Accelerating Farmer Adoption of Edge-of-Field Conservation Practices on Leakiest Fields: Economic Experiment Evidence that Reduces Hypothetical Bias

Date:
Sep 2025
Investigators:
  • Jacqueline Comito
  • Wendong Zhang
  • Cathy Kling

Issue

Water quality in the Mississippi River Basin is severely impacted by nutrient pollution, with nearly half of U.S. lakes showing signs of degradation. Despite the proven benefits of conservation practices like cover crops and edge-of-field (EOF) practices (e.g., saturated buffers and bioreactors), adoption rates remain critically low — less than 6% for cover crops and under 1% for EOF practices in Iowa. This threatens the region’s ability to meet the Gulf of Mexico Hypoxia Task Force’s goal of a 45% nitrogen reduction by 2035.

Objective

This project seeks to:

  1. generate accurate estimates of farmers’ willingness to adopt EOF practices under real-world policy conditions, and
  2. evaluate the effectiveness of peer messengers in outreach efforts to improve understanding and adoption of these practices.

Approach

Building on previous INRC-funded research, the team will conduct a mixed-method study combining surveys and qualitative interviews with 4,000 farmers in high-impact areas identified through spatial targeting. A split-sample survey will compare responses under hypothetical and real payment scenarios to assess adoption likelihood and willingness to accept (WTA). The study will also test the influence of male and female peer messengers on farmer decision-making, challenging conventional assumptions about gender dynamics in outreach. Listening sessions will further explore farmer perceptions, knowledge gaps and barriers to adoption. Findings will inform more effective conservation program design and outreach strategies, ultimately supporting broader adoption of EOF practices and improved water quality across the region.

Award Number:
2025-10

Project Update

December 2025

Watershed selection and sampling strategy:

  • The TNC team will help identify four HUC-10 watersheds in each of the four study states, and we already know two: Morse¬ Cicero HUC-10 watershed and one in NW Iowa; we still need to identify the particular ones in Minnesota and Illinois.
  • The sampling of 1000 farm households in each state will be all, likely 300-500, “leakiest fields” from the HUC-10 watershed identified from , a sample of the remaining parcels within this HUC-10 watershed, and a further-sparse sample of the farmland owners of other HUC-10 watersheds within the same HUC-8 subwatershed.

Overall treatments

  • Control
  • Real payments treatment where the respondents
  • Local agency soft nudge where the conservation professional or government official contact is provided for respondents to reach out.
  • Local agency strong nudge where the respondent is told that a government official such as NRCS or soil & conservation district staff will reach out.
  • [optional for Iowa] Some messenger effects? E.g., infographic about saturated buffers by female messengers.

Design of the real-payments treatment

  • The literature shows that as long as we ensure and effectively communicate with participants that a not-small percent of participants who said Yes to the adoption questions will be paid real payments for installing saturated buffers, we could achieve incentive-compatible design without having to pay every “yes-respondent” money.
  • We want to employ a sequential “waves” collection approach at least for the online respondent sample: for the first 100 respondents, we emphasize that first this is a high-stakes and real payments where we will actually fund at least 10-15 saturated buffer installations, and also everyone who says Yes will be paid to install the saturated buffers. For the next 400 respondents, we will still emphasize the real payments nature but clarify that only likely 20 percent of Yes-respondents will be chosen; and for another 500 respondents, we will clarify that only 5-10% of Yes-respondents will get paid to install saturated buffers. In a way, we will test to what extent varying the percent of consequential actual payments alter the perceived “realness” of respondents regarding the design.

IRB

  • We will wait for the rough survey design to be completed to submit to IRB for survey (and that will go through Cornell) – likely in February 2026 and likely an expedited review rather than exempt review.
  • Comito will submit a separate IRB for the listening session portion.