ESTIMATION STRATEGIES AND THE JUDGMENT OF EVENT FREQUENCY

Processes underlying judgments of absolute event frequency were investigated in 3 experiments. In all 3, word pairs consisting of a target (a category label, e.g., CITY) and context (a category exemplar, e.g., London) were presented in a different- or same-context study list. In the different-context condition, each target was paired with a new context on each presentation; in the same-context condition, a target always appeared with the same context. Verbal protocols (Experiment 1) and response times (Experiments 2 and 3) indicate that multiple estimation strategies were used and that strategy selection was related to memory contents. In particular, different-context participants often enumerated, and same-context participants did not. Also, because range information only affected same-context estimates (Experiment 3), it appears that a numerical conversion process was necessary when nonenumeration strategies were used.