Probability matching does not decrease under cognitive load: A preregistered failure to replicate

• Published in: Memory & cognition Vol. 47; no. 3; pp. 511 - 518
• Main Authors: Schulze, Christin, James, Greta, Koehler, Derek J., Newell, Ben R.
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.04.2019; Springer Nature B.V
• Subjects: Accuracy; Adolescent; Adult; Bayesian analysis; Behavioral Science and Psychology; Choice Behavior - physiology; Cognition & reasoning; Cognitive ability; Cognitive load; Cognitive Psychology; College students; Decision making; Earnings; Executive Function - physiology; Experimental psychology; Experiments; Female; Humans; Language; Male; Memory; Memory, Short-Term - physiology; Mental task performance; Probability; Psychology; Psychomotor Performance - physiology; Quantitative psychology; Science; Short term memory; Statistical analysis; Telescopes; Young Adult; Probability matching; back task; Working memory load; Probability maximizing; n-back task
• ISSN: 0090-502X; 1532-5946; 1532-5946
• DOI: 10.3758/s13421-018-0888-3

Does taxing cognitive resources improve people’s choices in repeated binary prediction? Wolford, Newman, Miller, and Wig ( 2004 , Canadian Journal of Experimental Psychology , 58 , 221–228) found that a secondary verbal working memory task, which competed for cognitive resources with a repeated binary choice task, steered participants toward adopting the optimal strategy, namely, probability maximizing . By contrast, under single-task conditions, an inferior strategy prevailed, namely, probability matching . We conducted a preregistered direct replication of Experiment 1 in Wolford et al. ( 2004 ) with a sample of participants more than 5 times larger than the original sample. We did not find a statistically significant effect of cognitive load on strategy selection in repeated binary choice. Moreover, in many cases, Bayesian analyses, which were performed in addition to conventional methods of null hypothesis significance testing, yielded substantial evidence in favor of the absence of cognitive load effects on choice behavior. Thus, we found no reliable support for the claim that taxing cognitive resources leads to improved decision-making in repeated binary prediction.