Results of Spline Modeling
It is possible that WTP for coastal adaptation outcomes may not be a linear function of the measure(s) of length of residency. That is, the impact of length of residency on the WTP for a particular attribute may increase or decrease as the level of length of residency changes. For example, the negative marginal change in the WTP for Homes associated with length of residency (as suggested by some of the previous results) might change, or even cease to exist beyond a specific threshold (or between specific thresholds) in the measure(s) of length of residency. Beyond a given break point, the marginal change in the WTP to protect homes may be positive and higher for longtime residents, who are more likely to have experienced …show more content…
This is achieved by using linear splines that allow for measures of length of residency to have different effects on the WTPs. For both measures of length of residency (years of residency and duration of ownership), I estimate separate models using two knots and one knot. The placement of the knots (for the two-knot models) was at S=9 and S=20, which creates three intervals between 0 and 77 for years of residency, and three intervals between 0 and 55 for years of property ownership. For the one-knot models, the placement of the knot was at S=20 years. In order to keep the number of observations within each segment relatively close, the choice of the knots was, to some extent, arbitrary. The estimated coefficients on the created variables (in all models) measure the slopes for the intervals rather than the change in the slopes from the preceding …show more content…
These results are reported in Table 3.14 and 3.15. Based on both measures of length of residency, estimates of the marginal change in the WTP to avoid the status quo (Neither) and to protect homes (Homes) are statistically significant for the first segment only (Residency-1 and Ownership-1). With the positive signs, these estimates indicate the marginal decline in the average WTP to avoid the status quo and to protect homes for each of the first twenty years of residency or property ownership. Beyond this break point, the estimated marginal changes in the WTP for these attributes is, statistically, not different from zero. These results are intuitive based on the fact that major flooding events (Nor’easter, Irene and Sandy) in the region occurred in recent years (2010, 2011 and 2012) that are contained in the first interval. Such events may heighten the perception of flood risks, and may subsequently cause WTPs to be
It is possible that WTP for coastal adaptation outcomes may not be a linear function of the measure(s) of length of residency. That is, the impact of length of residency on the WTP for a particular attribute may increase or decrease as the level of length of residency changes. For example, the negative marginal change in the WTP for Homes associated with length of residency (as suggested by some of the previous results) might change, or even cease to exist beyond a specific threshold (or between specific thresholds) in the measure(s) of length of residency. Beyond a given break point, the marginal change in the WTP to protect homes may be positive and higher for longtime residents, who are more likely to have experienced …show more content…
This is achieved by using linear splines that allow for measures of length of residency to have different effects on the WTPs. For both measures of length of residency (years of residency and duration of ownership), I estimate separate models using two knots and one knot. The placement of the knots (for the two-knot models) was at S=9 and S=20, which creates three intervals between 0 and 77 for years of residency, and three intervals between 0 and 55 for years of property ownership. For the one-knot models, the placement of the knot was at S=20 years. In order to keep the number of observations within each segment relatively close, the choice of the knots was, to some extent, arbitrary. The estimated coefficients on the created variables (in all models) measure the slopes for the intervals rather than the change in the slopes from the preceding …show more content…
These results are reported in Table 3.14 and 3.15. Based on both measures of length of residency, estimates of the marginal change in the WTP to avoid the status quo (Neither) and to protect homes (Homes) are statistically significant for the first segment only (Residency-1 and Ownership-1). With the positive signs, these estimates indicate the marginal decline in the average WTP to avoid the status quo and to protect homes for each of the first twenty years of residency or property ownership. Beyond this break point, the estimated marginal changes in the WTP for these attributes is, statistically, not different from zero. These results are intuitive based on the fact that major flooding events (Nor’easter, Irene and Sandy) in the region occurred in recent years (2010, 2011 and 2012) that are contained in the first interval. Such events may heighten the perception of flood risks, and may subsequently cause WTPs to be