7 ms·
There seems to be a trend in management towards an unwarranted belief in the effectiveness and importance of superficial quantitative analysis, and two compleme
by mannykannot 10y ago
There seems to be a trend in management towards an unwarranted belief in the effectiveness and importance of superficial quantitative analysis, and two complementary myths that go along with the trend: a) any number you can assign to something is significant; b) anything you cannot assign a number to is not significant.
Anyone approaching recruiting with this mindset is neither using critical thinking, nor likely to hire people who can.
- dilemma 10y agoWhat gets measured gets managed.
- keithpeter 10y agoAnd often there are titanic struggles to manage noise. I mean changes of 1% with sample sizes of 200 or so (education). I especially cherish 'business systems' that produce percentages accurate to two decimal places.
- sixdimensional 10y agoIt's almost as if we have regressed in some ways to the "principles of scientific management" (Frederick Winslow Taylor) with our focus on quantitative over qualitative (and not balanced between the two).
- angry_octet 10y agoThere has been an uptick in user of discredited techniques to screen in recruitment, stuff like MBTI (Myers-Briggs Type Indicator), lots of pseudo science without measuring validity. It seems like HR have been told to find people with soft skills and so they manufacture something that will give a rating and then rank candidates by that.
- naveen99 10y ago> mbti discredited Do you have any good citations? A majority of hacker news readers are INTJ and intp, 2 types that generally find mbti useful in understanding themselves and others. I doubt hr puts much stock in mbti anyway, because it would be an easy test to game if it was officially considered.
- angry_octet 10y agoGoogle 'mtbi validity'. It is very possible HN readers are INTJ because they see e.g. introversion as one of their traits -- but it doesn't mean it is true! Poor construct validity is one of the problems. For example, if I ask you, "Are you creative in how you solve problems?" the answers may only slightly correlate with creative problem solving. A very accessible article: https://www.psychologytoday.com/blog/give-and-take/201309/goodbye-mbti-the-fad-won-t-die https://www.psychologytoday.com/blog/give-and-take/201309/go... http://indiana.edu/~jobtalk/Articles/develop/mbti.pdf http://indiana.edu/~jobtalk/Articles/develop/mbti.pdf http://epublications.bond.edu.au/cgi/viewcontent.cgi?article=1026&context=hss_pubs http://epublications.bond.edu.au/cgi/viewcontent.cgi?article... Overall, MBTI is more reliable than astrology, and has proven to generate significant sums for the owner, CPP.
- naveen99 10y agoThanks for the references, I read them. > inconsistency of results Atleast on hacker news, the test is fairly consistent. There have been several surveys here. > lack of statistical analyzability That is the academic psychologist's problem. > 89 / 100 fortune 100 companies use it internally A lot of people hear recommend blindly investing in s&p 500. I trust fortune 100 companies more than psychology associations and bloggers. > college board chose not to use it Yes. Because it is one static test that can be memorized to produce desired result by test taker. But if one takes it for their own understanding, there is no reason to cheat (and it becomes useful) Not convinced that these two blog posts and 1 white paper opinion piece count as discrediting mbti. I will look for a meta review. In the end, I think personality testing, similar to iq testing is not politically correct. But it can be useful, as a self test atleast. Companies prefer to use it but keep the process non transparent. They like the flexibility of being able to say "personality fit" without the transparency of a test because one it doesn't open up liability and two the test is gameable in an adversarial situation. Atleast I haven't seen any claims of racial differences in mbti, which might make it useful as a way to find engineer / rational types in underrepresented groups in software.
- daniel-levin 10y agoAlso, when a metric becomes a target, it ceases to be a good metric, regardless of its original efficacy. This is widely known as Goodhart's law [0] [0] https://en.wikipedia.org/wiki/Goodhart%27s_law https://en.wikipedia.org/wiki/Goodhart%27s_law
- Cacti 10y agoThis is not just a trend in management. Many, many people labor under the assumption that if you throw enough (or any, for that matter) math at something you must be doing science, and therefore it's inherently better than the alternative. I mean, how many times have you heard "Well, sure, it's not exact, but it's better than what we have" in reference to some model that was made up from a pile of assumptions? No, no, no, it's not better than the alternative, it's exactly the same, that's the entire point! Just because you threw some statistics at it doesn't mean you're any better off than you were before.