Quantifying the subjective cost of self-control in humans

Abstract

Since Odysseus committed to resisting the Sirens, mechanisms to limit self-control failure have been a central feature of human behavior. Psychologists have long argued that the use of self-control is an effortful process and, more recently, that its failure arises when the cognitive costs of self-control outweigh its perceived benefits. In a similar way, economists have argued that sophisticated choosers can adopt "precommitment strategies" that tie the hands of their future selves in order to reduce these costs. Yet, we still lack an empirical tool to quantify and demonstrate the cost of self-control. Here, we develop and validate an economic decision-making task to quantify the subjective cost of self-control by determining the monetary cost a person is willing to incur in order to eliminate the need for self-control. We find that humans will pay to avoid having to exert self-control in a way that scales with increasing levels of temptation and that these costs appear to be modulated both by motivational incentives and stress exposure. Our psychophysical approach allows us to index moment-to-moment self-control costs at the within-subject level, validating important theoretical work across multiple disciplines and opening avenues of self-control research in healthy and clinical populations.

Keywords: decision-making; motivation; precommitment; self-control; stress.

Fig. 1.

Illustration of the self-control decision task. Participants reported their willingness to pay to avoid a tempting food reward both before the food was present (Top) and, periodically over a 30-min period, with added direct exposure to the food (Bottom).

Fig. 2.

Study 1. (A) Bids over time for control group (bids 1–10). (B) Proportion of subjects in Study 1 that consumed the tempting food during the study. Error bars denote SEM.

Fig. 3.

Study 2. (A) Bids to avoid exposure to the tempting food over time in participants, for which a $15 monetary loss was imposed for consuming the food (depicted in green; Incentive group) and for those where no monetary loss was imposed (depicted in gray; Control group). (B) Proportion of subjects in Study 2 that consumed the tempting food during the study.

Fig. 4.

Study 3. (A) Bids to avoid exposure to the tempting food over time for participants that underwent a physiological stress manipulation (depicted in red; Stress group) and for non-stressed participants (depicted in gray; Control group). (B) Proportion of subjects that consumed the tempting food during the study.

Fig. 5.

Study 4. (A) Bids to avoid exposure to the tempting food over time in stressed participants for which a $15 monetary loss was imposed for consuming the food (depicted in orange; Incentive+Stress group) and for those with the same penalty imposed but no stress induction (depicted in green; Incentive group; Study 2). The control group from Study 1 (no incentive or stress) is depicted in gray for reference. (B) Proportion of subjects from Study 4 that consumed the tempting food during the study.

Fig. 6.

Mean bid for subjects who demonstrated self-control “failures” (23%, depicted in purple) and those who did not (77%; depicted in blue) collapsed across Study 1 (control group) and 3 (stress group). Those participants who consumed the tempting food during the study revealed a higher willingness to pay to avoid control. No participants from Study 2 or 4 ($15 penalty groups) consumed the tempting food. Error bars denote SEM.

Fig. 7.

Individual difference correlations. (A) Perceived stress was positively correlated with average bidding behavior across participants. (B) Length of diet was negatively correlated with average bidding behavior across participants. *denotes Spearman ranked correlation.

Fig. 8.

Willingness to pay to avoid foods that varied in temptation level and amount of time required to spend with food. Bids scaled with increasing time with and temptation level of the food. Error bars depict SEM.