ABSTRACT
In a manner inconsistent with the laws of probability, the size of a drawn sample affects judgments of probability. For example, people find it more likely to draw a winning ticket out of a bowl with 10 winning and 90 blank tickets than out of a bowl with 1 winning and 9 blank tickets. More generally, holding the objective likelihood of a random event constant, the size of the sample from which a single element (e.g., a lottery ticket) is drawn has been found to affect its subjective likelihood (Denes-Raj & Epstein, 1994; Denes-Raj, Epstein, & Cole, 1995; Kirkpatrick & Epstein, 1992; Miller, Turnbull, & McFarland, 1989; Pelham, Sumarta, & Myaskovsky, 1994). The main question raised in this chapter is whether subjective experience is part of the process mediating this sample size effect on subjective likelihood. Consider Miller and colleagues (1989) “cookie jar” paradigm as another example. These investigators asked their participants to imagine the following scenario: their child prefers chocolate chip cookies over oatmeal cookies. Because they keep both cookie types in one jar, in order to avoid having the oatmeal cookies be left over all the time, they have just implemented the rule that the child draw one cookie at a time from the jar with his or her eyes closed. Upon being told the new rule, the child goes to the kitchen, returning excitedly with a chocolate chip cookie shortly after and claiming he or she drew it according to the new rule. The question to the experimental participants was how suspicious they would be that the child had peeked. Participants were more suspicious when the ratio of chocolate chip cookies to oatmeal cookies in the jar was 1 to 19 rather than 10 to 190. According to these authors, after knowing the outcome, that is, post hoc, a 1 in 19 chance seemed subjectively less likely than a 10 in 190 chance of drawing one chocolate chip cookie.
