Fuzzy-Trace Theory Explains Time Pressure Results

This post is based on a paper:  “Intuition and analytic processes in probabilistic reasoning: The role of time pressure,” authored by Sarah Furlan, Franca Agnoli, and Valerie F. Reyna. Valerie Reyna is, of course, the primary creator of fuzzy-trace theory. Reyna’s papers tend to do a good job of summing up the state of the decision making art and fitting in her ideas.

The authors note that although there are many points of disagreement, theorists generally agree that there are heuristic processes (Type 1) that are fast, automatic, unconscious, and require low effort. Many adult judgment biases are considered a consequence of these fast heuristic responses, also called default responses, because they are the first responses that come to mind. Type 1 processes are a central feature of intuitive thinking, requiring little cognitive effort or control. In contrast, analytic (Type 2) processes are considered slow, conscious, deliberate, and effortful, and they place demands on central working memory resources. Furlan, Agnoli, and Reyna assert that Type 2 processes are thought to be related to individual differences in cognitive capacity and Type 1 processes are thought to be independent of cognitive ability, a position challenged by the research presented in their paper. I was surprised by the given that intuitive abilities were unrelated to overall intelligence and cognitive abilities as set up by typical dual process theories.

According to fuzzy-trace theory, intuitive reasoning arises from qualitative-heuristic processes operating on the gist or essential meaning of a problem. Quantitative-analytic processes are more detail oriented and operate on verbatim representations of a problem. With increasing age, education, and practice in a problem domain, people become increasingly skilled at extracting and processing gist and increasingly reliant on intuition. Thus, according to the others, fuzzy-trace theory places intuition at the apex of development, unlike other dual process theories. Experts are able to grasp the gist of a situation quickly within their domain of expertise, whereas novices employ cognitively demanding analytic processes that manipulate verbatim elements of the problem description. Whereas other dual-process theories attribute much intelligent behavior to skilled analytic reasoning, fuzzy-trace theory emphasizes the role of intuitive reasoning. This paper explores the relative contributions of heuristic and analytic processing to intelligent behavior and how conflicts are resolved.

Previous research has explored this issue. The Cognitive Reflection Test is a strong predictor of performance on rational thinking tasks in which automatic processes applied by default are expected to yield a non-normative result. CRT was interpreted as measuring participants’ readiness to engage Type 2 processes that would override automatic Type 1 processes. Peters concluded that high-numerate participants are more likely to retrieve and use appropriate numerical principles, which is relevant to verbatim or algorithmic processes. They also found that higher numerate participants draw more effective meaning from numbers, which is consistent with intuitive gist-based reasoning.

Imposing time pressure during the performance of a task offers a way to substantially constrain the roles of the two types of processes and assess their relationships to abilities. When a task is self-paced, both types of processes can contribute to performance, but with all other factors equal, time pressure diminishes the opportunity for contributions from slow algorithmic processes. Furlan, Agnoli, and Reyna investigated the role of fast and slow processes in a probabilistic reasoning task consisting of three problems. In each problem, participants were asked to choose between two containers of marbles for a chance to draw a winning marble. The ratios of winning to losing marbles were manipulated, requiring that participants judge which ratio was larger or whether they were the same size. In one problem the ratios were 9/10 and 90/100. The other two of the three problems were  more difficult to solve, but one of the ratios was always 9/10. The other ratio was 85/95 in one problem and 95/105 in the other problem. According to the authors, these ratios were chosen to create comparison problems that would be difficult to solve using algorithmic or verbatim processes.

People who have a lot of experience and practice with numbers may, however, already know that adding a positive constant to the numerator and denominator of a fraction yields a larger number and subtracting a constant yields a smaller number. They could draw on this knowledge to respond quickly. If these problems are solved by fast gist-related processes, then time pressure should not have much effect on accuracy. The focus of this research is on relationships among performance on ratio comparisons, cognitive abilities, numeracy, cognitive reflection, need for cognition, and faith in intuition. These relationships may depend on the levels and range of cognitive and numeric abilities in the sample population. For this reason, the researchers investigated these relationships in three populations expected to have different levels of abilities.

The results of the study were that time pressure had no significant effect on accuracy in this probabilistic reasoning task. Averaging over all three sample populations, high school and two colleges, participants responded correctly on 65% of problems when self paced and 64% when under time pressure. Under time pressure participants were unlikely to compare the ratios computationally and were compelled to rely on intuitive gist processes.

Individual differences associated with rational thought (i.e., intelligence, numeracy, and cognitive reflection) were positively correlated with accuracy in the probabilistic reasoning task under time pressure but not without time pressure, and this result was obtained in all three sample populations.  Participants who employ correct fast gist-based processing should have accuracy relatively unimpaired by time pressure, as we found. Furthermore, the tendency to operate on fuzzy gist (i.e., approximation), as opposed to precise numbers, increases with experience. It appears that good performance under time pressure was the result of fast intuitive processes, and yet performance was correlated with intelligence, cognitive reflection, and objective numeracy.

To some extent, the selected task seems too perfectly adapted to prove the point. The results also fit the basic idea of parallel constraint satisfaction theory.  That is that intuition or the automatic system comes to conclusions, while the conscious analytic system goes out and finds new information. Intuition is more or less linear and can be fooled by non-linear situations. The CRT includes non-linear situations, while the problem explored here is linear and poses no problem for intuition. I am hopeful that Reyna will explore how embodied prediction fits with her fuzzy-trace theory.