Validity and reliability of the contingent valuation method : a study of willingness to pay for insecticide-treated nets in Nigeria
Objectives: To contribute to knowledge on the reliability and validity of the contingent valuation method (CVM) and explore the role of context-specific CVM question formats in Southeast Nigeria. Other objectives were to determine the factors that will explain actual willingness to pay (WTP) for insecticide-treated nets (ITNs). Methods: There was an extensive review of theoretical, methodological and empirical literature. A novel WTP question format that mimics price-taking behaviour in south- east Nigeria (called the structured haggling technique (SH)) was developed and compared with the bidding game (BG) and binary with follow-up technique (BWFU). The comparisons were for inter-rater and test-retest reliability, content, construct and criterion validity and the study conducted in three villages in Nigeria. Stated WTP was determined using a questionnaire administered to 810 household heads, while actual WTP was evaluated by offering the ITNs for sale to all respondents after one month of the first survey. Findings: There were considerable gaps in the literature regarding the reliability and validity of the CVM. In the empirical study, BG, BWFU and SH elicited reliable and valid estimates of WTP. The SH was the most content valid, while the BG and SH were the most construct-valid for ITNs and re-treatment respectively. The BG and SH were similarly criterion-valid while the BWFU was the least criterion-valid. All question formats were similar for tests of reliability. There were genuine reasons for divergences between the stated and actual WTP and for test and retest. Low-income status and physical accessibility were the major impediments to ITNs acquisition. Conclusion: The CVM could be used to elicit valid and reliable WTP estimates in the study area, but it was not clearly proven that better content-valid question formats would lead to more valid and reliable estimates of WTP. It is necessary to further determine how the validity and reliability of the SH and other WTP question formats could be improved. Finally, future studies should establish the content validity of question formats in settings where they will be used, and use bigger sample sizes, along with allowing less time between the survey and administering the criterion, for comparing stated and actual WTP.