Academic journal article Memory & Cognition

Decision Biases in Intertemporal Choice and Choice under Uncertainty: Testing a Common Account

Academic journal article Memory & Cognition

Decision Biases in Intertemporal Choice and Choice under Uncertainty: Testing a Common Account

Article excerpt

Two experiments were performed to test a psychophysical account of parallels between biases in risky choice and intertemporal choice. Experiment 1 demonstrated the common difference effect in intertemporal choice and the common ratio effect in risky choice. As was predicted, these two biases were uncorrelated with each other, although each was correlated across monetary/health domains. This result is consistent with the supposition that these two biases result from psychophysical properties of two different dimensions (time and probability, respectively). Experiment 2 examined the magnitude effect in intertemporal choice and the peanuts effect in risky choice. These two biases were correlated with each other but were uncorrelated across monetary/health domains. This result is consistent with the supposition that these two biases result from psychophysical properties of the same dimension (utility of money or health).

Some decisions involve uncertain outcomes. For example, selecting an entrée at a restaurant entails uncertainty about how delicious each menu option would be. Other decisions involve delayed outcomes. For example, initiating an exercise program entails a delay until the positive results of exercise are observed. Both decision making under uncertainty and intertemporal choice have been the topic of much research. In the present article, we analyze potential parallels between these two types of choices.

Decision making under uncertainty can be characterized by risk preferences. A preference for a lottery over its expected value (EV) (e.g., preferring a 50% chance of $ 10 over $5 for sure) is considered risk seeking, whereas a preference for the expected value over the lottery (e.g., preferring the $5 for sure) is known as risk aversion. Intertemporal choice can be characterized by time preferences. A preference for a positive outcome now rather than for an equivalent outcome later (e.g., preferring $ 10 now to $10 in 1 month) is called a positive time preference, and the extent of the time preference can be quantified as a discount rate, or the percentage of increase in the magnitude of the payout needed to offset a delay. For example, a 20% monthly discount rate would mean that $ 10 now is equally preferable to $12 in 1 month (Frederick, Loewenstein, & O'Donoghue, 2003).

There are apparent parallels between choice under uncertainty and intertemporal choice-that is, seemingly analogous phenomena that could be explained if risk preferences and time preferences are driven by common factors (see Green & Myerson, 2004). A number of researchers have suggested that time and delay are parallel (e.g., Keren & Roelofsma, 1995; Rachlin, Logue, Gibbon, & Frankel, 1986) or have analogous effects on choice (Chapman, 1997; Gafni & Torrance, 1984; Myerson, Green, Hanson, Holt, & Estle, 2003).

The study of both choice under uncertainty and intertemporal choice has centered around the analysis of decision biases, or deviations from normative theory. Risky choice biases are deviations from expected utility theory (Neumann & Morgenstern, 1953). Intertemporal choice biases are deviations from discounted utility theory (Koopsman, 1960). Prelec and Loewenstein (1991 ; Loewenstein & Prelec, 1992) noted that some risky choice biases parallel intertemporal choice biases.

The purpose of the present experiments was to provide an empirical test of a particular account of the apparent parallels between biases in risky choice and biases in intertemporal choice. That is, we tested the idea that biases in the two domains that appear similar are the result of the same underlying mechanism. We first will describe two pairs of biases; we then will present a psychophysical account for parallels between the biases and, finally, an empirical test devised to evaluate this account.

Pairs of Parallel Biases

Common difference and common ratio effects. The common difference effect is a bias in intertemporal choice in which adding a constant delay to both choice options shifts preference from the smaller, sooner outcome to the larger, later outcome (Christensen-Szalanski, 1984; Green, Fristoe, & Myerson, 1994; Kirby & Herrnstein, 1995). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.