Breaking habits
How to cite
Vollaard, Ben; Soest, Daan van. 2019. "Breaking habits." AEA RCT Registry. AEARCTR-0003822. https://doi.org/10.1257/rct.3822-1.0
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
We provide rare evidence for how a temporary incentive can kick-start an improved habit. We conducted a natural field experiment involving 70,000 households in an urban area. We study the extent to which they separate their waste, a behavior that is prone to habit formation. The incentive consisted of a letter informing households that disposal of non-separated waste is illegal and punishable by a fine, followed by an intensive and highly salient crackdown of four weeks. The crackdown had a large and instantaneous effect on household behavior. Most of the effect was still apparent up to seven months later. The stable, higher rate of separating waste is likely to have been sustained by a new, improved habit: crowding out of intrinsic motivation by the transactional motive to avoid punishment is found to be only transitory. In contrast to the threat of punishment, two behavioral interventions, one conveying the injunctive norm, another one the descriptive norm, are not found to have any effect.
Intervention
The treatment consists of informing all households by way of a letter from the municipality of Tilburg that not separating waste is illegal and punishable by a fine of €90, followed by a one-month, intensive and highly salient enforcement campaign.
Experimental design
We randomize the timing that a garbage collection route is to receive the treatment over a period of eight months. Randomization involved 60 of the 65 routes; the first five were pre-selected by the municipality because of implementation considerations. This phase-in design came down to randomizing the order in which the twelve remaining collection routes for a specific day of the week were to be treated (e.g. the twelve Monday routes; one Monday route was left out of randomization).
To increase the chance that on average similar routes were treated at different times, we used stratified random ordering of the timing of the treatment based on two variables: the share of residual waste in total waste pre-treatment (September-November 2014), and the percentage of terraced houses on a route in 2014. Both variables are indicators of the (lack of) potential for separation of waste. Once the ordering of the twelve routes for each day of the week was determined, we combined them into groups of five for a complete week of treatment in the Tilburg area.
Design details
Papers from this trial
- Van Soest, Daan, and Ben Vollaard, 2019, Breaking habits, unpublished manuscript, Tilburg University. https://surfdrive.surf.nl/files/index.php/s/5aN3W3raOrdtI4k
Cited by
No citation data available for this registration yet.