Measuring the Use of Recognition with the Multinomial r-Model: An Application to Aging and Individual Differences

Abstract

It has been suggested that individuals use simple decision strategies for comparative judgments. According to the recognition heuristic (RH; Goldstein & Gigerenzer, 1999), people infer that a recognized object scores higher on a criterion, if one of two objects is recognized, but not the other. Hilbig, Erdfelder, and Pohl (2010) have argued that previous research lacked process-pure estimates of RH use and rigorous model testing and proposed the r-model, a multinomial processing tree (MPT) model. Addressing these methodological issues, we present a first MPT analysis of differences in RH use between younger and older adults. Model-based analyses indicated that in both age groups the RH was used adaptively more often in the environment with higher recognition cue validity (cities), as opposed to a domain with lower cue validity (diseases). The validity of further knowledge or recognition as decision cues did not differ between age groups. Moreover, we examined the model estimates on the individual level by applying a Bayesian hierarchical approach and compared these estimates with behavioral indices and measures derived from signal-detection theory. The resulting comparisons with standard RH-adherence rates indicated high correlations. Further implications are discussed.


Back to Table of Contents