Using reminder messages to increase recycling behaviour in Peru

How to cite

Fuhrmann-Riebel, Hanna; Verschoor, Arjan; D'Exelle, Ben; Tonke, Sebastian; Vargas, Kristian López. 2021. "Using reminder messages to increase recycling behaviour in Peru." AEA RCT Registry. AEARCTR-0007780. https://doi.org/10.1257/rct.7780-8.1

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

Hanna Fuhrmann-Riebel* , Arjan Verschoor , Ben D'Exelle , Sebastian Tonke , Kristian López Vargas

* Primary investigator

Registry ID
AEARCTR-0007780
DOI
https://doi.org/10.1257/rct.7780-8.1
First registered
Last updated
Status
Completed
Trial dates
2021-06-14 – 2023-09-30
Countries
Peru (Lima)
Registry record
http://www.socialscienceregistry.org/trials/7780

Abstract

The municipality of Miraflores in Lima, Peru, has established a recycling programme, in which households can participate voluntarily and free of charge. Although households sign up to the programme voluntarily, few recycle regularly, pointing towards a gap between people’s intention to recycle and their actual recycling behaviour. Reminders can help people to follow through with their intentions by addressing the problem of limited attention. In this study, we conduct a randomized controlled trial (RCT) to test whether sms reminders can increase recycling behaviour of households. In particular, we will contrast the effect of continuous vs. interrupted vs. restarted reminders on households’ recycling behaviour. We will measure recycling behaviour of households over 12 weeks in total. While the first three weeks will serve as a baseline measure, the subsequent nine weeks will be our intervention period where households will be randomly assigned to either i) a control group that does not receive any reminders, ii) a group that receives continuous reminders over the whole nine weeks (continuous treatment), iii) a group that receives reminders only for the first three weeks (interrupted treatment), iv) or a group that receives reminders for the first three weeks and for the last three weeks, with a three weeks’ pause in between (restarted treatment). Our design will allow us to analyse the effects of continuous reminders on households’ recycling behaviour, persistence of reminder effects after the intervention has ended, as well as restart effects when the intervention is taken up again. We will further observe whether our intervention has positive spillover effects on the recycling behaviour of neighbouring households.

Intervention

We conduct a randomized controlled trial (RCT) to test whether simple reminder messages can increase recycling behaviour of households. We are interested in understanding the effects of continuous vs. interrupted vs. restarted reminders on households’ recycling behaviour. In addition to analysing the recycling behaviour of our sample of households, we will measure whether our intervention has social spillover effects on neighbouring households. This means, we will observe whether neighbouring households of treated households will increase their recycling activity as well in response to increased recycling activity of the treated households.

Experimental design

We will measure households’ recycling behaviour over a total time period of 12 weeks, where the first three weeks will serve as a baseline measure and the subsequent nine weeks will be our intervention period.

We will randomly distribute our sample into four groups of equal size. The four groups will consist of three treatment groups and one control group, which will receive the reminders during the intervention period as follows (after a three weeks’ baseline period):
i) Control group (T0): does not receive any reminders
ii) Continuous treatment group (T1): receives continuous reminders over the whole nine weeks
iii) Interrupted treatment group (T2): receives reminders only for the first three weeks
iv) Restarted treatment group (T3): receives reminders for the first three weeks and for the last three weeks, with a three weeks’ pause in between

The reminders will be sent to households during the respective weeks in the early mornings of the collection days via sms through a Peruvian sms provider system.

Design details

Randomization method
The randomization will be done in office by a computer.
Randomization unit
The randomization will be performed at the individual household level.
Clusters
N/A
Observations
Our total sample size will be 1392 households.
Arms
We will randomly distribute our sample into four groups of equal size. Given that our overall sample consists of 1392 households, this means there will be 348 households per group/treatment arm (i.e. 348 in control group, 348 in continuous treatment group, 348 in interrupted treatment group, 348 in restarted treatment group).
Primary outcomes
Our primary outcomes will be the recycling behaviour of our sample of households during the intervention period. This means every week: • Whether a household recycles on the collection day (yes/no) • How many bags are recycled by a household per collection day (count variable) For the analysis, we will look at the behaviour each week as well as at aggregate measures over blocks of three weeks and over the whole intervention period.
Secondary outcomes
Conditional on finding the expected treatment effects, our secondary outcomes will be the social spillover effects of our intervention on neighbouring households during the intervention period. We will use the same outcome variables defined above, though applied to the neighbours.

Cited by

No citation data available for this registration yet.

Keywords