3 ms·
> In short: one in five respondents say they have received explicit instruction for a ‘no white men’ policy. Another one in five say they’ve been implicitly ins
by gustavus 3y ago
> In short: one in five respondents say they have received explicit instruction for a ‘no white men’ policy. Another one in five say they’ve been implicitly instructed to follow this kind of a policy.
- xg15 3y ago... meanwhile the entire rest of the article is how you can promote diversity without having that kind of policy.
- 34679 3y agoThere's only one right way: to acknowledge that your skin color and sexual orientation are irrelevant to your ability to perform your job. The hiring process should never involve these attributes.
- xg15 3y agoI think what you can reasonably do is to monitor how many employees of a certain race/gender/etc are in a particular job - and if the distribution is significantly unlike the general distribution of the population, see that as an indicator that something is wrong. That doesn't mean that try the first available thing to "fix" the distribution, but it would be a good starting point to ask further questions why it is so skewed - and then try to identify root causes. I.e., if you only hire by skills and then find that you still hire 80% men, even though they only make up 50% of the population, then obviously something else is going wrong. The question is what.
- EForEndeavour 3y agoContinuing this line of thought: - I feel like you'd have to compare demographics of your successful applicants to something more specific than general population. Maybe demographics of your applicant pool? - What do you do if the root cause turns out to be "upstream" of your company's hiring process, e.g., social conditioning from a young age throughout education that introduces disparities in field of study across demographic categories?
- xg15 3y ago> - What do you do if the root cause turns out to be "upstream" of your company's hiring process, e.g., social conditioning from a young age throughout education that introduces disparities in field of study across demographic categories? Yes, absolutely. It's a large-scale problem and probably has to be solved at a societal level.
- raxxorraxor 3y agoYou first have to determine if that really is a problem though. And put evidence forth before you start to actively discriminate.
- benmmurphy 3y agothis is not even something necessarily wrong in your hiring policy or even in your industry. you could have lower numbers of certain groups because other industries or companies are carrying out discrimination in favour of those groups.
- neotrope 3y ago> if the distribution is significantly unlike the general distribution of the population, see that as an indicator that something is wrong. most technical roles are >80% men (e.g., google) but companies don't provide breakdowns by functional roles to avoid conversation with people who think like this. are you ageist if 12% of your talent isn't above the age of 60? once you open a debate about available talent, it ceases to be a useful metric.
- theironhammer 3y agoSo if the local daycare is staffed by all women there's something wrong? Thus, diversity hire to the rescue?
- seanmcdirmid 3y agoKindergarten teachers also. Almost 100% women. What is wrong with this world?
- raxxorraxor 3y ago> I think what you can reasonably do is to monitor how many employees of a certain race/gender/etc are in a particular job Obviously we can do that technically but maybe we shouldn't. We did monitor it and that lead to people having some most arcane conclusions they derived from this data. If you hire 80% for IT tech jobs, you probably just hire the average available on the job market. Some deduced this has to be discrimination and maliciously smeared the whole industry. This should not be repeated. So no, people maybe should keep their ethnicity private if they don't want it to be abused by busy bodies making everyone's live worse.
- houseatrielah 3y agoIf diversity is the goal: there are going to be fewer white men because they don't count as diverse. It's tautological. Anything else is speaking out of both sides of the mouth.
- xg15 3y ago> because they don't count as diverse Well yes. If your fruit salad consists of 80% apples and 20% pears, you're not going to get a better mix by adding even more apples. That's basic math, even if the apples might not like it. That being said, I keep reading about growing pies and not-zero-sum games, so you could apply the same logic here as well: if the percentage of women (vs men) or blacks (vs whites) etc should be increased, that doesn't necessarily mean that the absolute counts of men has to be reduced or that from this day on, no more men have to be employed for new jobs. It just means that employment numbers of women have to grow faster than those of men. Also, if there are root causes that can be addressed, such as different training biographies, it could be possible to archive that change without mandating any kind of specific employment policies.
- houseatrielah 3y agoIts weird though, because there's no combination where some HR manager is going to say, hmm, lets get a white guy in there. 100% (non-white) same-ethnicity is very diverse according to HR metrics, but its the opposite of the meaning of the word.