Linking process and measurement models of recognition-based decisions.

Abstract

When making inferences about pairs of objects, one of which is recognized and the other is not, the recognition heuristic states that participants choose the recognized object in a noncompensatory way without considering any further knowledge. In contrast, information-integration theories such as parallel constraint satisfaction (PCS) assume that recognition is merely one of many cues that is integrated with further knowledge in a compensatory way. To test both process models against each other without manipulating recognition or further knowledge, we include response times into the r-model, a popular multinomial processing tree model for memory-based decisions. Essentially, this response-time-extended r-model allows to test a crucial prediction of PCS, namely, that the integration of recognition-congruent knowledge leads to faster decisions compared to the consideration of recognition only-even though more information is processed. In contrast, decisions due to recognition-heuristic use are predicted to be faster than decisions affected by any further knowledge. Using the classical German-cities example, simulations show that the novel measurement model discriminates between both process models based on choices, decision times, and recognition judgments only. In a reanalysis of 29 data sets including more than 400,000 individual trials, noncompensatory choices of the recognized option were estimated to be slower than choices due to recognition-congruent knowledge. This corroborates the parallel information-integration account of memory-based decisions, according to which decisions become faster when the coherence of the available information increases. (PsycINFO Database Record

Some Metatheoretical Reflections on Adaptive Decision Making and the Strategy Selection Problem

“…Other researchers might respond that studies such as the aforementioned ones by Glöckner and Bröder (, )—as well as others (e.g., Bröder & Eichler, ) —do not fall in the scope of the recognition heuristic, for instance, by examining the heuristic in inferences from givens as opposed to inferences from memory or by providing additional cue knowledge about unrecognized objects (see Oliver Vitouch's essay in Marewski et al, , for an overview of the history of specification and scope of the recognition heuristic; see also Gigerenzer & Goldstein, , Marewski & Mehlhorn, , Pohl, ). However, no such arguments may be raised against Heck and Erdfelder's () re‐analyses of existing data on inferences from memory. And there is broad consensus that the wealth of studies triggered by the recognition heuristic controversy has led to inherently interesting insights on how recognition influences decision making, both within and outside the domains that fall into the scope of the recognition heuristic.…”

“…One way to solve this problem is to take into account further dependent measures. A comprehensive meta‐analysis of studies in the standard task using natural semantic knowledge took into account not only decisions but also response times (Heck & Erdfelder, ). For the large majority of participants, a single‐mechanism model seemed to account for the data better than the recognition heuristic: specifically, responses based on mere recognition were slower than those using recognition‐compatible knowledge, but faster than those with incompatible knowledge.…”

“…For instance, much past research on the recognition heuristic—including our own work (e.g., Pachur, Bröder, & Marewski, )—might not take into account sufficiently the workings of the cognitive capacities—most notably memory—key to the recognition and other memory‐based heuristics (for exceptions, see e.g., Castela & Erdfelder, ; Erdfelder, Küpper‐Tetzel, & Mattern, ; Heck & Erdfelder, ; Schooler & Hertwig, ; Pachur, ). Independently of what views one entertains on the empirical evidence for different models of recognition‐based inference, the mere fact that recognition and further knowledge are confounded is, on its own, a property of the world, and one that is likely picked up by human memory, and one that can, hence, shape the ways in which people make inferences from recognition and other information with different memory‐based decision mechanisms (be they labeled “single” or “multi”).…”

“…All of the model predictions are made from environmental data, namely, the frequency patterns of objects in the world. Such and more recent ACT‐R modeling work for different decision mechanisms (e.g., Link, Marewski, & Schooler, ) suggests that it is insightful to focus research more on modeling how those environmental regularities and other elements of environments shape the workings of the cognitive system (Anderson & Schooler, ; Bhatia, ; Heck & Erdfelder, ; Landauer & Dumais, ) and to conduct experimental research that goes beyond “eliminating” environmental “confounds” in artificial experimental settings.…”

“…J. N. M.: The following is a comment that is not directed towards Heck and Erdfelder's () work (which represents the most exciting paper on the recognition heuristic I have seen in recent years! ), but that aims to stress a point that has previously been sometimes overlooked: The recognition heuristic does not make, on its own, any predictions about response times, making such predictions requires a series of auxiliary assumptions, and here, different assumption can lead to different response time predictions (Marewski & Mehlhorn, ).…”