Handling Protest Responses in Contingent Valuation Surveys

Area of Science:

• Health Economics
• Environmental Economics
• Econometrics

Background:

• Protest responses in contingent valuation (CV) studies, where individuals refuse to state their value for health gains, are often excluded from analyses.
• Excluding protesters can introduce bias if they differ from non-protesters in ways that predict their responses.
• The Heckman selection model, commonly used to adjust for protesters, may have implausible assumptions in this context.

Purpose of the Study:

• To present and evaluate a multiple imputation (MI) approach for addressing protest responses in CV studies.
• To compare the performance of MI with the Heckman selection model.
• To estimate mean willingness-to-pay (WTP) for a Quality Adjusted Life Year (QALY) gain using the EuroVaQ study data.

Main Methods:

• A simulation study was conducted using data from the multinational EuroVaQ study, assessing MI and Heckman models under various realistic scenarios.
• Scenarios included different proportions of missing data and non-response mechanisms.
• The methods were illustrated using the EuroVaQ study to estimate mean WTP for a QALY gain.

Main Results:

• MI demonstrated lower bias and mean squared error compared to the Heckman approach across all simulated scenarios.
• The Heckman approach led to significant under- or overestimation of mean WTP, potentially due to violations of the normality assumption.
• The case study indicated that protesters have a lower mean WTP for a QALY gain, with results varying based on the method used to handle protesters.

Conclusions:

• Multiple imputation (MI) is an appropriate and more accurate method for handling protest responses in contingent valuation (CV) studies.
• MI provides more reliable estimates of willingness-to-pay (WTP) compared to the Heckman selection model.
• The findings highlight the importance of appropriate statistical methods for analyzing CV data with protest responses.