THE DENOMINATOR NEGLECT IN DECISION-MAKING

Abstract

IntroductionIf you want to buy a lottery ticket, would you choose from a pile in which one out of ten is a winner, or would you choose from a pile in which nine out of a hundred are winners? If you prefer to choose from the second pile, you probably pay more attention to the numerator than to the denominator. This phenomenon is not rare; it is called denominator neglect, or ratio-bias phenomenon (in the present study I use the term denominator neglect). Denominator neglect is focusing on relative frequencies of numerators (Reyna & Brainerd, 2008), paying attention to numerators (number of times an event has happened) and inadequate attention to denominators (overall number of opportunities for an event to happen) (e.g., Garcia-Retamero, Galesic, & Gigerenzer, 2010; Okan, Garcia-Retamero, Cokely, & Maldonado, 2011; Passerini, Macchi, & Bagassi, 2012), for instance, whether nine winning lottery tickets is few or many depends on the overall number of lottery tickets.According to Alonso and Fernandez-Berrocal (2003) one reason of denominator neglect is that people better understand whole numbers than ratios (people automatically code frequencies) and second reason is that people better understand low numbers than large numbers. Low numbers are more concrete than large numbers and, furthermore, people have a social circle of around 100-200 people, so understanding probabilities with larger denominator may be difficult (Garcia-Retamero & Galesic, 2011). For example, in terms of the occurrence of side effects, people have problems imagining the incidence of 1-in-1000, 1-in-10000, etc. Understanding numerical information seems to be important because it can lead to accurate judgement and decision making, especially in medical treatment when people focus more on treated and non-treated people who die and they do not consider the overall number of patients. In another study, Alonso and Fernandez-Berrocal (2003) found that people made irrational decisions; they chose options with less objective probability due to higher numerator than in the alternative option.Pacini and Epstein (1999) examined some determinants of denominator neglect and they found that participants scoring low in rational thinking made more suboptimal responses in Jellybean tasks. In another study, Alonso and Fernandez-Berrocal (2003) found that people who chose the suboptimal options had lower score in rationality measured by Need for Cognition Scale, and Toplak, West, and Stanovich (2013) found that cognitive reflection predicted resistance to denominator neglect.In previous studies (Alonso & Fernandez-Berrocal, 2003; Garcia-Retamero & Galesic, 2011; Pacini & Epstein, 1999; Toplak, West, & Stanovich, 2013) denominator neglect was described in terms of its frequencies and its predictors. In the present study, the main aim was to replicate the above-mentioned findings: a) to examine the degree of denominator neglect and b) to verify possible relations of denominator neglect with cognitive ability, motive for rational thinking integration, preference of cognitive style, and cognitive reflection. The present study enriched denominator neglect research by using the Jellybean task with equal versus unequal and small versus large probabilities, next, by using self-reported scales as well as performance task to measure cognitive abilities, and by motivation of participants for optimal (maximal) performance.In Study 1 simple Jellybean task was used - 3 tasks with unequal low probabilities and relationship between denominator neglect and cognitive ability and motive for rational integration was verified. In Study 2 more complex Jellybean task was used (14 items) and the effect of equality of probability (equal vs. unequal) and degree of probability (low vs. high) were examined, as well as relationships between denominator neglect and motive for rational integration and thinking dispositions. In contrast to Study 1, participants were incentivized by extra credits dependent on the accuracy of their choices in the Jellybean task. …