Implicit and explicit preferences for conventional and organic food

How to cite

Essl, Andrea; Dadic, Hana; Zumthurm, Samuel. 2025. "Implicit and explicit preferences for conventional and organic food." AEA RCT Registry. AEARCTR-0015729. https://doi.org/10.1257/rct.15729-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Andrea Essl* , Hana Dadic , Samuel Zumthurm

* Primary investigator

Registry ID
AEARCTR-0015729
DOI
https://doi.org/10.1257/rct.15729-1.0
First registered
Last updated
Status
In development
Trial dates
2025-01-01 – 2026-09-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/15729

Abstract

Cunningham and de Quidt (2024) introduced a theoretical framework for identifying implicit preferences. We apply this framework to empirically examine explicit and implicit preferences in food consumption. While many individuals express explicit preferences for organic food products likely due to self and social image concerns, their purchasing behavior often reflects preferences for cheaper conventional alternatives. To explore this discrepancy, we will conduct an incentivized laboratory experiment where participants state their willingness to pay (WTP) for cooked, colored eggs presented side-by-side in three joint evaluations. By systematically varying product attributes (organic versus conventional and an additional attribute, the painting types), we test whether individuals adhere to explicit preferences for organic eggs when the eggs differ only with respect to one attribute (organic versus conventional) but revert to implicit preferences for the cheaper, conventional eggs when an additional attribute (different painting types) dilutes their reasoning. This dilution weakens the signaling motive, allowing participants to justify their evaluation differently and enabling them to act on their implicit preferences. The theoretical framework assumes that we observe a participant’s revealed preferences, i.e., within-subject data. To address order effects, we will also use between-subject data by systematically varying the first joint evaluation. Furthermore, as people may publicly prefer organic eggs (explicit preference) but privately favor conventional eggs (implicit preference), the experiment will consist of two contextual treatments (between-subject design). A public treatment where individual evaluations will be revealed at the end of the experiment to the other participants and a private treatment where the individual evaluations remain concealed.

Intervention

In an incentivized laboratory experiment, we will (1) systematically vary two product attributes to identify potential implicit and explicit preferences and we will (2) manipulate the decision context, where participants indicate their willingness-to-pay (WTP) for eggs with different attributes either in a private or in a public treatment.

1) The first varying factor are two binary product attributes (production method: organic versus conventional and painting technique: opaque versus marbled). The product attributes will systematically vary. Three joint evaluations of products are needed to reveal implicit preferences (within-subject-design). In one joint evaluation the products only vary in the production method (simple comparison). A higher evaluation of one product compared to the other product can be clearly attributed to a preference for a production method. In two joint evaluations the products vary in the production method and the coloring technique (complex comparisons). Here, a higher evaluation of one product compared to the other product cannot be clearly attributed to a preference for a production method or a coloring technique. This is called dilution and allows participants to act according to their implicit preferences. To mitigate potential order and consistency effects, we will also use between-subject data by systematically varying the first joint evaluation, such that each of the three joint evaluation is made with the equally likely as the first evaluation across sessions. This allows us to test whether the whole sample has implicit and explicit preferences.
2) The second varying factor is the contextual setting (between-subject). In the private treatment, participants' decisions remain anonymous. In the public treatment, participants know that one of their WTP decisions (randomly selected for payout) will be publicly revealed to all study participants.

We will elicit participants’ WTP using the Becker–DeGroot–Marschak (BDM) method. Participants receive an endowment and know that one of the products will randomly be drawn for their payout. If their stated WTP exceeds a randomly drawn price, they purchase the product at that randomly drawn price and receive the product and the payment (endowment minus randomly drawn price) at the end of the study; otherwise, they do not receive the product but keep the entire endowment.

