Follow up sessions 2024 to Reducing Negative Production Externalities Using Integrated Consumer Ratings

Last registered on September 17, 2024

Pre-Trial

Trial Information

General Information

Title
Follow up sessions 2024 to Reducing Negative Production Externalities Using Integrated Consumer Ratings
RCT ID
AEARCTR-0014370
Initial registration date
September 15, 2024

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
September 17, 2024, 1:50 PM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Primary Investigator

Affiliation
University of Bergen, department of economics

Other Primary Investigator(s)

PI Affiliation
BI business school Oslo and Stavanger

Additional Trial Information

Status
In development
Start date
2024-09-16
End date
2024-12-31
Secondary IDs
D82
Prior work
This trial is based on or builds upon one or more prior RCTs.
Abstract
Consumer purchase decisions for goods and services increasingly take place on a platform that features a vendor reputation indicator based on past consumer experience of the product quality. At the same time, there has been a growth of added certification indicators to signal other aspects of a product or its production methods, such as indicators of environmental friendliness (ecolabels). The idea is that information provided via these extra indicators function as ‘green nudges’ to persuade consumers to buy more responsible ‘low externality’ products, which would in turn put market pressure on firms to invest in more responsible production methods.

A question arises in what forms information regarding production externalities can be presented to increase the market pressure on firms to reduce them. Specifically, will any pressure on producers always have to come only from ‘activist’ consumers who are willing to pay for externality reduction, or are there ways in which firms could feel pressure to reduce externalities from all consumers?
In this project, we investigate how information regarding production externalities (ecolabels) can be presented to create market pressure on firms to reduce them. Specifically and novelly, we ask whether reducing the dimensionality of different certificates and consumer experience ratings into one rating, can result in firms feeling pressure to reduce externalities from all consumers, not just ‘activist’ consumers. We address this question in two ways.

First, we set up a theoretical model where firms privately choose current product quality and production externality level. Consumers learn about firms’ past behavior through a product quality rating and an ecolabel, or via a single rating which combines product quality and externality concerns. When consumers separately observe both a product quality rating and an externality, firms either ignore the externality dimension, or invest in externality reduction, but only sell to ‘activist’ consumers. In contrast, if the two concerns are combined into one indicator, there exists an equilibrium in which, regardless of there existing enough ‘activist’ consumers, a firm both invests in high product quality and low negative externality.

Second, we plan to test our hypotheses by running an incentivized economics laboratory experiment. In the experiment, we will create a small dynamic market setup similar to that in our theoretical model. We plan two treatments: One where consumers observe a firm’s reputation reflected in both a product quality and product externality rating, and one where the consumers observe only one combined rating to inform their choices.

We believe that results from this project will have important implications for how information to consumers can be presented in order to increase the market pressure on firms to reduce negative externalities.
External Link(s)

Registration Citation

Citation
de Haan, Thomas and Magnus Knutsen. 2024. "Follow up sessions 2024 to Reducing Negative Production Externalities Using Integrated Consumer Ratings." AEA RCT Registry. September 17. https://doi.org/10.1257/rct.14370-1.0
Experimental Details

Interventions

Intervention(s)
The experiment is planned to again be run at the economic experimental lab at BI Oslo and data collection is planned to start from September 16th, 2024.
The experiment will be run using zTree software (Fishbacher 2007). In the experiment participants are randomly assigned the role of producers and consumers and divided in matching groups of 6 (3 Producers, 3 Consumers). They will be randomly matched within their matching groups over approximately 48 rounds.

In each round, one producer is randomly matched with one consumer. Producers and consumers are of a “Regular” type, with 90% probability or a “Niche” type, with 10% probability (each individually determined). Each round a producer chooses whether to invest in one or both product “Attributes”

The cost of investing in Attribute A is 15 points for both producer types. The cost of investing in Attribute B is 10 points for a Regular producer and 0 points for a Niche producer. Each round a consumer also chooses whether to buy the offered product in a round. Both consumer types earn 100 points if they purchase from a producer who has invested in Attribute A. Niche consumers lose 50 points if they purchase from a producer who has not invested in Attribute B while Regular consumers´ earnings are unaffected by this.

Consumers do not observe the current round investment decisions made the producer they are matched. Consumers do observe ratings of the producer they are matched with in a round (from round 3 and onwards after the start of the experiment and after a “ratings reset”). The ratings summarize average investment behavior of the producer in previous rounds.

There will be two treatments and treatments will be randomized also within an experimental session. In one treatment the consumers will observe two separate ratings, one summarizing the past relative frequency of investment in each attribute separately. In a second treatment the consumers only observe one rating for each matched producer which will be a weighted sum of the past relative frequency of investment into both attributes.

The experiment lasts a number of rounds. How many rounds exactly is randomly determined. The experiment will last for approximately 48 rounds and most likely (with >80% probability) between 30 and 70 rounds. The experiment will run over 4 stretches of rounds, where in between, three times the ratings will be reset. A stretch will last for at least for 5 rounds. After 5 rounds, there is after each round a 12.5% probability that the ratings are reset.

