A Field Experiment on Information from Senders with Shared or Divergent Identities

Last registered on August 25, 2025

View Trial History

Pre-Trial

Trial Information

General Information

Title

A Field Experiment on Information from Senders with Shared or Divergent Identities

RCT ID

AEARCTR-0016608

Initial registration date

August 22, 2025

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

August 25, 2025, 8:46 AM EDT

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

Locations

Country

Canada

Region

Ontario

Primary Investigator

Name

Siyu Wang

Affiliation

Wichita State University

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Tongzhe Li

PI Affiliation

University of Guelph

Contact Investigator

Additional Trial Information

Status

In development

Start date

2025-09-07

End date

2025-09-14

Keywords

Agriculture, Behavior, Environment & Energy

Additional Keywords

Field Experiment, Belief, Farmers' decisions, Auctions

JEL code(s)

Secondary IDs

D83

Prior work

This trial does not extend or rely on any prior RCTs.

Abstract

Information is often cited as a key driver of public opinion and behavioral change. Moreover, the identity of the messenger can shape how information is received, with evidence from the literature suggesting that echo chambers form around many issues. In this study, we examine three main questions:
1. Do information receivers prefer messages from identity-matched senders or from identity-mismatched senders?
2. Does identity matching, i.e. receiving messages from a sender who shares versus differs in identify measured by prioritization of profit versus environmental friendliness, influence participants’ beliefs about the effectiveness of a technology and their willingness to pay for it?
3. Does the stated priority of the message sender affect participants’ beliefs about the effectiveness of a technology and their willingness to pay for it?
To answer these questions, we will conduct a lab-in-the-field experiment at the 2025 Canada’s Outdoor Farm Show (COFS) in Ontario, Canada. Participants will first indicate their priorities regarding environmental versus profit-based dimensions of fertilizer management, a key practice for agricultural production. They will then receive information from a real agronomist, who represents advisors who prioritize either profitability or environmental friendliness. After viewing this information, participants will answer a question about the perceived effectiveness of and bid for a fertilizer management product using a cost-share Becker–DeGroot–Marschak (BDM) mechanism (Becker et al., 1964).

External Link(s)

Registration Citation

Citation

Li, Tongzhe and Siyu Wang. 2025. "A Field Experiment on Information from Senders with Shared or Divergent Identities." AEA RCT Registry. August 25. https://doi.org/10.1257/rct.16608-1.0

Sponsors & Partners

Experimental Details

Interventions

Intervention(s)

Timeline and Sample
Farmers will be recruited at COFS in Woodstock (Ontario) between September 9-11, 2025 to participate in a sender-selection task followed by an experimental auction in the format of cost-share BDM mechanism. Each participant will place a bid on a fertilizer management product, specifically a dual nitrogen inhibitor used in conjunction with fertilizers. COFS is one of the largest outdoor farm shows in Eastern Canada and an ideal venue to run a field experiment as it attracts more than 40,000 farmers annually.

Overview of Experiment
This experiment will establish the causal relationship with how information sources (see Table 1 for the specific treatments) affect farmers’ beliefs on the benefits of fertilizer management and their bids for relevant product. We are going to conduct this experiment over the course of three days. Each participant will bid on the required cost-share amount ∈[0%,100%] for the product. After submitting their bids, participants will go through a post-experiment questionnaire and answer demographic and farm-specific questions. Each farmer will receive $20 (CAD) for participating in the experiment. Python language will be used to program this experiment, and participants will use HTML pages shown on iPads to provide their inputs.

Table 1. Between-subjects treatments randomized at the individual level.

