Figure - uploaded by Grant Gardner
Content may be subject to copyright.
Contexts in source publication
Context 1
... also estimated a linear probability model of participation, where we regressed an indicator for survey response on average annual water use, total aquifer depletion from 2000 to 2019, a dummy variable for whether there were two or more filers for the annual water use reports to DWR during 2000-2019, and GMD dummies (see Table A4 in the supplementary appendix). Net of the GMD effects, we do not detect any evidence of average water use, aquifer depletion, or change in filer being a statistically significant predictor of survey participation. ...Context 2
... somewhat surprisingly, irrigators indicating "slightly concerned" use the least amount of water per year on average. Table 4 shows pairwise comparisons amongst all the concern categories. Using pairwise comparisons, we find that the differences in marginal effects between "slightly" and all other categories are statistically significant at 0.10 or better. ...Context 3
... somewhat surprisingly, the coefficient on "very concerned" is the smallest in magnitude relative to no concern. Examining Table 4, the pairwise difference between "very concerned" and all other levels of concern is statistically significant at 0.05 or better. The pairwise differences between the remaining categories for total acreage are not statistically significant. ...Context 4
... the pattern of acreage reduction is not necessarily monotonic across concern categories. Using pairwise comparison of the average marginal effects in Table 4, we find statistically significant differences between "slightly concerned" and "moderately concerned" and "very concerned" at 0.01 or better. Likewise, the pairwise difference between "somewhat concerned" and "moderately concerned" and "very concerned" is statistically significant at 0.01. ...Context 5
... findings support Hypothesis 3 that concern correlates with a history of lower water use. The pairwise difference in marginal effects are presented in Table 4. The difference between "slightly concerned" and "somewhat concerned" and "very concerned" is statistically significant at 0.01. ...Context 6
... next to total irrigated acreage, we again detect a pattern where irrigated acreage levels decline as the level of concern increases, supporting Hypothesis 1. Specifically, irrigators indicating "slightly," "somewhat," "moderately," or "very" concerned irrigate about 18.5, 51.1, 34.5, and 59.1 fewer acres per year than irrigators indicating no concern, respectively. Again, the reduction in acreage across categories of concern is not necessarily monotonic (Table 4). For example, the "very concerned" category has the coefficient of largest absolute magnitude but is not statistically different from the "somewhat concerned" category. ...Context 7
... next at corn irrigated acreage, we observe that irrigators indicating "slightly," "somewhat," "moderately," or "very" concerned irrigate about 29.1, 37.4, 25.6, and 50.7 fewer acres of corn per year than irrigators indicating no concern, respectively. These findings support Hypothesis 2. Using pairwise comparison of the average marginal effects, we find that irrigators indicating "very concerned" tend to irrigate the fewest number of corn acres overall, with the difference between all other groups statistically significant at 0.01 (Table 4). Corn provides the highest net returns to irrigated production but is also the most irrigation intensive. ...
