Evaluating three frameworks for the value of information: adaptation to task characteristics and probabilistic structure

Abstract

We identify, and provide an integration of, three frameworks for measuring the \ninformativeness of cues in a multiple-cue judgment task. Cues can be ranked by information \nvalue according to expected information gain (Bayesian framework), cue-outcome correlation \n(Correlational framework), or ecological validity (Ecological framework). In three \nexperiments, all frameworks significantly predicted information acquisition, with the \nCorrelational (then the Bayesian) framework being most successful. Additionally, \nparticipants adapted successfully to task characteristics (cue cost, time pressure, and \ninformation limitations) – altering the gross amount of information acquired, but not \nresponding to more subtle features of the cues’ information value that would have been \nbeneficial. Rational analyses of our task environments indicate that participants' behavior can \nbe considered successful from a boundedly rational standpoint.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 0 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.