3 ms·

Bureaucracies need not necessarily reduce the average. If they only truncated the tails - equally removing the top performers and the worst performers - then w

Bureaucracies need not necessarily reduce the average. If they only truncated the tails - equally removing the top performers and the worst performers - then we'd expect the average stays the same. If the distribution is right-tailed, with some absolutely stellar performers, then we'd lower the average by removing them. If left-tailed, with some horrendous performers (even legal liabilities), then we'd improve the average by removing them. Bureaucracies at the end of the day are just institutionalized process, and while process by definition reduces variance, it can both raise and lower the average.

abdullahkhalids 3y ago
You are sort of correct, but my claim was putting a lot of its weight on "such" (bureaucracies). I particularly mean bureaucracies who have no expertise and little understanding of the actual work. What happens is that they invent nonsensical metrics to measure performance. The high performers are hampered by these metrics, but the low performers just hack the metrics. For instance, my university's admins were fascinated by the average grade in a course, in a department, for a faculty. To be in their good graces, your students needed to have a 3.5 or higher. Well, what do you think was happening. The bad faculty just had lots of easy assignments in their courses to ensure this. No extra work needed for faculty. Their courses were worse than if they were not subject to this pressure - they might have just stuck to giving a reasonable mix of the textbook problems. The high performers were stuck between a rock and a hard place. Reduce how much you teach your students by adding easy assignments or take the good salary increment at the end of the year. To summarize: remember that the average and variance is for bad metrics that either don't measure what needs to be measured, or worse are anti-correlated with what needs to be measured (and improved). Addendum: actually, I don't even agree with the above framework of metrics. Teaching is an art, and very much like no one runs any sort of metrics on music to decide that Mozart is among the greats, there aren't many real metrics that can be used to identify good teaching from bad.

abdullahkhalids 3y ago

You are sort of correct, but my claim was putting a lot of its weight on "such" (bureaucracies). I particularly mean bureaucracies who have no expertise and little understanding of the actual work. What happens is that they invent nonsensical metrics to measure performance. The high performers are hampered by these metrics, but the low performers just hack the metrics. For instance, my university's admins were fascinated by the average grade in a course, in a department, for a faculty. To be in their good graces, your students needed to have a 3.5 or higher. Well, what do you think was happening. The bad faculty just had lots of easy assignments in their courses to ensure this. No extra work needed for faculty. Their courses were worse than if they were not subject to this pressure - they might have just stuck to giving a reasonable mix of the textbook problems. The high performers were stuck between a rock and a hard place. Reduce how much you teach your students by adding easy assignments or take the good salary increment at the end of the year. To summarize: remember that the average and variance is for bad metrics that either don't measure what needs to be measured, or worse are anti-correlated with what needs to be measured (and improved). Addendum: actually, I don't even agree with the above framework of metrics. Teaching is an art, and very much like no one runs any sort of metrics on music to decide that Mozart is among the greats, there aren't many real metrics that can be used to identify good teaching from bad.