Nudging to reduce littering in waste disposal areas – A field experiment in Austria’s community houses

How to cite

Grosch, Kerstin; Gangl, Katharina; Walter, Anna. 2020. "Nudging to reduce littering in waste disposal areas – A field experiment in Austria’s community houses." AEA RCT Registry. AEARCTR-0006108. https://doi.org/10.1257/rct.6108-1.1

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

Kerstin Grosch* , Katharina Gangl , Anna Walter

* Primary investigator

Registry ID
AEARCTR-0006108
DOI
https://doi.org/10.1257/rct.6108-1.1
First registered
Last updated
Status
Completed
Trial dates
2020-06-30 – 2020-09-11
Countries
Austria (Vienna)
JEL codes
C93, Q50, D91
Registry record
http://www.socialscienceregistry.org/trials/6108

Abstract

This field experiment tests an intervention to reduce littering in waste disposal areas in community houses in Vienna/Austria. Community houses differ in size from about 20 apartments in some houses to over 2000 apartments in other houses. The inhabitants take away their garbage bags in shared waste disposal areas. Waste disposal areas in community houses are often littered. This problem causes monetary costs for cleaning personnel as well as psychological costs for residents who feel disturbed by litter in this area. In this study, we test instruments to reduce littering in these waste disposal areas in a cost-effective way. We test four different nudges in this study. Currently, there is little evidence on how best to engage inhabitants of those community houses to preserve cleanliness in the shared waste disposal areas. In a randomized controlled trial, we test four interventions in around 560 waste disposal areas in over 90 different community houses. In a control group, there will be no poster and everything as is. The interventions are posters that will be placed above the garbage containers. Besides two classical nudges that work with pictures of landscapes or watching eyes and on a more subconscious level (system 1), there will be two posters focusing on raising awareness for the injunctive norm and the negative monetary consequences of littering (system 2). Our outcome variable is the tidiness/messiness of the waste disposal areas. We shoot pictures three times; before, shortly after hanging up the posters and after a couple of weeks. This way, we have a proxy on messiness at the baseline level, short-term effects, and long-time effects of the different interventions. The four treatments allow as to compare not only the intervention group to the control group but also make comparisons concerning effectiveness across different interventions.

Experimental design

This field experiment tests an intervention to reduce littering in waste disposal areas in community houses in Vienna/Austria. Community houses differ in size from about 20 apartments in some houses to over 2000 apartments in other houses. The inhabitants take away their garbage bags in shared waste disposal areas. Waste disposal areas in community houses are often littered. This problem causes monetary costs for cleaning personnel as well as psychological costs for residents who feel disturbed by litter in this area. In this study, we test instruments to reduce littering in these waste disposal areas in a cost-effective way. We test four different nudges in this study. Currently, there is little evidence on how best to engage inhabitants of those community houses to preserve cleanliness in the shared waste disposal areas. In a randomized controlled trial, we test four interventions in around 560 waste disposal areas in over 90 different community houses. In a control group, there will be no poster and everything as is. The interventions are posters that will be placed above the garbage containers. Besides two classical nudges that work with pictures of landscapes or watching eyes and on a more subconscious level (system 1), there will be two posters focusing on raising awareness for the injunctive norm and the negative monetary consequences of littering (system 2). Our outcome variable is the tidiness/messiness of the waste disposal areas. We shoot pictures three times; before, shortly after hanging up the posters and after a couple of weeks. This way, we have a proxy on messiness at the baseline level, short-term effects, and long-time effects of the different interventions. The four treatments allow as to compare not only the intervention group to the control group but also make comparisons concerning effectiveness across different interventions.

Design details

Randomization method
The service agent for the community homes provided us with a list of 98 community houses. We randomize the treatments within community houses. We assign a unique number to each waste disposal area that identifies the community house and the specific waste disposal area. In the list, we sort the waste disposal areas by community houses and assign the treatment randomly. For this, we use a random number generator by the software excel between 1 and 5 whereas each number identifies one of the four treatment groups or the control group. We stratify by outdoor or indoor waste disposal area since we suspect that the effectiveness of the interventions may be heterogeneous across the two conditions. To ensure a balanced sample concerning treatment allocation, we kept updating the randomization until a distribution test (chi2) signaled that treatment conditions are assigned randomly across indoor and outdoor waste disposal areas.
Randomization unit
We randomized within community houses and stratified by indoor/outdoor waste disposal areas.
Clusters
about 95 community houses
Observations
about 455 waste disposal areas
Arms
about 90 waste disposal areas in each treatment group and in the control group, respectively.
Primary outcomes
Outcome variable: floor_rating Description of the outcome variable: Independent research assistants rate the floor pictures of the waste disposal areas (the treatment is not visible on the picture); the rating is based on perception from 1 to 7 whereas 1 indicates a sparkling clean floor and 7 a complete mess. With the numbers in between the research assistants can grade based on a pre-defined categorization. We employ two independent research assistants and take the average of the two floor ratings for the analysis.

Cited by

No citation data available for this registration yet.

Keywords