Norm-based feedback on household waste recycling: a field experiment

How to cite

Ek, Claes; Söderberg, Magnus. 2019. "Norm-based feedback on household waste recycling: a field experiment." AEA RCT Registry. AEARCTR-0003301. https://doi.org/10.1257/rct.3301-2.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Claes Ek* , Magnus Söderberg

* Primary investigator

Registry ID
AEARCTR-0003301
DOI
https://doi.org/10.1257/rct.3301-2.0
First registered
Last updated
Status
Completed
Trial dates
2018-05-01 – 2020-12-31
Countries
Sweden (Partille and Varberg municipality)
Registry record
http://www.socialscienceregistry.org/trials/3301

Abstract

The last decade has seen a marked upsurge of research into ‘norm-based feedback’ interventions, which aim to reduce the consumption of some resource by providing users with information about their use compared to e.g. neighbors in their community. As far as we are aware, ours is the first study on the effect of providing such feedback within the waste domain. We run separate randomized controlled trials of accurate and household-specific norm-based feedback on residual (unsorted) waste in two municipalities in western Sweden, Partille and Varberg. In total, some 20,000 single-family households participate, about three-quarters of which are located in Varberg. Our main feedback condition, which we term “static” feedback, replicates the by now standard Opower-style norm and involves comparing each household’s own residual waste (in kilograms per person) with an average among roughly 100 nearby households. We also explore differently presented, “dynamic” feedback motivated by psychological evidence that norms focusing on changes in behavior, rather than snapshot pictures of consumption differences within a population, are more effective at reducing use.

Experimental design

We run separate trials of norm-based feedback on household residual waste (the unsorted waste fraction) in two municipalities in western Sweden, Partille and Varberg. In total, some 20,000 single-family houses participate in these studies, about three-quarters of which are located in Varberg. As in previous studies, the aim of the experiment is to examine whether treated households that receive feedback reduce their generation of residual waste compared to a control group that does not receive feedback.

In both municipalities, households are divided roughly equally into three treatment arms, including a control group. However, the two treatments differ across our participant municipalities.

Design details

Randomization method
Random number generator.
Randomization unit
Cluster randomization with blocking: three clusters per block. Treatment perfectly correlated within cluster. Each block/cluster is a contiguous set of households.
Clusters
Partille: 55*3 = 165 Varberg: 172*3 = 516
Observations
Partille: 5,756 Varberg: 15,723 Due to some households opting out of the study, we expect our actual data set to cover roughly 5,000 in Partille and 15,000 in Varberg.
Arms
Partille: 55 clusters/treatment arm Varberg: 172 clusters/treatment arm
Primary outcomes
Collected residual waste (kgs/person)
Secondary outcomes
Collected food waste (kgs). If possible, electricity use by households in Varberg

Cited by

No citation data available for this registration yet.

Keywords