Information Acquisition and Sustainable Consumption

The intervention offers consumers access to individualized information about the sustainability of a purchased product and close substitutes. The intervention focuses on plastic water bottles from two different brands that differ in their share of recycled material (brand A: 100% recycled material; brand B: 0-25% recycled material).

2022-09-23

2022-12-03

Whether a consumer (in the treatment "Info") acquires information about the recycled material of the purchased plastic water bottle.
Whether a consumer chooses a voucher for brand A (i.e., for water bottles that are made 100% from recycled plastic). Details of the voucher decision are explained in the Experimental Design.

Time spent on the decision screen (for consumers in treatment “Info”).
Time spent on the information screen (for consumers in treatment “Info” who acquired the information).
In addition to these pre-registered secondary outcome variables, we investigate the purchases made during and after the voucher period (e.g., whether any bottle from brand A was purchased).

Participants are users of an app from a large supermarket chain. We focus on plastic water bottles from two brands (A and B) with sizes 0.33L, 0.5L, 1L, 1.5L, 2L, and 6.25L. After having uploaded a purchase with at least one of the considered water bottles, participants are assigned to one of two treatments (only the first time the product is bought during the intervention period):
1. "No Info": Participants receive a "thank you" message for purchasing the product.
2. "Info": Participants receive a "thank you" message as in the treatment "No Info" but are additionally asked whether they would like to know more about the amount of recycled material in this and other bottles. They can click on "Know more" or "Close." If they click on the former, they receive information about the share of recycled material for four water bottles (including the one they have just uploaded).

Directly after having been assigned to a treatment (and seeing the messages associated with the treatment), all participants face a voucher decision. In the voucher decision, participants choose for which brand (A or B) they would like to receive a €5 voucher (or they can close the window without making a choice). Finally, they enter their email address and loyalty card number to receive the voucher.

Consumers are randomly assigned (with equal probabilities) to the treatment “No Info” or “Info” upon the first time they upload one of the considered products to the app during the intervention period. The randomization is implemented by the app developer.

Individual

No

1,000

1,000

No Info: ~50%
Info: ~50%

In the following, we present the ex-post MDEs (with 80% power and a significance level of 5%) in absolute value, based on simulations and conditional on the realized sample sizes. The size of the MDE in absolute value depends on whether the true effect/difference is positive or negative. Below, we always show the MDE for the more restrictive scenario (i.e., the MDE for the alternative scenario is weakly smaller than the MDE presented). 1. MDE for difference in information acquisition in treatment “Info” by purchasing history (consumers from brand A vs. consumers from brand B; using one of the two as the baseline): ~13pp (0.28 standard deviations) 2. MDE for average treatment effect of intervention on choosing sustainable voucher (using treatment “No info” as the baseline): ~7pp (0.14 standard deviations) 3. MDE for difference in conditional average treatment effect of intervention on choosing sustainable voucher by purchasing history (consumers from brand A vs. consumers from brand B, assuming only one consumer type is affected by the intervention, using consumer-type-specific values from treatment “No info” as the baseline, difference standardized by overall standard deviation in treatment “No info”): ~15pp (0.31 standard deviations)