Abstract
Can others detect the emotional consequences of our personal choices? Here, we investigate whether third-party observers can detect the emotional benefits of two factors shown to influence self-reported happiness: the speed with which people make decisions and the generosity of spending choices. Participants were randomly assigned to purchase a goody bag either for themselves or for a sick child, and to choose the contents of this goody bag either as quickly as possible, or by taking as much time as needed. Then, participants reported their current emotional state and were rated for happiness by a research assistant blind to their spending condition. Analyses revealed that purchasing a gift for someone else not only improved participants’ self-reported mood, but that observers could detect these affective differences as well. Observers also rated participants who made their spending decision more quickly as happier, although participants did not report these emotional differences.
Acknowledgment
This work was supported by a grant from the Canadian Institute for Advanced Research (CIFAR) to the first author.
Notes
1. Although we would have liked to have multiple research assistants present to calculate reliability, we feared that this would arouse suspicion for a computer-based study. However, a single informant’s happiness ratings appear to provide valid information because coder ratings were positively correlated with self-reported emotion, even though coders were blind to condition. Moreover, because previous research has shown that multiple coders using a similar coding system demonstrate high reliability (see Aknin et al., Citation2012), adding additional coders may have added little value to the accuracy of these ratings. Finally, the use of one coder provides a conservative test of our research question. If single coder ratings present noisier data, they should be less likely to differ by condition; our results demonstrate the opposite.
2. The main effect of spending target remained relatively unchanged if we controlled for state happiness only (F(1,112) = 4.11, p < 0.05), trait happiness only (F(1,112) = 3.92, p = 0.05), or removed any measure of baseline happiness (F(1,112) = 3.49, p = 0.06).
3. As with the self-report affect ratings, this effect remained significant when we controlled for state happiness only (F(1,113) = 5.90, p < 0.02), trait happiness only (F(1,113) = 5.92, p < 0.02), or removed any measure of baseline happiness (F(1,113) = 5.91, p < 0.02).