Valuing Honey from Ecosystem Service Systems

2026-08-01

2026-09-30

Choices of honey

Self-reported certainty about reference price
Self-reported credibility of information received
Perceived importance of price, environmental benefits, and energy benefits in purchasing decisions
Stated environmental and climate change attitudes (measured using Six Americas Super Short Survey questions)
Environmental altruism (measured by a dictator game)
Self-reported reference price for honey
Beliefs about others' willingness to pay for ecosystem service honey
Self-other comparison of willingness to pay

This study uses a combination of survey questions and an incentive-compatible purchasing experiment to measure consumer preferences for honey produced from different ecosystem service practices. Participants first answer questions about their price expectations, then are randomly assigned to receive external price information based on their expected price or no price information, and finally make a series of real purchasing decisions for honey at varying prices. The survey also asks questions to elicit participants' climate change and environmental attitudes and views. The study examines how price expectations and external price information influence consumer choices and willingness to pay, and whether stated environmental attitudes align with actual purchasing behavior.

This study uses a combination of survey questions and an incentive-compatible purchasing experiment to measure consumer preferences for honey produced from different ecosystem service practices. Participants complete a survey, make a series of purchasing decisions with real money, receive information about environmental benefits, and then make another series of purchasing decisions. The study examines how information affects consumer choices and whether stated environmental attitudes align with actual purchasing behavior.

Not available

Participants will be automatically assigned to one of the three experimental groups upon starting the online survey on the Qualtrics platform.

Individual participants. Each participant is randomly assigned to one experimental group independently. There is no clustering at the session or location level.

No

Not applicable (individual-level randomization, no clustering).

1000 participants.

330-350 for each experimental group.

We conducted an ex‑ante Monte Carlo simulation (1,000 loops) using our 12‑task choice design, accounting for individual‑level clustering. The target parameter is the utility coefficient (β) on the logit scale, with the unobserved error variance normalized to π²/3 ≈ 3.29. To account for taste heterogeneity, baseline parameters were informed by a pilot study (N = 19; 304 observations). The Minimum Detectable Effect Size for our primary outcomes—the interaction effects between reference price treatments and honey attribute (Prairie Strip and Solar Farm)—is set to a conservative small effect size (Cohen’s d = 0.20), which corresponds to a fixed structural mean shift of approximately 0.36 on the logit scale. Simulation results show that at a 5% significance level, a sample of 1000 respondents achieves the statistical power of 80% across all interaction parameters.