Thus for averting expenditures as a whole in both the survey and experimental data, C is the total contribution to all three averting funds (TAVERT), whereas the overall expenditure is TWTP. The authors generated a new variable, LNAVERTODDS, to represent the log of this odds ratio.
As Table 6 shows, risk information, subjective risk, the existence of private insurance, amenity scores, efficacy scores, income, education, party affiliation, age, and household size had roughly the same direction of effect on participation and level of significance as they did on TWTP and averting share.
Other significant factors that positively influenced TWTP included risk attitude (H3), age, Republican Party affiliation, and income.
In contrast with TWTP, those who were informed that they fell within a high- or moderate-risk group significantly decreased the share of expenditures allocated to both private and public risk averting activities, whereas those informed of their low-risk group status significantly increased that share.
The authors predict that provision of risk information would result in an ordering effect that causes those located in high-risk zones to report higher TWTP for insurance and averting activities relative to those in lower risk zones.
The authors expected perceived efficacy of averting activities to influence participation rates, TWTP, and averting share in a positive manner in the survey setting.
Across settings, the authors found that income affected TWTP in a positive manner.