Rural Drinking Water Habits, Information Provision, and Pollution Exposure

The Midwestern landscape is among the most altered and intensively managed ecosystems in the world. A perennial challenge associated with annual row crop production is its impacts on water quality. Public and private organizations have dedicated tremendous resources to improve on-farm practices, retire land, and restore wetlands. Despite these efforts, health-based drinking water quality violations continue to affect millions of Americans annually. Little is known about the degree to which these violations expose rural residents to unsafe drinking water, what efforts rural residents take to avoid possible exposure, and how information provision might improve the welfare of rural residents. We aim to fill in this knowledge gap. In this project, we propose to study rural households’ drinking water patterns, pollution avoidance behavior, and willingness to pay for pollution information. Specifically, we propose to (i) field a large-scale, longitudinal survey of rural households' drinking water sources and bottled water/filter purchases; (ii) study the impacts of providing households with in-home nitrate test kits on these purchases; (iii) elicit households’ willingness to pay for in-home nitrate pollution test kits, comparing stated versus revealed preference elicitation methods; and (iv) engage rural communities and create multi-faceted extension and education programming to improve knowledge and awareness of drinking water challenges and solutions. Results will be used to guide effective policy-making for water quality programs throughout the United States.

Methods: Primary data will be collected using household surveys. Supplemental data will include socio-demographic information from the U.S.Census and publicly available data on households' participation in state-sponsored drinking water testing programs. We will
study the impact of pollution information on households' survey responses using a randomized control trial, i.e., a random
subset of households' will be assigned to receive a nitrate test kit in between the first and second survey round. Willingness to
pay estimates will be elicited from the subset of homes that receive nitrate test kits in the second survey round. Survey
responses will be evaluated using standard statistical and econometric techniques. Results will include summary statistics of
key outcomes of interest (e.g., frequency of bottled water and water filter purchases), estimated impacts of pollution information
provision on those outcomes, and average willingness to pay estimates for more nitrate test kits.

External Link(s)

Registration Citation

Citation

Benning, Jamie et al. 2019. "Rural Drinking Water Habits, Information Provision, and Pollution Exposure." AEA RCT Registry. April 17. https://doi.org/10.1257/rct.4099-1.0

Former Citation

Benning, Jamie et al. 2019. "Rural Drinking Water Habits, Information Provision, and Pollution Exposure." AEA RCT Registry. April 17. https://www.socialscienceregistry.org/trials/4099/history/45089

Sponsors & Partners

Experimental Details

Interventions

Intervention(s)

We will select up to fifteen rural Iowa counties based on their (a) having a large number of potential households that rely on private wells; and, (b) being in areas susceptible to nitrate leaching. We will obtain and digitize plat maps from all counties to construct our initial sample.
We will send an initial survey to all households. The survey will include an incentive payment of $2/survey to maximize response rates. All surveys will be followed with a reminder postcard and a second survey to non-respondent households. This follows the best practices in the survey design literature (Dillman et al. 2009).

In the second phase, we will randomly select half of the responding households from the initial survey and send them in-home, nitrate test kits. The test kits will include contact information for local officials to obtain more precise water testing opportunities.

All households that receive the initial test kits will receive a reminder of best practices regarding water quality testing. We will also ask households if they would like to order test strips and their willingness to pay for more strips. A sub-sample will have an option to order test strips, while the remaining households will only state how much they would hypothetically pay for strips if offered the opportunity to purchase them. We will also conduct three listening sessions with 30 to 40 of the respondents to gather qualitative data to enhance the survey findings.

The last phase of the program will re-survey all homes that responded to the first survey round. By surveying all households, we will test whether households that received the water testing kit systematically changed their water use habits. Responses from the pre-treatment survey will be used to control baseline differences in water use behavior and improve statistical precisions.

Intervention Start Date

2019-04-15

Intervention End Date

2020-05-31

Primary Outcomes

Primary Outcomes (end points)

Key outcomes of interest include testing whether households that receive nitrate test kits differentially respond to the follow-up survey with regarding their: (i) well water uses - drinking water, cooking, laundry, landscaping, showering; (ii) alternative water sources - bottled water, water coolers, others; (iii) water filter use; (iv) bottled water purchases; (v) well water testing; (vi) water quality concerns and perceptions; and (vii) participation in state-funded water quality testing programs.

Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)

Secondary Outcomes (explanation)

Experimental Design

Half of the households that respond to our initial survey will be selected to receive simple in-home, nitrate test kits. The test kits will include contact information for local officials to obtain more precise water testing opportunities.

Experimental Design Details

Randomization Method

All households that respond to our initial survey will be eligible to receive a nitrate test kits. Random number generators will be used to block-randomize households into treatment and control by county. Treatment and control households will be compared in each county based on household demographics (Education; Income block; Renters) collected in the first-round survey to ensure balance. The samples will be re-randomized if samples are unbalanced.

Randomization Unit

Randomization unit is the household.

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

16 Iowa counties

Sample size: planned number of observations

Around 18,000 households will be surveyed. Based on previous survey responses, we anticipate 9,000 households will respond to the initial survey.

Sample size (or number of clusters) by treatment arms

Half of the first-round respondents will be assigned to treatment. The anticipated treatment and control group sample size is 4,500 households.

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

Preliminary power calculations suggest an MDE around 5% for bottled water purchase responses, 12% for water filter use responses, and 4% for well-testing responses.

Supporting Documents and Materials

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

Request Information

IRB