4 ms·

I'm studying statistics just at the moment, and for anyone who is curious and has an open mind, the most likely (in a Bayesian sense) reason that the OP has pos

I'm studying statistics just at the moment, and for anyone who is curious and has an open mind, the most likely (in a Bayesian sense) reason that the OP has posted this comment is because they do not have access to a single female Senior Software Engineer who they feel meet their requirements. (This is more likely than an alternative explanation such as that the best candidate for the position OP wishes to fill is female, but OP refuses to hire a female.) For anyone who is curious and has an open mind, since women are half of the population and attain a large number of degrees (as opposed to many decades ago when they were not allowed in education at all), why does OP not have access to female candidates OP feeels meets their requirements? Since I am studying statistics, I looked up the answer in a lecture about women, collected in a book called the Past, Present, and Future of Statistical Science[1] There are interesting comments about sexism within the science of statistics, for example: Thwarted employment search after college Having discarded high school and college teaching, actuarial science, and medicine, I sought employment after college graduation in 1960. I was aware of only two methods to find a job: look in the newspapers’ “Help Wanted” sections and talk with employers at job fairs on campus. The newspaper route proved fruitless. Younger readers may not be aware that newspapers had separate “Help Wanted Female” and “Help Wanted Male” sections until the late 1960s or early 1970s when such practice eventually was ruled to be illegal sex discrimination. In 1960 advertised positions using math skills and interest were in “Help Wanted Male,” and I assumed that it would be futile to apply. Job interviews on campus with employers played out similarly; all positions were segregated by gender and all technical positions were for males. One vignette, among many, illustrates the employment culture for women in the US in 1960. When I registered for an interview on campus with IBM, I was required to take a math aptitude test. The IBM interviewer commented that he had never seen such a high score from any applicant and offered me either a secretarial or an entry sales position. I countered that I was interested in their advertised technical positions that required a math background, especially given my score on their math aptitude test, but he simply said that those positions were for males. End of conversation. - Clearly, this process explains why OP does not have access to female applicants he feels are qualified. Perhaps OP is judging the history of candidates and applicants, but many other people judged their gender in the past when excluding them. What should OP do? Of course, OP can't force others who excluded women from going back in history and including them instead. Is there anything OP can do now to try to rectify the injustice? I think that one thing that OP could potentially do is follow the process outlined in the company. He could hire a woman for the role. I think it is the right thing to do in OP's situation and given the history of the industry. [1] if you search for this title on Google one of the top results is the full PDF so I think it's okay to post it here: https://higherlogicdownload.s3.amazonaws.com/AMSTAT/71a758c7-5229-4729-bb64-caf9d1cf855f/UploadedImages/past-present-future-copss.pdf https://higherlogicdownload.s3.amazonaws.com/AMSTAT/71a758c7...

