Working Paper: CEPR ID: DP15716
Authors: Jonas P. Kaiser; Alexander K. Koch; Julia Nafziger
Abstract: We test in an experiment whether and why non-binding, self-set goals are effective even though individuals can easily revise such goals.Depending on the treatment, subjects either set a goal a few days before working on an online task or right at the start of the task. In the former case, they may or may not be explicitly asked to revise their goal at the start of the task. Consistent with the hypothesis that goals are self-regulation tools, we observe that goals set before the task are larger than goals set at the start of the task. And they are effective:Subjects work more when a goal was set a few days before the task than when it was set at the start of the task. Importantly, these results arise even though subjects revise their initial goals. They do so no matter whether goal revision is made explicit or not -- suggesting that goal revision is an important factor for goal non-achievement.
Keywords: self-control; present-biased preferences; goals; goal revision; commitment devices; real effort; online experiment
JEL Codes: D03; D81; D91
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Initial goal serves as a reference point (L21) | Influences subsequent performance (D29) |
| Revised goals do not significantly differ from goals set immediately before the task (L21) | Lesser impact of initial goal on final performance outcomes (D29) |
| Goals set five days before the task (L21) | Higher levels of effort (D29) |
| Goals set immediately before the task (L21) | Lower levels of effort (D29) |
| Higher initial goal (L21) | Greater effort (D29) |