If we want to improve something, one common strategy is to create measurements to track whether we are getting closer to our goal or not. However, if we measure what is easy instead of what is actually meaningful for the end goal, we can go down a dangerous path.

In a recent Freakonomics episode where they talked about policing in the US, one critique is that their police departments tend to measure things like the number of arrests. When the end goal is to reduce crime, simply arresting more people does not necessarily mean you’re solving the problem.

Another context where this happens is in academia. We want to reward people for producing valuable knowledge that will benefit society, but that’s really hard to quantify. Instead we resort to bean counting like the number of first/last author manuscripts, h-value, impact factor etc.