ago
>It seems to exclude the possibility of women who don't identify as belonging to an underrepresented group. Absolutely. If I had taken this survey, I wouldn't have checked "I feel underrepresented because I'm gay" because... well, it's literally never been an issue. In that sense, why would I care if I'm the only gay guy in a team of a million programmers? They all treat me just fine. Must I be "represented"? I can imagine the numbers, for that reason, simply won't reflect the number of non-heterosexual people. It could be higher.
- swfsql 8y ago
  Yes, they used "_do you consider yourself_ underrepresented". And there's also the implicit "and _do you feel like this matters_ for this survey". Some people may feel underrepresented but don't feel like this should matter on this survey, so they'd be likely to answer "No". So a decrease in the "underrepresented" statistics doesn't necessarily mean it's dealing with the actual situation.. It wouldn't necessarily imply anything "bad" or "undesired".
  - Retra 8y ago
    And this is why every survey I've ever taken has pissed me off and now I don't take them at all. I've never felt that the information gleaned from a survey was honestly useful for making decisions.
    - smolder 8y ago
      This is typical of political surveys and other surveys where there is a "preferrable" outcome, but I don't think in this case they are trying to get a specific outcome. Maybe feedback can make subsequent surveys better and reduce the ambiguity of results.
      - Retra 8y ago
        It's typical of all surveys. You either need to go into extensive detail to remove the ambiguity of the questions, or you need to allow people to state what they think in plain language. Taking surveys to gather information is like using record sales to measure what kind of music is best; there are a billion confounding factors and none of them are covered by the abjectly simple metric used.
forapurpose 8y ago
I'm not sure what population the survey intends to look at; it's not Rust users because 25% of respondents aren't using it (per the first question). We don't know how the survey was conducted. We don't know if women or others are overrepresented or underrepresented in the results - and that's relative to an undefined population. We know very little, and thus we can't begin to know the reasoning behind the quoted comment. > women who don't identify as belonging to an underrepresented group The existence of significant numbers of such people seems to be the object of much speculation on HN, possibly because it would undermine attempts to increase the representation of women. However, I have not seen evidence that such a population exists. Also, underrepresentation in this case is much more a matter of math and not one of chosen identity.

by mrec 8y ago

dep_b 8y ago

It was a bit weird to me as well. There's like 7.x% of gay people which would leave 0.4% for non-gay people that feel that they belong to some kind of minority. It makes gay people look like having a ton of issues at once and I can't imagine that's true. Or is it: "Yes I'm a woman, but I don't feel underrepresented"?

lern_too_spel 8y ago
Women are definitely underrepresented in technology relative to their representation in the human population at large. LGBT are not underrepresented in my experience. If anything, a higher percentage of technologists are LGBT vs. the general population.

Vinnl 8y ago

I think being a woman means (and certainly is intended to mean) you're automatically part of an under-represented group - namely, that of women. There are (far) fewer women among Rust users than in the general population, which you could say is the (a?) definition of under-represented. The percentages can add up to >100% because you can be part of multiple groups, e.g. a gay woman.

mrec 8y ago
1. Yes, but if the "women" stats only come from a question about self-identifying as underrepresented, your "automatically" doesn't kick in. If women don't check that box, you don't know they're there. 2. I get that, I'm just surprised it's so pronounced. With most statistical distributions, when 0 (flavours of underrepresentation) covers 92% of responses, you'd expect 1 to cover a major chunk of the remainder, and that doesn't appear to be the case here.

swfsql 8y ago

> as the number of women is still lower than the industry average of women in programming fields I think averages shouldn't be goals by themselves. And there's no good reason on why should Rust approach this Hypothetical Average, since Rust Lang != Hypothetical Average Lang. I mean that I guess being closer to this hyp. avg. doesn't make it "better", whatever that means..

ancarda 8y ago

>It seems to exclude the possibility of women who don't identify as belonging to an underrepresented group. Absolutely. If I had taken this survey, I wouldn't have checked "I feel underrepresented because I'm gay" because... well, it's literally never been an issue. In that sense, why would I care if I'm the only gay guy in a team of a million programmers? They all treat me just fine. Must I be "represented"? I can imagine the numbers, for that reason, simply won't reflect the number of non-heterosexual people. It could be higher.

swfsql 8y ago
Yes, they used "_do you consider yourself_ underrepresented". And there's also the implicit "and _do you feel like this matters_ for this survey". Some people may feel underrepresented but don't feel like this should matter on this survey, so they'd be likely to answer "No". So a decrease in the "underrepresented" statistics doesn't necessarily mean it's dealing with the actual situation.. It wouldn't necessarily imply anything "bad" or "undesired".
- Retra 8y ago
  And this is why every survey I've ever taken has pissed me off and now I don't take them at all. I've never felt that the information gleaned from a survey was honestly useful for making decisions.
  - smolder 8y ago
    This is typical of political surveys and other surveys where there is a "preferrable" outcome, but I don't think in this case they are trying to get a specific outcome. Maybe feedback can make subsequent surveys better and reduce the ambiguity of results.
    - Retra 8y ago
      It's typical of all surveys. You either need to go into extensive detail to remove the ambiguity of the questions, or you need to allow people to state what they think in plain language. Taking surveys to gather information is like using record sales to measure what kind of music is best; there are a billion confounding factors and none of them are covered by the abjectly simple metric used.

forapurpose 8y ago

I'm not sure what population the survey intends to look at; it's not Rust users because 25% of respondents aren't using it (per the first question). We don't know how the survey was conducted. We don't know if women or others are overrepresented or underrepresented in the results - and that's relative to an undefined population. We know very little, and thus we can't begin to know the reasoning behind the quoted comment. > women who don't identify as belonging to an underrepresented group The existence of significant numbers of such people seems to be the object of much speculation on HN, possibly because it would undermine attempts to increase the representation of women. However, I have not seen evidence that such a population exists. Also, underrepresentation in this case is much more a matter of math and not one of chosen identity.