The unbearable burden of executive load on cognitive reflection: A validation of dual process theory

The unbearable burden of executive load on cognitive reflection: A validation of dual process theory Eric D. Johnson (eric.johnson@ub.edu) Department of Basic Psychology and IR3C, University of Barcelona, Barcelona, Spain Elisabet Tubau (etubau@ub.edu) Department of Basic Psychology and IR3C, University of Barcelona, Barcelona, Spain Wim De Neys (wim.de-neys@parisdescartes.fr) CNRS, LaPsyDE (CNRS Unit 8042), Paris Descartes University, Paris, France Abstract The Cognitive Reflection Test (CRT) is increasingly employed to measure the tendency of an individual to override a prepotent but incorrect response and to subsequently engage in further reflection. This interplay between fast intuitive responding and resource demanding reflection has been offered as a paradigmatic example of dual process theories of thinking. Despite its growing popularity both for dual process theories and as an easily deployed measure of intelligence, the basic assumption that the CRT relies on executive resources remains generally circumstantial. The present study directly tested these dual process assumptions by presenting the standard bat-and-ball problem and a no-conflict control version while manipulating executive resources with a secondary load task. With the no-conflict control problems, accuracy was uniformly at ceiling in no load, low load, and high load conditions. In sharp contrast, in the standard conflict problems accuracy clearly declined with increasingly load. These findings validate dual process assumptions by providing direct causal evidence that correctly resolving the bat-and-ball problem draws on executive resources. Keywords: Cognitive reflection; bat-and-ball problem; Dual process; Executive resources; Reasoning; Decision making; Bias Introduction In the face of difficulty, human reasoners often appear to forego the effortful processing that may be required and opt instead for less demanding intuitive responses (Kahneman, 2011). While many fast and frugal heuristics are no doubt adaptive in complex and reoccurring environments (Gigerenzer, 2007), thinking fast can also lead to quite embarrassingly erroneous responses in less routine settings. Quickly consider the following example: A bat and a ball together cost $1.10. The bat costs $1 more than the ball. How much does the ball cost? Intuitively, the answer “10 cents” quickly springs to mind. In fact, typically around 80% of university students, including those from elite schools such as MIT and Harvard, respond with this intuitive—but incorrect—answer (e.g. Bourgeois-Gironde & Vanderhenst, 2009; Frederick, 2005). If the bat costs $1 more than a 10-cent ball, the bat itself must cost $1.10. Summing up, a $1.10 bat + a $0.10 ball would equal $1.20, not $1.10 as stated in the problem. Does this imply that highly educated young adults think that ‘110 + 10’ = ‘110’? Of course not. Rather, it suggests that even educated reasoners often do not invest the necessary effort to correct their initial intuition, and instead settle for a quickly derived response. The above bat-and-ball problem 1 is extremely popular amongst researchers, and extremely difficult for participants. This popularity and difficulty are highly intertwined. First, high correlations are often observed between performance on this problem and other measures of intelligence or cognitive capacity, making it a tempting short alternative to more cumbersome measures such as standard IQ or working memory tests (Toplak, West, Stanovich, 2011). Second, like many traditional judgment and reasoning tasks arising out the heuristics-and-biases era, the above bat-and-ball problem is a paradigmatic example of a situation whose initially primed response is incorrect and which therefore must be overridden in order to correctly solve the task. This latter point has made the bat-and-ball problem particularly attractive for dual process theories of thinking and reasoning (Kahneman, 2011; Kahneman & Frederick, 2002; Thompson, 2009). According to dual processes theories, human reasoning is characterized by two systems, or processing types, which respond to encountered information in different ways. A first intuitive type thinking (System 1) automatically triggers responses based on prior knowledge/beliefs or via a number of simplifying heuristics, whereas a second more deliberate type (System 2) demands the more controlled utilization of limited executive resources. The bat-and-ball problem, together with two other items, forms part of the Cognitive Reflection Test (CRT; Frederick, 2005). For a recently extended test, see Toplak, West & Stanovich (2013).