<?xml version="1.0" encoding="UTF-8"?>
<feed xmlns="http://www.w3.org/2005/Atom" xmlns:dw="https://www.dreamwidth.org">
  <id>tag:dreamwidth.org,2016-12-24:2596992</id>
  <title>mykyta_p</title>
  <subtitle>mykyta_p</subtitle>
  <author>
    <name>mykyta_p</name>
  </author>
  <link rel="alternate" type="text/html" href="https://mprotsenko.dreamwidth.org/"/>
  <link rel="self" type="text/xml" href="https://mprotsenko.dreamwidth.org/data/atom"/>
  <updated>2022-01-26T02:10:08Z</updated>
  <dw:journal username="mprotsenko" type="personal"/>
  <entry>
    <id>tag:dreamwidth.org,2016-12-24:2596992:12070</id>
    <link rel="alternate" type="text/html" href="https://mprotsenko.dreamwidth.org/12070.html"/>
    <link rel="self" type="text/xml" href="https://mprotsenko.dreamwidth.org/data/atom/?itemid=12070"/>
    <title>A definition of equitable approach to hiring</title>
    <published>2022-01-26T00:14:23Z</published>
    <updated>2022-01-26T02:10:08Z</updated>
    <category term="equity"/>
    <dw:security>public</dw:security>
    <dw:reply-count>14</dw:reply-count>
    <content type="html">Defining equity in hiring (and everywhere else) is not as easy as it may sound - "equity" is not an established term and may be carry different meanings for different people.&lt;br /&gt;&lt;br /&gt;&lt;br /&gt;I can see at least three ways to define equitable hiring process:&lt;br /&gt;&lt;br /&gt;1) All groups are proportionally represented.&lt;br /&gt;&lt;br /&gt;2) Candidates from all groups have the same false negative rate (fit candidates are not hired equally across all groups).&lt;br /&gt;&lt;br /&gt;3) Candidates from all groups have the same false positive rate (unfit candidates are hired equally across all groups).&lt;br /&gt;&lt;br /&gt;The funny fact is that implementing #2 and #3 looks suspiciously similar to affirmative action. Why?&lt;br /&gt;&lt;br /&gt;&lt;br /&gt;&lt;br /&gt;&lt;br /&gt;Let say we have an imperfect proxy metric that defines future success and this metric has different distribution for different groups.&lt;br /&gt;&lt;br /&gt;A really simplified example would be standard tests' scores - let's say we have group A with a limited resources to practice tests, tutors etc. The more talented students will still score higher.&lt;br /&gt;&lt;br /&gt;On the other hand, in group B, where students have access to everything they can dream of, even dumb students can score higher than the best students from the group A.&lt;br /&gt;&lt;br /&gt;However, in each respective group the score can still be a predictor of a long-term success - the best students from each group will achieve similar results in their careers. (Yes, yes, I am going on a limb here, but this is a simplified example - so please bear with me.)&lt;br /&gt;&lt;br /&gt;&lt;br /&gt;&lt;br /&gt;If we define equity as #2 and #3 - then we would need to define different thresholds for different groups. It may sound counterituitive - but different thresholds can lead to equalizing opportunities for equally talented people with different access to resources, &lt;b&gt;when we use an imperfect proxy metric to measure a probability of a future performance&lt;/b&gt;.&lt;br /&gt;&lt;br /&gt;&lt;br /&gt;Any flaws in this logic? Comments/feedback are welcome.&lt;br /&gt;&lt;br /&gt;&lt;img src="https://www.dreamwidth.org/tools/commentcount?user=mprotsenko&amp;ditemid=12070" width="30" height="12" alt="comment count unavailable" style="vertical-align: middle;"/&gt; comments</content>
  </entry>
</feed>
