Public Perception of Agricultural Pollution and Gulf of Mexico Hypoxia

Abstract

Hypoxia is a prevalent problem is the Gulf of Mexico, and non-point-source pollution is believed to be the primary cause. Because agriculture is believed to be the leading single contributor to non-point source pollution, mitigating agricultural pollution is important. This study analyzes public perceptions about Gulf of Mexico hypoxia and addresses public willingness to pay to implement agricultural technology to decrease agricultural pollution to reduce the severity of hypoxia in the Gulf of Mexico. Results suggest that public awareness of the problem is low, but the public does express some willingness to pay to mitigate pollution.

Keywords:

• agricultural pollution
• hypoxia
• public perception
• willingness to pay

Notes

This research was funded by the Advanced Spatial Technologies and Agriculture grant program, Mississippi State University.

1. While random digit dialing mitigates biases generated by unlisted numbers and the birthday screening question mitigates biases generated by tendencies in who answers the phone, this, like all other surveys, is subject to self-selection and non-response biases. As is shown below, the resulting sample appears representative, and this technique has been shown to generate no item non-response bias (Hudson et al., 2004)

2. The 95% response rate refers to the percentage of people who agreed to participate in the survey after screening for age. This number does not include fax machines, answering machines, disconnected numbers, or instances where there were no eligible respondents available to answer the questionnaire.

^aApproximated from Census data for those over 18 years of age.

^bApproximated from Census data based on percentages completing high school and college.

3. For this analysis, we use regional designations defined by the Census. Southeast: Mississippi, Alabama, Florida, Georgia, Arkansas, Louisiana, Tennessee, North Carolina, and South Carolina. Northeast: Kentucky, Virginia, West Virginia, Maryland, Delaware, Pennsylvania, New Jersey, Massachusetts, Rhode Island, New York, Connecticut, Vermont, New Hampshire, and Maine. Midwest: Ohio, Indiana, Michigan, Wisconsin, Illinois, Missouri, Kansas, Nebraska, Colorado, North Dakota, South Dakota, Wyoming, Montana, Minnesota, and Iowa. Southwest: Oklahoma, Texas, New Mexico, Arizona, California, Nevada, and Utah. Northwest: Oregon, Washington, and Idaho.

^aPercentages are row percentages. That is, percentages should be read across each row.

^*Statistically significant at the 0.10 level.

^aRegional dummy variables = 1 if respondent is from that region, = 0 if otherwise; Midwest is the base region.

^bSmaller than 0.0000 in absolute value.

^cEthnicity dummy variables = 1 if respondent is of that ethnicity, = 0 if otherwise; all other ethnicities besides White and African American are the base.

4. Specifically, the use of square miles or kilometers was most confusing to the focus group. These focus group participants asked questions about the relative size compared to some defined geographic area, leading to the wording used in the question.

5. For example, a respondent would be told that if all farmers in the region employed precision agriculture, it would be expected that a 5% reduction in agricultural runoff would be generated, resulting in a 4% reduction in the size of the hypoxic zone.

6. It may seem somewhat irrelevant to ask respondents about technical relationships. The objective of the question is not to ascertain accuracy. Rather, the objective of the question is to see whether respondents were sensitive to different program outcomes. It would be more difficult to convince taxpayers to support a program they do not view as realistic.

^aRegional dummy variables = 1 if respondent is from that region, = 0 if otherwise; Midwest is the base region.

^bSmaller than 0.0000 in absolute value.

^cEthnicity dummy variables = 1 if respondent is of that ethnicity, = 0 if otherwise; all other ethnicities besides white and African American are the base.

^aRegional dummy variables = 1 if respondent is from that region, = 0 if otherwise; Midwest is the base region.

^bSmaller than 0.0000 in absolute value.

^cEthnicity dummy variables = 1 if respondent is of that ethnicity, = 0 if otherwise; all other ethnicities besides white and African American are the base.

7. It should be noted that this WTP analysis is nonparametric and does not estimate differences in WTP based on sociodemographic factors. The purpose of this analysis is to provide a general picture of the average WTP, not to analyze differences on the basis of location or sociodemographic variables. For an example of a more complete analysis on a smaller geographic sample, see Hite, Hudson, and Intarapapong (2002).