Highlighed notable changes in the design compared to the initial data collection from 2022
• There is no general negative externality connected to the production decisions of the producers like there was before related to the producers’ decisions to invest in “Damage Reduction”. Instead, producers are asked to invest in two possible product attributes, “Attribute A” and, ”Attribute B”.
• Also instead, there are two types of producers and two types of consumers. “Regular” producers, and “Niche” producers with the difference that niche producers can invest in product attribute B for free. And also “Regular” consumers and “Niche” consumers, where the difference is that niche consumers suffer a 50 point loss if they purchase a product by a producer who has not invested in attribute B.
• Instead of a 6-pointed-line market research graph depicting the relationship between past reputation scores observed by consumers and their likelihood to have purchased a product, we instead display two sets of reputation indicators, one depicting the average producer rating(s) for producers who’s product went on to be purchased in a round by the consumer and one depicting the average producer ratings in the case when the consumer did not purchase.
Intervention Start Date
2024-09-16
Intervention End Date
2024-12-31

Primary Outcomes

Primary Outcomes (end points)
We hypothesize that the market setup where seller reputation is summarized in a single reputation rating will converge to an outcome with a higher likelihood for a producer to invest in both attributes than in the market setup where seller reputation is denoted in two separate ratings.
Primary Outcomes (explanation)
The way we will test the hypothesis is to look at the average rate of investment into each attribute, particularly expecting a difference for attribute B, per matching group, taken over all periods (although the termination is determined randomly per matching group as described above). We will run a non-parametric Mann-Whitney test to compare the investment rates between matching groups of the different treatments.

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The experiment is planned to again be run at the economic experimental lab at BI Oslo and data collection is planned to start from September 16th, 2024.
The experiment will be run using zTree software (Fishbacher 2007). In the experiment participants are randomly assigned the role of producers and consumers and divided in matching groups of 6 (3 Producers, 3 Consumers). They will be randomly matched within their matching groups over approximately 48 rounds.

In each round, one producer is randomly matched with one consumer. Producers and consumers are of a “Regular” type, with 90% probability or a “Niche” type, with 10% probability (each individually determined). Each round a producer chooses whether to invest in one or both product “Attributes”

The cost of investing in Attribute A is 15 points for both producer types. The cost of investing in Attribute B is 10 points for a Regular producer and 0 points for a Niche producer. Each round a consumer also chooses whether to buy the offered product in a round. Both consumer types earn 100 points if they purchase from a producer who has invested in Attribute A. Niche consumers lose 50 points if they purchase from a producer who has not invested in Attribute B while Regular consumers´ earnings are unaffected by this.

Consumers do not observe the current round investment decisions made the producer they are matched. Consumers do observe ratings of the producer they are matched with in a round (from round 3 and onwards after the start of the experiment and after a “ratings reset”). The ratings summarize average investment behavior of the producer in previous rounds.

There will be two treatments and treatments will be randomized also within an experimental session. In one treatment the consumers will observe two separate ratings, one summarizing the past relative frequency of investment in each attribute separately. In a second treatment the consumers only observe one rating for each matched producer which will be a weighted sum of the past relative frequency of investment into both attributes.

The experiment lasts a number of rounds. How many rounds exactly is randomly determined. The experiment will last for approximately 48 rounds and most likely (with >80% probability) between 30 and 70 rounds. The experiment will run over 4 stretches of rounds, where in between, three times the ratings will be reset. A stretch will last for at least for 5 rounds. After 5 rounds, there is after each round a 12.5% probability that the ratings are reset.

Highlighed notable changes in the design compared to the initial data collection from 2022
• There is no general negative externality connected to the production decisions of the producers like there was before related to the producers’ decisions to invest in “Damage Reduction”. Instead, producers are asked to invest in two possible product attributes, “Attribute A” and, ”Attribute B”.
• Also instead, there are two types of producers and two types of consumers. “Regular” producers, and “Niche” producers with the difference that niche producers can invest in product attribute B for free. And also “Regular” consumers and “Niche” consumers, where the difference is that niche consumers suffer a 50 point loss if they purchase a product by a producer who has not invested in attribute B.
• Instead of a 6-pointed-line market research graph depicting the relationship between past reputation scores observed by consumers and their likelihood to have purchased a product, we instead display two sets of reputation indicators, one depicting the average producer rating(s) for producers who’s product went on to be purchased in a round by the consumer and one depicting the average producer ratings in the case when the consumer did not purchase.
Experimental Design Details
Not available
Randomization Method
Randomisation is done via both the computer (to determine the number of rounds in a session) and via a randomized entry procedure in the lab to assign people to seats.
Randomization Unit
The relevant data point will be matching groups consisting of 6 participants (3 in the role of producer, 3 in the role of consumer)
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
We plan to collect between 10-15 matching group data points per treatment
Sample size: planned number of observations
We plan to collect between 10-15 matching group data points per treatment, this will likely have us recruit between 100 and 150 participants
Sample size (or number of clusters) by treatment arms
We plan to collect between 10-15 matching group data points per treatment
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
See attached pre plan document
IRB

Institutional Review Boards (IRBs)

IRB Name
IRB Approval Date
IRB Approval Number
Analysis Plan

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information