Attitudes and Information Effects in Contingent Valuation of Natural Resources

Abstract

This thesis investigated the effects of photographic and text information on respondents' attitudes and willingness-to-pay for a proposed beach protection scheme in the erosion prone Gold Coast region on the east coast of Australia. The research developed two alternative expectancy-value attitude-behaviour models to test residents' attitudes toward relevant targets and behavioural intention, expressed through stated willingness-to-pay, and compared the proposed models with one established attitude-behaviour model. The thesis set out to investigate three central research questions; one question relating to the effects of information on attitudes and willingness-to-pay, and two questions relating to the relationships between attitudes and willingness-to-pay. It was hypothesised that photographs that depicted severe erosion damage would result in more positive attitudes toward, and greater willingness-to-pay for, beach protection than photographs that showed only mild levels of erosion damage. Positive relationships were hypothesised between variables representing attitudes toward beach erosion, attitude toward beach protection, attitude toward paying for beach protection, and willingness-to-pay. Finally, it was hypothesised that the relationships between attitudes and willingness-to-pay could be adequately explained by the proposed attitude-behaviour models. The thesis describes how seven information treatments and eight attitude measurement scales were developed and tested in a pilot experiment before use in a survey of homeowners in the region of interest. Analysis of variance showed that, while respondent's attitude toward beach protection was affected by the information treatments, their willingness-to-pay for the proposed program was insensitive to information. There were no significant effects that could be attributed exclusively to text descriptions of the good but there were significant effects that could be attributed to photographic information treatments. However, none of the effects on attitudes resulted in significant effects on the behavioural intention expressed in stated willingness-to-pay. Analysis of respondents with low previous knowledge of the proposed good revealed more extensive information effects on attitudes, but still not on willingness-to-pay, and this suggests that high levels of previous knowledge in a large proportion of the sample had a moderating effect on attitude change caused by the information treatments. Regression analysis showed that seven of the eight attitude and behaviour variables in the proposed attitude-behaviour model were significant predictors of willingness-to-pay. In the final phase of the analysis, goodness-of-fit indices, estimated using Structural Equation Modelling, indicated a good fit between the data and the attitude-behaviour models tested. Standardised coefficients on the model indicated that perceived behavioural control, expected utility of outcomes, and subjective norms all had strong direct relationships with stated willingness-to-pay, and strong indirect relationships on willingness-to-pay via attitudes toward payment. These results are consistent with the relationships proposed in attitude-behaviour models and the moderating effects of these variables explain why significant information treatment effects were observed on attitude to beach protection but not on willingness-to-pay. This research showed that respondent's willingness-to-pay in a contingent valuation experiment is quite insensitive to photographic treatments when previous knowledge is high and that costly and time consuming testing procedures, recommended by authorities, may not be necessary under these conditions. It also demonstrated that measures of attitude, consistent with an attitude-behaviour model, can be collected easily in a contingent valuation study and can contribute to understanding of participant responses and to identification of protest responses.