nervlord 5y ago
Past discrimination doesn't justify discrimination in the present.
- logicallee 5y ago
  >Past discrimination doesn't justify discrimination in the present. You show a kind of logical thinking, since it is in fact a kind of discrimination in the present. However, you can verify for yourself that this is correct through something like the following. It's pseudocode, to help you reason correctly: // assume all we know about the candidate is whether they are "A" or "B" and their years of experience. // generate 10,000 candidates candidates = 10,000 x random (a or b) for each candidate trueability = random (1 to 100) // their true ability which we don't know is a random number 1 to 100 for each candidate years of experience = random (1 to 20) // however, let's corrupt our data. Let's artificially reduce the years of experience for all candidates who are "A". // let's add sexism for each candidate: if candidate gender == "A" then years of experience = 0 Under this scenario, is using the proxy "years of experience" the best way to get trueability? Bag of candidates = 10 random candidates from all candidates How can you judge the best candidate if the most salient characteristic (years of experience) is artificially constrained? This is the situation under the true historical context I quoted above: >The IBM interviewer commented that he had never seen such a high score from any applicant and offered me either a secretarial or an entry sales position. I countered that I was interested in their advertised technical positions that required a math background, especially given my score on their math aptitude test, but he simply said that those positions were for males. End of conversation. This means that my outrageous pseudocode with this corrupt line that caps talented candidates at 0 years of experience - actually describes real life. Continuing to use the proxy rather than get the trueability is not justified. Under such a scenario, it is no longer possible to use the salient proxy characteristic to judge the candidate. I want you to really get this, so I'll use a final motivating example. - Should a company hire violent rapists? Of course not. They are a threat to other employees. Okay, so what if someone accused you of being a violent rapist even though you were innocent? That would suck! Imagine if you tried to appeal but were told: "Sorry. We understand you were falsely accused in the past, but a false accusation in the past does NOT justify us spending resources on determining whether the accusation is true or not. It is easiest for us to use the proxy of not hiring anyone like you." I use this example because if I were a woman denied a job just because I wasn't a man, I would feel it is a gross injustice similar to a false accusation. So, you should use that feeling to realize that the "discrimination" you're complaining about is in fact righting a wrong. You are saying: "Righting past wrongs does not justify spending present resources". This is a moral question. Personally, I feel I have enough resources to spend an extra few seconds each candidate to see their true ability, rather than applying a heuristic based on past discrimination. In my experience this also gives me access to better teams faster. If that comes at the cost that I have to get to know candidates it seems like a small price to pay.

nervlord 5y ago

Past discrimination doesn't justify discrimination in the present.

logicallee 5y ago
>Past discrimination doesn't justify discrimination in the present. You show a kind of logical thinking, since it is in fact a kind of discrimination in the present. However, you can verify for yourself that this is correct through something like the following. It's pseudocode, to help you reason correctly: // assume all we know about the candidate is whether they are "A" or "B" and their years of experience. // generate 10,000 candidates candidates = 10,000 x random (a or b) for each candidate trueability = random (1 to 100) // their true ability which we don't know is a random number 1 to 100 for each candidate years of experience = random (1 to 20) // however, let's corrupt our data. Let's artificially reduce the years of experience for all candidates who are "A". // let's add sexism for each candidate: if candidate gender == "A" then years of experience = 0 Under this scenario, is using the proxy "years of experience" the best way to get trueability? Bag of candidates = 10 random candidates from all candidates How can you judge the best candidate if the most salient characteristic (years of experience) is artificially constrained? This is the situation under the true historical context I quoted above: >The IBM interviewer commented that he had never seen such a high score from any applicant and offered me either a secretarial or an entry sales position. I countered that I was interested in their advertised technical positions that required a math background, especially given my score on their math aptitude test, but he simply said that those positions were for males. End of conversation. This means that my outrageous pseudocode with this corrupt line that caps talented candidates at 0 years of experience - actually describes real life. Continuing to use the proxy rather than get the trueability is not justified. Under such a scenario, it is no longer possible to use the salient proxy characteristic to judge the candidate. I want you to really get this, so I'll use a final motivating example. - Should a company hire violent rapists? Of course not. They are a threat to other employees. Okay, so what if someone accused you of being a violent rapist even though you were innocent? That would suck! Imagine if you tried to appeal but were told: "Sorry. We understand you were falsely accused in the past, but a false accusation in the past does NOT justify us spending resources on determining whether the accusation is true or not. It is easiest for us to use the proxy of not hiring anyone like you." I use this example because if I were a woman denied a job just because I wasn't a man, I would feel it is a gross injustice similar to a false accusation. So, you should use that feeling to realize that the "discrimination" you're complaining about is in fact righting a wrong. You are saying: "Righting past wrongs does not justify spending present resources". This is a moral question. Personally, I feel I have enough resources to spend an extra few seconds each candidate to see their true ability, rather than applying a heuristic based on past discrimination. In my experience this also gives me access to better teams faster. If that comes at the cost that I have to get to know candidates it seems like a small price to pay.