How Much Do Incidental Values Affect the Judgment of Time?

In a recent article, Ungemach, Stewart, and Reimers (2011) showed that incidental values from the environment influence people’s decisions about money, risk, and time. They suggested that people compute the subjective magnitude of a target quantity by sampling items from memory and comparing the target item with this sample to establish its rank in the set (Stewart, Chater, & Brown, 2006). Items encountered more recently are more likely to enter the memory sample and thus influence judgment, even if the items are incidental to the judgment at hand. I focused on Ungemach et al.’s (2011) demonstration of the influence of incidental values on temporal judgments. In their Study 3, Ungemach et al. had participants think about how they would spend their next birthday before completing a delay-discounting task in which they chose between a smaller amount of money to be awarded in 3 months and a larger amount to be awarded in 9 months. Participants whose birthdays fell between these time points showed steeper discounting (i.e., they regarded the reward in 9 months as less valuable than did other participants), an outcome consistent with the idea that the incidental temporal value (time until next birthday) entered the memory sample and, in a rank-based evaluation, made the difference between 3 months and 9 months seem more pronounced. I share Ungemach et al.’s (2011) theoretical perspective but was unable to replicate their significant effect in the 12 studies I conducted. In Studies 1 through 9, I had participants rank or rate possible birthday plans before completing various judgment tasks. Participants in the inside group had birthdays that fell between 3 and 9 months in the future; participants in the outside group had birthdays that fell either less than 3 months or more than 9 months in the future. In Study 1, participants imagined receiving a $75 voucher that would become active 3 months in the future and indicated how much extra money they would demand if they had to wait until 9 months in the future before using it (Zauberman, Kim, Malkoc, & Bettman, 2009). Study 2 was similar but used a pen-and-paper task. In Study 3, the point at which participants became indifferent between receiving $100 in 9 months and a smaller amount in 3 months was established by a method of limits (Sun & Li, 2010). In Study 4, participants were given the choice between receiving $100 in 9 months and a smaller, unspecified amount in 3 months and were asked how much the smaller amount needed to be before they would find both options equally attractive. Study 5 addressed the same question using a stepwise decision procedure identical to that used by Ungemach et al. Study 6 replicated Study 5 in the lab. Study 7 was similar to Study 6 but used wording taken directly from Ungemach et al. In Study 8, participants simply rated their perception of the time difference between events 3 months and 9 months in the future on a scale from 1 (very small) to 20 (very large). Study 9 was identical to Study 8, except that participants overtly indicated the timing of their next birthday relative to events occurring 3 months and 9 months in the future before rating their perception of the time difference between events 3 months and 9 months away. Studies 10 and 11 used a different type of incidental time value: Participants read a news item about an event occurring in 1 week (inside group) or 6 months (outside group) before choosing between a smaller amount to be awarded immediately and a larger amount to be awarded in 1 month. (See the Supplemental Material available online for methodological details of all the current studies.) The results are summarized in Table 1. None are significant. The procedure and results of Studies 5 through 7 most closely mimic those of Ungemach et al. (2011), but the effect is much smaller and shows no sign of generalizing to other measures of temporal judgment and choice or to other kinds of incidental time value. A random-effects meta-analysis (including Ungemach et al.’s data) yielded an overall effect size of –0.003, with a 95% confidence interval that included zero (−0.112 to 0.106; Viechtbauer, 2010). Applying the Bayesian meta-analytic approach of Rouder and Morey (2011) suggested that the combined evidence favors the null hypothesis by a factor of 28.7 to 1 (see Matthews, 2011). To check the validity of my studies, I conducted an additional study in which participants rated the attractiveness of