Passive learning: a critique by example

In models of learning by experimentation that exhibit signal dependence, a benchmark using a passive learner has been proposed. The use of this benchmark is flawed – first, passive learning does not disentangle the effects of knowing that beliefs, as well as other state variables, might change, and we address this issue directly by introducing a naïve learner. Secondly, and more tellingly, passive learning does not do what it is supposed to do, namely help measure the gains from active experimentation; the naïve learner enables us to illustrate this point in the context of a particular example.