*2.1. Data*

The data were from a survey of irrigators in Mississippi conducted by the Survey Research Laboratory at the Social Science Research Center at Mississippi State University [23]. A telephone-based survey secured a total of 148 completed interviews in Mississippi from a total 2216 telephone numbers acquired (861 were disconnected or inaccessible after 10 attempts) with an overall cooperation rate of 27.6 percent. The sample is representative of the Delta area of Mississippi with 131 respondents residing in the area and 3 more in neighboring counties. The survey instrument contained questions on growers' characteristics, cultural practices, irrigation management practices, and perceptions and attitudes regarding groundwater availability.

Except for irrigated area, years of education, and pumping cost, the variables included in the analysis were coded as categorical or indicator (dummy) variables. The pumping cost is calculated as [24]:

$$p = \theta\_{\ell} p\_{\ell} d\_{\ell} \tag{1}$$

where *pe* is the price of the energy source for the power unit, *d* is depth to water used as a proxy for pumping lift, and *θ<sup>e</sup>* is the amount of energy from source *e* needed to lift a cubic meter of water a distance of one meter. The distance to water was obtained from the U.S. Geological Survey [25] based on the respondent's claimed county of residence. Average energy prices were obtained from the U.S. Energy Information Administration (EIA).

Cracking soils is the percentage of soils with clay content dominated by smectite. Such soils crack on the surface when a moist soil shrinks due to drying. The data on soil composition is accessible through the USDA-NRCS Websoil survey [26]. This is an important control variable. For example, surge irrigation programs are more difficult to manage in this type of soil because the programming of the alternating cycles is more complicated. The typical program relies on visual cues to switch from

one side of the field to the next based on the time water takes to reach the tail of the section. In cracking soils, water infiltrates through the cracks, and actual wetting occurs by 3 or more meters ahead of the wetting on the surface. The program is still applicable and carries the same water savings potential, but becomes harder for the farmer to realize.

Number of years of education was calculated based on a question that was originally categorical. The assigned values were as follows: 10 for less than completed high school, 12 for completed H.S., 13 for some college or vocational program, 14 for completed Associate's degree, 16 for completed Bachelor's, 18 for completed Master's, and 20 for more than Master's. This transformation is helpful in the estimation and interpretation of regression results with a sample that is small relative to the number of variables considered.

The variable GW problem is a dummy variable based on the combination of categorical responses to two different questions in the survey: "In your opinion, do you have a groundwater shortage problem on your farm?" and "In your opinion, do you have a groundwater shortage problem in your state?" Lastly, conservation program is a dummy variable based on the combination of responses to four different questions that would have otherwise yield 19 response categories (see Appendix A).
