I think there is a deeper unreasonableness. What would a truly scientific hiring process look like? What might a large organisation, such as the British Civil Service do to make sure that it hires the right people?
One idea is to score candidates on many measures (100?) and make use guessed weights to guide initial hiring decisions. A decade later the organisation can compare the initial assessments against the hires subsequent progress up the organisation. This leads to a re-assessment of the weights.
Notice that the new weights were computed from the poputation of hires, not the population of candidates, and the hires were chosen using the old weights. The new weights are causing the organisation to hire people that it would have rejected in the past and it has no solid data on how those people would have performed in the organisation - it didn't hire them! There are other problems. The whole approach has too many parameters to calibrate it reliably from the data available. The linearity implicit in a weighted sum is implausible, a more sophisticated model with the number of variables squared or cubed is clearly required. The time scale on which the success of the hiring decisions becomes apparent is long, decades. It is the time scale on which things change. So the new scientific weights might be no better than rebooting the whole process with new, guessed weights.
My point is that organisations don't have substantial reasons for their hiring decisions. They have a process because they are forced to chose, and they consequently have procedural reasons for their hiring decisions.
A small organisation, such as Fog Creek, that has scientifically trained personnel, may be well aware that it hires the wrong people. Those in charge know that they hire the wrong people and that they cannot fix the problem. They also know that business is competitive, they only have to do better than their competitors, and since the problem is pretty much impossible, their competitors cannot fix it either.
The serious answer to the hiring question is "We don't know how to chose. We make wild guesses and then pretend that we have good reasons for them so that we don't feel too bad afterwards. The whole topic is really embarrassing. You shouldn't have asked."
One idea is to score candidates on many measures (100?) and make use guessed weights to guide initial hiring decisions. A decade later the organisation can compare the initial assessments against the hires subsequent progress up the organisation. This leads to a re-assessment of the weights.
Notice that the new weights were computed from the poputation of hires, not the population of candidates, and the hires were chosen using the old weights. The new weights are causing the organisation to hire people that it would have rejected in the past and it has no solid data on how those people would have performed in the organisation - it didn't hire them! There are other problems. The whole approach has too many parameters to calibrate it reliably from the data available. The linearity implicit in a weighted sum is implausible, a more sophisticated model with the number of variables squared or cubed is clearly required. The time scale on which the success of the hiring decisions becomes apparent is long, decades. It is the time scale on which things change. So the new scientific weights might be no better than rebooting the whole process with new, guessed weights.
My point is that organisations don't have substantial reasons for their hiring decisions. They have a process because they are forced to chose, and they consequently have procedural reasons for their hiring decisions.
A small organisation, such as Fog Creek, that has scientifically trained personnel, may be well aware that it hires the wrong people. Those in charge know that they hire the wrong people and that they cannot fix the problem. They also know that business is competitive, they only have to do better than their competitors, and since the problem is pretty much impossible, their competitors cannot fix it either.
The serious answer to the hiring question is "We don't know how to chose. We make wild guesses and then pretend that we have good reasons for them so that we don't feel too bad afterwards. The whole topic is really embarrassing. You shouldn't have asked."