By situating the concept of “effort” in an adaptive context, Kurzban et al.'s target article represents a major advance toward a scientifically mature conception of self-control. We agree that perseverance depends less on reservoirs of a depletable resource than on favourable cost-benefit computations. In what follows we will show how this approach can illuminate recent findings concerning the relationship between religion and self-control. In doing so we highlight a central currency of cost-benefit computation – the currency of social approval or reputation. However, we also raise doubts about the claim that the phenomenology of effort is a signal that resources are better deployed elsewhere.

The notion that God can replenish one's “strength” to face hardships and resist temptations is common in scripture (e.g., Isaiah 40:31; Peter 5:10) and in popular culture (e.g., Clapton Reference Clapton1974). Consistent with this idea, Rounding et al. (Reference Rounding, Lee, Jacobson and Ji2012) suggest that “invoking religious beliefs may provide important psychological ‘nutrients’… to restore self-control resources … after their depletion” (pp. 640–41). These authors report that participants primed with religion endured more discomfort, delayed gratification longer, persisted longer in an “ego-depletion” test, and performed better on a Stroop task than did control participants. Although Rounding et al. (see also Baumeister et al. Reference Baumeister, Bratslavsky, Muraven and Tice1998; McCullough & Willoughby Reference McCullough and Willoughby2009; Preston et al. Reference Preston, Ritter and Hernandez2010) favour a resource-depletion explanation for these results, the cost-benefit approach of Kurzban et al. affords a more compelling interpretation.

Imagine a man who, approached by a beggar requesting spare change, demurs. He may be unable to donate, having no change on him, or he may be unmotivated to donate. Later, another beggar approaches the man as he passes a church. This time the man digs in his pocket and hands the beggar some coins. Whatever effect the religious context has had, it has not magicked money into the man's pocket, strengthening his financial reserves; more likely it has increased the salience of certain incentives (e.g., the approval of supernatural agents), giving him a reason to donate money he already possessed. Similarly, we contend that the effects Rounding et al. report reflect changes in the implicit incentive structure of relevant tasks, giving participants more reason to wait, persist, and endure – not greater reserves of the ability to do so.

Rounding and colleagues' delayed-gratification study particularly illustrates why we favour a cost-benefit explanation for such priming effects. After priming, participants chose between returning the next day to collect a $5 honorarium, or returning in seven days to collect $6. Participants primed with religion more frequently chose to wait. Rounding and colleagues' contention that religious priming replenished self-control resources suggests that control participants were unable (rather than unmotivated) to wait, implying that participants who waited discounted the value of the delayed $6 less than those who did not wait. However, applying Mazur's (Reference Mazur, Commons, Mazur, Commons and Rachlin1987) discounting formula to the choice indicates that participants who waited must have had miniscule temporal discounting rates (k ≤ .0286). Such rates are generally observed in studies involving larger dollar amounts (Green et al. Reference Green, Myerson and McFadden1997); small amounts (<$10) generate much higher discounting (k = .132; Harrison & McKay Reference Harrison and McKay2012). If religious primes reduce discounting to approximately 20% of typically observed rates, this effect should be easily detected. We tested this possibility in a follow-up study (Harrison & McKay Reference Harrison and McKay2013), but reduced the possible influence of social desirability.

In Rounding and colleagues' study, it was obvious that electing to wait would make participants appear more patient. We used a discounting measure that was opaque in this respect – that is, it was difficult for participants to discern the socially desirable response. We randomly assigned 42 participants to a neutral or a religious prime group; participants completed temporal discounting tasks prior to and immediately following the priming task. Although experimentally induced changes to discounting rates have been detected using similar research designs, in our study religious priming had no effect on subsequent discounting.

So why would primed participants in Rounding et al. (Reference Rounding, Lee, Jacobson and Ji2012) have elected to wait? We agree with Norenzayan and colleagues (Gervais & Norenzayan Reference Gervais and Norenzayan2012; Shariff & Norenzayan Reference Shariff and Norenzayan2007) that the religious primes probably activated the notion of surveillance (supernatural or otherwise), triggering cognitions that regulate behaviour in the service of reputation management (Haley & Fessler Reference Haley and Fessler2005). As a consequence, primed participants were implicitly motivated to signal their patience and persistence to potential observers. In Rounding et al.'s delayed-gratification study the relevant signal was the decision to wait the extra six days – their other studies involved other signals.

What is not clear is whether, in altering subjective incentives to exert “self-control,” religious priming actually alters the experience of effort. Kurzban et al. argue that the phenomenology of effort reflects the opportunity cost of continuing the current course of action; so if perceived incentives to persist with current tasks are increased, those tasks should feel less aversive. However, the recently described “martyrdom effect” suggests that in certain contexts people will deliberately seek out effortful experiences (Olivola & Shafir Reference Olivola and Shafir2013). Such situations (e.g., self-imposed penance) may simply reflect cultural distortions of biological function – exceptions that prove the “effort as deterrent” rule. However, there are other reasons to doubt that the feeling of effort constitutes a signal to pursue alternative goals (Hockey Reference Hockey and Ackerman2011). For one thing, this claim entails the prediction that if there were absolutely no other useful tasks one could engage in (imagine a person in solitary confinement whose only option is to figure out how to escape), the current course could be pursued indefinitely without fatigue. However, the necessity of sleep suggests that resources cannot be endlessly deployed.

This issue notwithstanding, we find Kurzban et al.'s approach extremely compelling, and view their cost-benefit analysis as providing a parsimonious explanation for recent findings regarding the effect of religion on self-control. With apologies to Eric Clapton, our view is that religion affects not the strength “to carry on” but the expected utility of doing so.