Promote plastic recycling using behavioral insights.

This project aims to increase recycling behavior collaborating with local authorities and waste-pickers cooperatives. It will compare the efficacy of a standard educational intervention and behaviorally informed ones. If successful, this research will be valuable because it will hypothetically prove two things: first that behaviorally informed interventions are more effective in increasing plastic recycling behavior than standard educational campaigns in Belo Horizonte (Brazil). Secondly it will prove which behaviorally informed letter is the most effective to increase recycling behavior in a context where behavioral interventions in the recycling realm have never been carried out. 

In Latin America and The Caribbean (LAC) only a few countries use formal recycling methods as part of their municipal solid waste policies. In consequence, a large proportion of the 160 million tons of waste generated every year in the region end up in landfills and dumps. Partly due to this lack of formality in recycling and an unequal socio-economic reality, in LAC there is a group of professionals who make a living, collecting and selling recyclables: the waste-pickers. 

Informal names and numbers of waste-pickers vary greatly across countries, from the 100.000 “cartoneros” in Peru (Linzner & Lange, 2013) to the 387.000 “catadores” in Brazil (S. Dias & Silva, 2017). Despite this heterogony, however, there is one thing several authors agree on, waste-pickers in LAC are not only important actors in waste collection but also in environmental protection and economic development. This intervention will test a behavioral approach that has already been successful in Peru, Guatemala and Argentina to increase tax compliance. 

There are seven waste-pickers cooperatives in Belo Horizonte who act as service providers to the city council in recycling schemes. It is in the best interest for both parties to increase recycling rates: the city council has recycling goals they have to achieve and the cooperatives, besides a fix income from their contracts with the city council,  receive a higher income the more recyclables they are able to collect. 

Collaboration with local authorities and waste-pickers cooperatives will be constant. The draft of the educational letters, data collection and how to assess the success of the intervention will be done in cooperation. 

This project uses the BASIC methodology from the OECD. To increase plastic recycling rates we will systematically modify the content of the educative letters sent to the citizens in Belo Horizonte. Based on previous research on sustainable behavior we know that: 

• •       When a desirable behavior wants to be promoted but is not yet normative, such is the case of this research proposal since less than 1% of the waste is recycled, highlighting the dynamic norm, i.e. how other people’s behavior is changing over time, is an effective strategy (Rare & Behavioural insights Team, 2019; Sparkman & Walton, 2017).  

  •       A meta-review of over 140 randomized control trials that use goal setting to drive behavioral change, found that environmental goals have the largest effect sizes (Epton, Currie, & Armitage, 2017). The authors also conclude that in several domains, goals are particularly effective when they are difficult, set publicly and part of a group goal.

  We will use 4 different models of letters to conduct the educational campaign: 

  •    L1: Standard educational campaign about recyclables. This is the intervention group corresponding to H1. 

• •    L2: The same letter as in L1 will be sent adding the in a salient place: “In the last six months more than X kg/month have been recycled by the neighbors of the X neighborhood. Please keep helping us!”. This condition corresponds to  H2. 

  •    L3: The same letter as in L1 will be sent adding in a salient place of the letter: “As a neighborhood we aim to recycle X kg  per month. Please help us to achieve this goal.” This condition corresponds to  H3.

  •    L4: The same letter as in L1 but adding the same dynamic norm of L2 and the same goal of L3 in salient places of the letter.  

  We will have a control group that will not receive any letters and will provide a measure of recycling rates in the absence of any intervention. The pretest will help us to have a reliable and sensible measurement of the dependent variable, the amount of recyclables collected per area. 

   

This research proposal aims to increase recycling behavior collaborating with local authorities and waste-pickers cooperatives. It will compare the efficacy of a standard educational intervention and behaviorally informed ones. If successful, this research will be valuable because it will hypothetically prove two things: first that behaviorally informed interventions are more effective in increasing recycling behavior than standard educational campaigns in Belo Horizonte (Brazil). Secondly it will prove which behaviorally informed letter is the most effective to increase recycling behavior in a context where behavioral interventions in the recycling realm have never been carried out. 

These results would also be beneficial for the waste-picking communities in LAC that have a low and volatile income contingent on how many recyclables they are able to collect and sell. (Auler, Nakashima, & Cuman, 2014; Colombijn, 2017; Dantas, 2017; S. Dias & Silva, 2017; Linzner & Lange, 2013)

A truck will pick up recyclables from a certain subarea of a neighborhood before dropping them at a waste-pickers cooperative. In collaboration with local authorities and cooperatives we can delimit these areas and assess both current recycling rates and population size. Once we have done this analysis, we will select those that are more similar in population size and current recycling rates. Due to the nature of this experiment, we cannot plan a sample size to achieve a certain power since we depend on the data obtained in the field. We will aim, however, for the largest sample size possible. 

From the clusters found, we will randomly select five groups: four will be treated and one will be the control group. Note that a treatment group will receive only one type of the letters defined in the methodology.  

The intervention will last 6 months. During the first 3 months we will delimit areas and assess current recycling rates and sample size. We will then randomize the areas and send the letters. 3 months afterwards, we will collect the amount of recyclables per area in kg as it is the best approximation to measure recycling behavior. Data will be excluded if noise is found, for example, a truck changed its route and is now picking up recyclables from different streets. 

In collaboration with the University of Bath. 

The team is a mix of researchers and practitioners with international experience.