Experimental design

In the laboratory experiment, participants will repeatedly indicate their WTP for product pairs in three joint evaluations, where two products are presented next to each other. Participants indicate their WTP for both products separately, resulting in six WTP entries per participant. Across three joint evaluations, the food products will systematically vary in the production method (conventional versus organic) and in the coloring technique (opaque versus marbled). In one joint evaluation, the two products only vary in the production method (simple comparison). Participants state their WTP for conventional marbled colored eggs and organic marbled colored eggs, allowing clear attribution of preferences to the production method. In two joint evaluations, both production method and painting type vary (complex comparison). In one complex joint evaluation, participants have to indicate their WTP for conventional marbled colored eggs and organic opaque colored eggs. In the other complex joint evaluation, participants have to indicate their WTP for organic marbled colored eggs and conventional opaque colored eggs. Since both attributes differ, any preference cannot be clearly attributed to a preference for a production method or a painting type. To minimize order and consistency effects, participant complete a filler task (identifying the most frequently occurring letter in a sequence) between each joint evaluation. In addition, the order of the three joint evaluations is randomized, ensuring that each appears equally likely as the first comparison. This design allows for between-subject analysis as a robustness check.

In addition, we aim to examine whether the presence of implicit and explicit preferences are influenced by the contextual setting, specifically, how participants behave when their WTPs are made private versus public. To explore this, we will use a between-subject design with two treatments: 1) In the public treatment, at the end of the experiment, one joint evaluation is randomly selected, and participants’ stated WTP for both products is disclosed to the other participants in the same session. Participants are informed at the beginning that any of their three joint evaluations could be selected and made visible. In the private treatment, the procedure is identical, except that participants' responses remain anonymous. Participants are informed that we use a double-blind design for the payout - a person in another room will prepare participants’ payout (product and/or cash) in an envelope and leave the room after everything is prepared. Participants will collect their payout autonomously without anyone else knowing what they received.

Furthermore, we elicit beliefs on descriptive and injunctive social norms. Belief elicitation will be incentivized and takes place after participants stated their WTP for all products. Participants will also complete a survey on self-report scales (e.g., social image concerns, self-image concerns, liking of products, attitude towards organic, frequency of buying organic, reasons to buy organic, barriers to buy organic, estimating food market prices, whether they know people participating in the same session) and answer demographic questions. We will examine whether intransitive choices are influenced by factors like social signaling, self-signaling and beliefs on social norms. Further, we will also measure reaction time in the joint evaluations, that is, the time participants need to indicate their WTP for both products within a joint evaluation.

Hypotheses

Dilution:

H1: Participants have explicit preferences for organic eggs and implicit preferences for conventional eggs.

H1a: Participants have a higher WTP for organic marbled colored eggs when they are compared to conventional marbled colored eggs (simple comparison) than when they are compared to conventional opaque colored eggs (complex comparison).

H1b: Participants have a higher WTP for conventional marbled colored eggs when they are compared to organic opaque colored eggs (complex comparison) than when they are compared to organic marbled colored eggs (simple comparison).

Signaling:

H2: Participants’ differences between implicit and explicit preferences are stronger in the public treatment than in the private treatment.

H2a: There is an interaction between dilution and decision context such that participants decrease their WTP for organic marbled colored eggs from the simple comparison to the complex comparison stronger when they are in the public treatment than when they are in the private treatment.

H2b: There is an interaction between dilution and decision context such that participants increase their WTP for conventional marbled colored eggs from the simple comparison to the complex decision stronger when they are in the public treatment than when they are in the private treatment.

Cunningham, T., & de Quidt, J. (2024). Implicit Preferences. https://jondequidt.com/pdfs/paper_implicit.pdf

Design details

Randomization method
Computer (lab experiment)
Randomization unit
Experimental Session
Clusters
The number of participants planned for the experiment is about 312 people. Participants who report a WTP of zero in all six WTP entries are excluded from the analyse
Observations
The number of participants planned for the experiment is about 312 people. Participants who report a WTP of zero in all six WTP entries are excluded from the analyses.
Arms
About 156 participants per treatment
Primary outcomes
Willingness-to-pay (WTP)
Secondary outcomes
Social norms, self-signaling, social-signaling

Cited by

No citation data available for this registration yet.

Keywords