Heuristics and biases: interactions among numeracy, ability, and reflectiveness predict normative responding.

Paul A Klaczynski

doi:10.3389/fpsyg.2014.00665

Abstract

In Stanovich's (2009a, 2011) dual-process theory, analytic processing occurs in the algorithmic and reflective minds. Thinking dispositions, indexes of reflective mind functioning, are believed to regulate operations at the algorithmic level, indexed by general cognitive ability. General limitations at the algorithmic level impose constraints on, and affect the adequacy of, specific strategies and abilities (e.g., numeracy). In a study of 216 undergraduates, the hypothesis that thinking dispositions and general ability moderate the relationship between numeracy (understanding of mathematical concepts and attention to numerical information) and normative responses on probabilistic heuristics and biases (HB) problems was tested. Although all three individual difference measures predicted normative responses, the numeracy-normative response association depended on thinking dispositions and general ability. Specifically, numeracy directly affected normative responding only at relatively high levels of thinking dispositions and general ability. At low levels of thinking dispositions, neither general ability nor numeric skills related to normative responses. Discussion focuses on the consistency of these findings with the hypothesis that the implementation of specific skills is constrained by limitations at both the reflective level and the algorithmic level, methodological limitations that prohibit definitive conclusions, and alternative explanations.

Highlights

When the standards against which they are evaluated are traditional norms, performance on heuristics and biases (HB) tasks is often poor (Kahneman et al, 1982; Reyna and Brainerd, 1995; Stanovich, 1999)
CONFLICT AND NO-CONFLICT PROBLEMS To examine whether normative responses were more frequent on N-CN problems than on CN problems, a multivariate analysis of variance, with normative scores on the four tasks as dependent variables and problem type (CN or N-CN) as a within-subjects variable, was conducted
The analyses were intended to determine whether noconflict scores on the four tasks were related to each other, conflict scores, and the individual difference measures (i.e., thinking dispositions (TD), general ability (GA), and numeracy)

Summary

Introduction

When the standards against which they are evaluated are traditional norms, performance on heuristics and biases (HB) tasks is often poor (Kahneman et al, 1982; Reyna and Brainerd, 1995; Stanovich, 1999). Underlying most views of the “normative/descriptive gap” (see Baron, 2008) is the assumption that rational thinking is “bounded” by information processing limitations (e.g., working memory, processing speed). In accord with this view, measured intelligence, generally assumed to index these processing limitations, relates positively to normative responses on several HB tasks. Thinking dispositions—relatively malleable cognitive styles, beliefs, intellectual values, and motivations to manage cognitive resources (e.g., expending effort, guarding against impulsivity, valuing deliberate thinking, openness to using different strategies)—often account for variance in performance independently from general ability (Stanovich and West, 1998, 2000; Klaczynski and Lavallee, 2005; West et al, 2008; Toplak et al, 2011)

Methods

Results

Conclusion