Recent investigations of adolescents’ beliefs about risk have led tosurprisingly optimistic conclusions: Teens’ self estimates of theirlikelihood of experiencing various life events not only correlate sensibly withrelevant risk factors (Fischhoff et al., 2000), but they also significantlypredict later experiencing the events (Bruine de Bruin et al., 2007). Using thesame dataset examined in previous investigations, the present study extendedthese analyses by comparing the predictive value of self estimates of risk tothat of traditional risk factors for each outcome. The analyses focused on theprediction of pregnancy, criminal arrest, and school enrollment. Three findingsemerged. First, traditional risk factor information tended to out-predict selfassessments of risk, even when the risk factors included crude, potentiallyunreliable measures (e.g., a simple tally of self-reported criminal history) andwhen the risk factors were aggregated in a nonoptimal way (i.e., unitweighting). Second, despite the previously reported correlations between selfestimates and outcomes, perceived invulnerability was a problem among the youth:Over half of the teens who became pregnant, half of those who were not enrolledin school, and nearly a third of those who were arrested had, one year earlier,indicated a 0% chance of experiencing these outcomes. Finally, adding selfestimates of risk to the other risk factor information produced only small gainsin predictive accuracy. These analyses point to the need for greater educationabout the situations and behaviors that lead to negative outcomes.

Subjects donate individually (control group) or in pairs (treatment group). Thosein pairs reveal their donation decision to each other. Average donations in thetreatment group are significantly higher than in the control group. Pairedsubjects have the opportunity to revise their donation decision afterdiscussion. Pair members shift toward each others’ initial decisions.Subjects are happier with their decision when their donations are larger, butthose in pairs are less happy, controlling for amount donated. These findingssuggest reluctant altruism due to peer pressure in charitable giving.

Decision makers often save money for a specific goal by forgoing discretionaryconsumption and instead putting the money toward the savings goal. Wehypothesized that reference points can be exploited to enhance this type ofsaving. In two hypothetical scenario studies, subjects made judgments of theirlikelihood to forgo a small expenditure in order to put the money toward thesavings goal. In Experiment 1, judgments were higher if the savings goal waspresented as composed of weekly subgoals (e.g., save $60 per week to buya $180 iPod). Experiment 2 replicated this finding and demonstrated thatthe subgoal manipulation increased judgments of likelihood to save money onlywhen the money saved from the foregone consumption would allow the decisionmaker to meet the weekly subgoal exactly (not under or overshoot it). Theseresults suggest a reference point mechanism and point to ways that behavioraldecision research can be harnessed to improve economic behaviors.

We report a mathematical result that casts doubt on the possibility ofrecalibration of probabilities using calibration curves. We then discuss how tointerpret this result in the light of behavioral research.

Physicians expect a treatment to be more effective when its clinical outcomes aredescribed as relative rather than as absolute risk reductions. We examinedwhether effects of presentation method (relative vs. absolute risk reduction)remain when physicians are provided the baseline risk information, a vital pieceof statistical information omitted in previous studies. Using a between-subjectsdesign, ninety five physicians were presented the risk reduction associated witha fictitious treatment for hypertension either as an absolute risk reduction oras a relative risk reduction, with or without including baseline riskinformation. Physicians reported that the treatment would be more effective andthat they would be more willing to prescribe it when its risk reduction waspresented to them in relative rather than in absolute terms. The relative riskreduction was perceived as more effective than absolute risk reduction even whenthe baseline risk information was explicitly reported. We recommend thatinformation about absolute risk reduction be made available to physicians in thereporting of clinical outcomes. Moreover, health professionals should becognizant of the potential biasing effects of risk information presented inrelative risk terms.

Previous research on the role of recognition in decision-making in inferentialchoice has focussed on the Recognition Heuristic (RH), which proposes that insituations where recognition is predictive of a decision criterion, recognizedobjects will be chosen over unrecognized ones, regardless of any other availablerelevant information. In the current study we examine the role of recognition inpreferential choice, in which subjects had to choose one of a pair of consumerobjects that were presented with quality ratings (positive, neutral, andnegative). The results showed that subjects’ choices were largely basedon recognition, as the famous brand was preferred even when additional starratings rendered it as less attractive. However, the additional information didaffect the proportion of chosen famous items, in particular in the cases whenstar ratings for the recognised brand were negative. This condition alsoresulted in longer response times compared to neutral and positive conditions.Thus, the current data do not point to a simple compensatory mechanism inpreferential choice: although choice is affected by additional information, itseems that recognition is employed as an initial important first step in thedecision-making process.

In this paper we test two hypotheses regarding magical thinking about theperceived likelihood of future events. The first is that people believe thatthose who “tempt fate” by failing to take necessary precautionsare more likely to suffer negative outcomes. The second is the“protection effect”, where reminding people of precautions theyhave taken leads them to see related risks as less likely. To this end, wedescribe the results from three attempted direct replications of a protectioneffect experiment reported in Tykocinski (2008) and two replications of atempting fate experiment reported in Risen and Gilovich (2008) in which we add atest of the protection effect. We did not replicate the protection effect butdid replicate the tempting fate effect.

Birnbaum (2011, 2012) questioned the iid (independent and identicallydistributed) sampling assumptions used by state-of-the-art statistical tests inRegenwetter, Dana and Davis-Stober’s (2010, 2011) analysis of the“linear order model”. Birnbaum (2012) cited, but did not use, atest of iid by Smith and Batchelder (2008) with analytically known properties.Instead, he created two new test statistics with unknown samplingdistributions.

Our rebuttal has five components: 1) We demonstrate that the Regenwetter et al.data pass Smith and Batchelder’s test of iid with flying colors. 2) Weprovide evidence from Monte Carlo simulations that Birnbaum’s (2012)proposed tests have unknown Type-I error rates, which depend on the actualchoice probabilities and on how data are coded as well as on the null hypothesisof iid sampling. 3) Birnbaum analyzed only a third of Regenwetter etal.’s data. We show that his two new tests fail to replicate on the othertwo-thirds of the data, within participants. 4) Birnbaum selectively picked dataof one respondent to suggest that choice probabilities may have changed partwayinto the experiment. Such nonstationarity could potentially cause a seeminglygood fit to be a Type-II error. We show that the linear order model fits equallywell if we allow for warm-up effects. 5) Using hypothetical data, Birnbaum(2012) claimed to show that “true-and-error” models for binarypattern probabilities overcome the alleged short-comings of Regenwetter etal.’s approach. We disprove this claim on the same data.

Prior studies have found that subjects prefer an improving sequence of incomeover a constant sequence, even if the constant sequence offers a largerpresent-discounted value. However, little is known about how these preferencesvary with the size of the wage payments. In each of four studies, we find apositive relationship between the preference for increasing payments and thesize of the payments. We find no evidence that our measure of the decreasingmarginal utility of money is associated with this relationship. Additionally, wefind weak evidence in support of a theoretical prediction that the differencebetween the preference for increasing wage payments and the preference forincreasing nonwage payments will be largest for intermediate amounts. We do notfind a relationship between the preference for increasing payments and thepreference for improving nonmonetary sequences. Finally, the relationshipbetween the preference for increasing payments and the size of the payments doesnot appear to be sensitive to the precise specification of the increases.