Treatment Receiver Identity/Preference Sender Identity
T1 A/B Prioritize Environmentally Friendliness Prioritize Environmentally Friendliness
T2 A/B Prioritize Profitability Prioritize Profitability
T3 A/B Prioritize Environmentally Friendliness Prioritize Profitability
T4 A/B Prioritize Profitability Prioritize Environmentally Friendliness

Ex-ante power analysis shows that to detect a treatment effect in the size of 5 percentage points in terms of bid difference in the cost-share auction with an uninformed untreated mean and a standard deviation of 10%, the minimum sample size needs to be 64 per treatment group; a total sample size of 512 (64 x 8) to detect the effect of different appeals based on pairwise comparisons. Since representative farmers are recruited at a specific event, we expect the sample size will depend on the attendance at the COFS. We aim for a minimum of 512 participants while trying to reach 750 participants if conditions permit. Power analyses are based on a 5% significance level with 80% power, which is standard practice in the literature.

Description of the incentive-compatibility component of the experiment

A cost-share BDM mechanism will be used to ensure incentive compatibility (i.e., each participant will be incentivized to bid their actual valuation). At the end of the experiment, participants will be randomly chosen via computer with a probability of one out of a hundred to have their bid evaluated. An “experimental market price” (representing the percentage of the cost to be paid by the participant, which is equivalent to 100% – the percentage of the cost to be subsidized by the experimenter) will be randomly selected from a distribution based on actual subsidies available in the real world. Each randomly selected participant’s bid for the product will be compared to the “experimental market price.” If the participant’s bid is equal or higher, then they will purchase the product at the “experimental market price.” If the bid is lower, then they will not purchase the product. Under this mechanism, participants don’t want to over-bid as there is no competition against others while over-bidding may result in paying a price higher than their actual valuation. Participants also don’t want to under-bid because in such a case they may forego the opportunity of purchasing the product at a price lower than their actual valuation (i.e., at a subsidized price through the cost-share mechanism).

Intervention (Hidden)

Intervention Start Date

2025-09-07

Intervention End Date

2025-09-14

Primary Outcomes

Primary Outcomes (end points)

(a) participants’ preferences regarding the identity of the message sender; (b) participants’ beliefs about the benefits of the nitrogen management product; and (c) participants’ bids for the product across the different between-subject treatment conditions.

Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)

demographic and farming-specific variables

Secondary Outcomes (explanation)

Experimental Design

Overview of Experiment
This experiment will establish the causal relationship with how information sources (see Table 1 for the specific treatments) affect farmers’ beliefs on the benefits of fertilizer management and their bids for relevant product. We are going to conduct this experiment over the course of three days. Each participant will bid on the required cost-share amount ∈[0%,100%] for the product. After submitting their bids, participants will go through a post-experiment questionnaire and answer demographic and farm-specific questions. Each farmer will receive $20 (CAD) for participating in the experiment. Python language will be used to program this experiment, and participants will use HTML pages shown on iPads to provide their inputs.

Table 1. Between-subjects treatments randomized at the individual level.

Treatment Receiver Identity/Preference Sender Identity
T1 A/B Prioritize Environmentally Friendliness Prioritize Environmentally Friendliness
T2 A/B Prioritize Profitability Prioritize Profitability
T3 A/B Prioritize Environmentally Friendliness Prioritize Profitability
T4 A/B Prioritize Profitability Prioritize Environmentally Friendliness

Experimental Design Details

Randomization Method

randomized by computer programming

Randomization Unit

individual

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

512 participants

Sample size: planned number of observations

512 participants

Sample size (or number of clusters) by treatment arms

64 participants per treatment

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

Ex-ante power analysis shows that to detect a treatment effect in the size of 5 percentage points in terms of bid difference in the cost-share auction with an uninformed untreated mean and a standard deviation of 10%, the minimum sample size needs to be 64 per treatment group; a total sample size of 512 (64 x 8) to detect the effect of different appeals based on pairwise comparisons. Since representative farmers are recruited at a specific event, we expect the sample size will depend on the attendance at the COFS. We aim for a minimum of 512 participants while trying to reach 750 participants if conditions permit. Power analyses are based on a 5% significance level with 80% power, which is standard practice in the literature.

Supporting Documents and Materials

IRB