Fluency heuristic: A model of how the mind exploits a by-product of information retrieval.

Abstract

Boundedly rational heuristics for inference can be surprisingly accurate and frugal for several reasons. They can exploit environmental structures, co-opt complex capacities, and elude effortful search by exploiting information that automatically arrives on the mental stage. The fluency heuristic is a prime example of a heuristic that makes the most of an automatic by-product of retrieval from memory, namely, retrieval fluency. In 4 experiments, the authors show that retrieval fluency can be a proxy for real-world quantities, that people can discriminate between two objects' retrieval fluencies, and that people's inferences are in line with the fluency heuristic (in particular fast inferences) and with experimentally manipulated fluency. The authors conclude that the fluency heuristic may be one tool in the mind's repertoire of strategies that artfully probes memory for encapsulated frequency information that can veridically reflect statistical regularities in the world.

The psychology of intelligence analysis: Drivers of prediction accuracy in world politics.

“…Word frequency is an indicator of a word's fluency; more frequently occurring words are easier to access than less frequently occurring words (e.g., Hertwig, Herzog, Schooler, & Reimer, 2008;McClelland & Rumelhart, 1981). Though low frequency words may be more likely to elicit TOTs than high frequency words (e.g., Burke et al, 1991), if participants are inclined to infer from a TOT state a heightened state of accessibility for the unretrieved target relative to when no TOT state is present, they may infer from the presence of a TOT state a greater likelihood of the unretrieved target being of higher frequency.…”

Magnitude comparison extended: How lack of knowledge informs comparative judgments under uncertainty.

“…Likewise, metric accuracy was poor (Mdn OME ϭ 0.71, IQR OME ϭ 0.34). It is also noteworthy that among all pairs of variables in Table 1, recognition times and knowledge ratings are correlated most highly (Mdn r S ϭ Ϫ.57, IQR r S ϭ .29), which should not be surprising as both have been used as measures of familiarity in the past (e.g., Brown & Siegler, 1992;Hertwig et al, 2008).…”

“…Footnote 1); (b) the recognition cue in a given domain has a low validity (Pachur & Hertwig, 2006;Pohl, 2006); and (c) a low retrieval fluency of the recognized item in a given pair signals that the recognition cue might not be predictive in that particular instance (Marewski, Gaissmaier, et al, 2010;Marewski & Schooler, 2011). This last condition presupposes that the assessment of the speed with which the recognized pair member is processed-a potential index of the criterion (Hertwig, Herzog, Schooler, & Reimer, 2008;Schooler & Hertwig, 2005)-is used as a cue to evaluate the applicability of the RH.…”

“…However, this is not true for the preceding recognition stage. Given that retrieval time can be correlated with the criterion (e.g., Hertwig et al, 2008), items with higher criterion values might be retrieved faster, which, in turn, would also affect the overall RT for the comparative judgment . This prediction is also consistent with recent ACT-R implementations of the RH (Marewski & Mehlhorn, 2011;Marewski & Schooler, 2011), which show that the recognition time of the recognized pair member and overall inference time are commonly positively correlated.…”

“…In Study 2, we used a repeated-set procedure, as has been the case in most prior research on the RH (but see Hertwig et al, 2008;Marewski & Schooler, 2011;Volz et al, 2006). That is, participants were presented with all possible pairs created from a relatively small set of items drawn from a larger item population.…”