The effect of “should” and “would” instructions on delay discounting of rewards for self and others

Abstract Would you prefer $50 now or $100 in 6 months? What if you made this decision for someone else—would you be more impulsive or more self‐controlled? Some studies suggest we are more impulsive when deciding for ourselves, whereas others suggest the reverse. This might be because some researchers ask participants what they would do whereas others ask what they should do. We investigated the impact of should/would decision type on delay discounting rate in choices for the self and for another person. We also examined the effect of condition order. In Experiment 1 (using a student sample), discounting rates were affected by the combination of decision frame and condition order. Decision frame had a bigger effect on choices in the second condition, perhaps because instructions became clearer when they could be contrasted with the previous set. Experiment 2 (using a Mechanical Turk sample) investigated this possibility by including all possible frames at the beginning of the session; this produced a more consistent would/should difference for choices for the self, but an order effect remained. Experiment 3 isolated task order by having participants complete the same choice task twice. Decisions were significantly more self‐controlled for the second iteration. Together, these results suggest that people are more self‐controlled when making should decisions and (less consistently) decisions for others and that having recently made delay‐amount trade‐off decisions also promotes self‐control. Rates of unsystematic data were unexpectedly high in Experiment 2, particularly among nonmaster Turk workers and those located in India.