4 ms·

An article that explains the perverse results of "Ban the Box", and then attempts to rebut it: https://onlabor.org/ban-the-box-and-perverse-consequences-part-i

An article that explains the perverse results of "Ban the Box", and then attempts to rebut it: https://onlabor.org/ban-the-box-and-perverse-consequences-part-i/ https://onlabor.org/ban-the-box-and-perverse-consequences-pa... > Ban the Box interacts with racial profiling because, unless it checks or asks, an employer has no information one way or the other about an applicant’s criminal record. > Without individualized information, employers apply racial stereotypes that place applicants of color under a cloud of suspicion. > But when employers check, they confirm not only who does have a record but also who does not, enabling employers to target the former for more accurate exclusion. > The racial exclusion from stereotyping when employers don’t check could exceed the racial exclusion from accurate screening when they do check. > That is exactly what happens, according to the studies.

conanbatt 9y ago
Wow I am impressed at this analysis and finding. I am currently debating with myself about the cost of hiding information, and cases where there is a deliberate attempt at hiding it are the ones that are more likely to counter my intuition that information truly should be absolutely free. Thanks for the contribution!
- throwaway613834 9y ago
  It needs more upvotes. I'm confused at how it's at the bottom of the comments here. It's an excellent one.
Veelox 9y ago
Tl;dr Since employers cannot ask about criminal history, they use stereotypes to make there decision instead. Studies show the stereotype filtering leads to worse outcomes for PoC. I came here to say this, thank you for finding the souce.
geofft 9y ago
Thanks, that data is useful. How are they determining race? Names? Is "ban the name" a sensible policy proposal? (Hold the name in escrow with someone who will contact references, do background checks, etc., until you're ready to make the offer.)
- sokoloff 9y ago
  I personally don't find "ban the name" to be a sensible policy. I google and LinkedIn anyone whom I'm about to interview. I want to be equipped with information to make that hour together as productive as possible for both sides. I get that my personal preference to do homework and be prepared prior to an interview might not outweigh a societal reason to prevent that.
  - geofft 9y ago
    Yeah, I think that is extremely likely to lead to subconscious bias. In the good case, you find something worth talking about with that person, and you have a productive conversation where you go deep onto a topic they should be familiar with and you figure out whether they're actually familiar or not, yes. In the not-so-good case, you find that they were accused and then cleared of charges for something odious, or that they sued their former employer for entirely justifiable reasons, or that they're an ardent supporter of a political party or religious view you just can't stand, or that they just tweeted that they're pregnant, or whatever. And even in the good case, you're asking this person something very different from what you're asking other candidates, which means you're likely to be influenced by something as irrelevant as other candidates not having a public record. (e.g., one candidate worked somewhere with an oppressive OSS policy, one didn't, you got to have a good conversation about a project the second candidate's GitHub but did not get as far talking about an internal project the first candidate did.) Which doesn't actually seem that good. I am incredibly curious to see data, if anyone has it, about who interviews most effectively (in the sense of saying "yes" to candidates that do well, being the one "no" on candidates who left the company quickly, being the one "yes" to candidates whose LinkedIn implies they ended up being successful, etc.) I suspect - but have no data - that asking consistent questions to each candidate leads to higher accuracy. But perhaps you miss out on the rare, extraordinary candidates.
    - sokoloff 9y ago
      > I think that is extremely likely to lead to subconscious bia I agree, but wonder if that process makes the interview more effective rather than less effective overall, bias notwithstanding. It's not like the average unbiased one hour interview has a stellar track record of predicting success, and attempts to create a standardized interview track break down beyond the mid-senior individual contributor or junior manager levels, IME. If I'm interviewing for a Director or VP, I want to have some background information on the candidate to ask specific questions rather than a generic, "Tell me about a time when..." that works at more junior levels.

conanbatt 9y ago

Wow I am impressed at this analysis and finding. I am currently debating with myself about the cost of hiding information, and cases where there is a deliberate attempt at hiding it are the ones that are more likely to counter my intuition that information truly should be absolutely free. Thanks for the contribution!

throwaway613834 9y ago
It needs more upvotes. I'm confused at how it's at the bottom of the comments here. It's an excellent one.

Veelox 9y ago

Tl;dr Since employers cannot ask about criminal history, they use stereotypes to make there decision instead. Studies show the stereotype filtering leads to worse outcomes for PoC. I came here to say this, thank you for finding the souce.

geofft 9y ago

Thanks, that data is useful. How are they determining race? Names? Is "ban the name" a sensible policy proposal? (Hold the name in escrow with someone who will contact references, do background checks, etc., until you're ready to make the offer.)

sokoloff 9y ago
I personally don't find "ban the name" to be a sensible policy. I google and LinkedIn anyone whom I'm about to interview. I want to be equipped with information to make that hour together as productive as possible for both sides. I get that my personal preference to do homework and be prepared prior to an interview might not outweigh a societal reason to prevent that.
- geofft 9y ago
  Yeah, I think that is extremely likely to lead to subconscious bias. In the good case, you find something worth talking about with that person, and you have a productive conversation where you go deep onto a topic they should be familiar with and you figure out whether they're actually familiar or not, yes. In the not-so-good case, you find that they were accused and then cleared of charges for something odious, or that they sued their former employer for entirely justifiable reasons, or that they're an ardent supporter of a political party or religious view you just can't stand, or that they just tweeted that they're pregnant, or whatever. And even in the good case, you're asking this person something very different from what you're asking other candidates, which means you're likely to be influenced by something as irrelevant as other candidates not having a public record. (e.g., one candidate worked somewhere with an oppressive OSS policy, one didn't, you got to have a good conversation about a project the second candidate's GitHub but did not get as far talking about an internal project the first candidate did.) Which doesn't actually seem that good. I am incredibly curious to see data, if anyone has it, about who interviews most effectively (in the sense of saying "yes" to candidates that do well, being the one "no" on candidates who left the company quickly, being the one "yes" to candidates whose LinkedIn implies they ended up being successful, etc.) I suspect - but have no data - that asking consistent questions to each candidate leads to higher accuracy. But perhaps you miss out on the rare, extraordinary candidates.
  - sokoloff 9y ago
    > I think that is extremely likely to lead to subconscious bia I agree, but wonder if that process makes the interview more effective rather than less effective overall, bias notwithstanding. It's not like the average unbiased one hour interview has a stellar track record of predicting success, and attempts to create a standardized interview track break down beyond the mid-senior individual contributor or junior manager levels, IME. If I'm interviewing for a Director or VP, I want to have some background information on the candidate to ask specific questions rather than a generic, "Tell me about a time when..." that works at more junior levels.