Models of Learning Sample Clauses

Models of Learning. ‌ Several recent models in economics offer a variation on the classical Bayesian agent. This paper fits best with Xxxxxxxx (2012), Xxxxx and Xxxxxx (1999), Schwartzstein (2014), and Xxxxxx (2014), in that each examines passive learning under cognitive limitations. Again, a behavioral assumption is not required in the current model. The sender might be the one collapsing the message into a binary signal. Other work examines learning in a more active environment, where an action is taken in multiple stages (e.g., Xxxxxx and Smorodinsky 2000), as in repeated games (see Xxxxx 2000 for an early survey). Here the only game is the vote in the committee setting, so no individual has the opportunity to learn from any action aside from votes. Inattention is a potential source of error in the model, but unlike the work of Xxxx (2003) or Xxxxxx and Xxxx (2015), attention is not derived from a direct constraint on cognition. The only explicit cost of inattention is in reducing the quality of the ultimate decision. Furthermore, since the stopping rule is exogenous, the opportunity cost of continuing to review information is assumed zero, unlike classic stopping rule models following Wald (1947). Lastly, Section VI presents a moving average model inspired by Xxxxxx (1993) and Xxxxxxxxxx and Xxxxxx (1992), though Cross (1973) introduces a similar model where the average moves over probabilities instead of states.

Models of Learning