Two field experiments on the effectiveness of behavioral nudges to reduce plastic bag consumption.

The first experiment will test different ways of asking individuals to actively choose to engage in behavior that can be damaging to the environment. In the literature, this is termed an "active and enhanced active choice" environment. In the context of this experiment, the shop owner will give the customer the option to choose to get a plastic bag, which will be phrased in different ways, reflecting the degree of active choice and the saliency between the use of plastic bags and its impact on the environment. Three treatments will include 1) opt-in, 2) active choice, and 3) enhanced active choice.

The second experiment will test the efficacy of reminder messages to bring a reusable bag(s), and the provision of a reusable bag. Theoretically, a reminder nudge falls between cognitive and moral nudges, and it increases an individual's attention to a decision and reduces forgetfulness (Carlsson et al., 2020). Nevertheless, a reminder can bring a moral cost if it draws attention to a decision that the decision-maker eschews (Damgaard & Gravert, 2018). For example, being reminded about a task can invoke guilt in a person. However, Sunstein (2017) pointed out that a reminder may fail to change behavior if it is long and poorly worded, but it can succeed if it is short and straightforward. This experiment involves two treatments, one receiving reminder messages and a second treatment receiving reminder messages and a reusable bag.

2022-08-05

2022-09-15

For the first experiment, we are primarily interested in two outcomes.
1. A binary outcome of whether a customer takes a plastic bag or not during shopping.
2. The total number of plastic bags customers take at each store.

For our second experiment, the outcome is the number of plastic bags each participating customer took during the duration of the experiment.

Experiment #1
We will use a two-level cluster-randomized trial. Shop owners will be recruited for the study, and the treatments will be randomized among the recruited shops. The treatments will be conveyed using signage at the registrars. Participating shop owners will be trained to initiate a conversation with each customer to direct their attention to the signage in front of the cash register. We will observe individual transactions and conduct customer surveys with a subset of the customers. The enumerators will stand at a position where they can observe and record three items for each customer: (1) whether or not the consumer took a plastic bag and, if so, how many?, (2) the amount of transaction, and (3) did the consumer bring any bag. This survey will collect information on other covariates necessary for the statistical analysis of the treatment effect.

Experiment #2
We will use a multi-site blocked randomized controlled trial with individual-level treatment randomization. Shops will be recruited in different areas (blocks) to participate in the study. During the experiment, we will approach every other individual who shops in those stores to participate in a consumer survey. After getting the consent for participation from the customer, the enumerator will give the customer a shopping voucher card. The voucher will carry a monetary value for all participants. In one treatment, randomly selected customers will receive vouchers with reminder messages; in a second treatment, customers will receive the messages along with a reusable bag for future use.

Experiment #1
We will use a two-level cluster-randomized trial. First, we will recruit shops in the bazaars where shop owners will be offered compensation for participation in the study. Second, we will randomize the treatment among the recruited shops. We will observe individual transactions and conduct consumer surveys from the participating shops. We will apply the treatment through simple signage installed in front of the cash register in the shop. Participating shop owners will also be trained to initiate a conversation with consumers to direct their attention to the signage in front of the cash register. The intervention will continue for a week at each shop. For logistical considerations, we will rotate shops into treatments in a randomized order over four different weeks.
We will collect two types of data: a) we will observe individual customers' choices during a shop transaction. Our enumerators will stand by in a position where they can observe three things (1) whether or not the consumer took a plastic bag and, if so, how many?, (2) the amount of transaction, and (3) did the consumer bring any bag; b) we will conduct consumer surveys for every nth person shopping in that shop depending on the number of customers come in every hour; b) the survey will collect information on other covariates necessary for the statistical analysis of the treatment effect.

Experiment #2
We will use a multi-site blocked randomized controlled trial with individual-level treatment randomization. First, we will approach individual shops in different areas (blocks) to participate in the study and offer participation incentives to the shop owners. Next, we will approach every other customer who shops in those shops to participate in a customer survey. We will inform participating customers that we are surveying what items the customer buys from this shop for the next three weeks. After getting the consent for participation from the customer, the enumerator will give the customer a shopping voucher card. The voucher will carry a monetary value for all participants. In one treatment, randomly selected customers will receive vouchers with reminder messages; in a second treatment, customers will receive the messages along with a pair of reusable bags for future use. The participating customers will be told they can redeem the value of the products using the voucher card if they shop from this shop within the next three weeks at least three times. The other condition is that no other person in the household can redeem the value except the participating customer. The shopping voucher card is a direct incentive for participation for the customers.

Experiment #1: Cluster randomization using individual stores as clusters
Experiment #2: a multi-site blocked randomized controlled trial with individual-level treatment randomization.

Experiment #1: Shops
Experiment #2: individual customers

Yes

Experiment #1: 40 retail shops from five different local markets in Dhaka
Experiment #2: 15 shops in three different local markets in Dhaka

Experiment #1: 400 customer surveys from 40 clusters and at least 100 observational data from each cluster Experiment #2: 120 customer surveys

Experiment #1: 10 clusters per treatment arm. (three treatments and a control)
Experiment #2: 40 individuals under each treatment arm (two treatments and control)

We expect a 15-20% reduction in plastic bag use among the retail customers under the treatment conditions compared to the control group.