Rebekah R Heinzen1, John F P Bridges. 1. Johns Hopkins Bloomberg School of Public Health, 615 North Wolfe Street, Baltimore, MD, USA. rheinzen@jhsph.edu
Abstract
OBJECTIVES: To compare four contingent valuation elicitation methods as a means to estimate the value of a pneumococcal vaccine in Bangladesh and to test if the elicitation methods are subject to bias and if they produce valid responses. METHODS: Three hundred sixty-one households with at least one child under 5 years of age were recruited in Dhaka, Bangladesh. Subjects were cluster-randomized to various elicitation methods: open-ended, dichotomous choice (at one of two asking prices), payment card (one of two cards with differing ranges). The dichotomous choice method was then followed up with a bidding game methodology, with the dichotomous choice price acting as the starting price for the bidding game. Analysis focused on summary statistics, demand curve estimation and multivariate regression analysis to test for validity and bias. RESULTS: Thirty-one households refused to participate, leaving a total of 330 participating households (a 91.4 percent response rate). Willingness to pay estimates varied significantly across the methods (p < .001), with average estimates varying between $2.34 and $18 (US). The open-ended elicitation method was found to produce highly inflated values that were insensitive to construct validity tests. The dichotomous choice method produced quantity (demand) estimates rather than value estimates, and there was some evidence of yea-saying. The payment card elicitation method was found to be affected by range bias. The bidding game elicitation method was found to be less sensitive to starting point bias and yea-saying. CONCLUSIONS: Different elicitation format do give rise to different demand curves; however, this may be partially due to the fact that they do not measure the same outcome. For example, the dichotomous choice format produces a demand curve, while the payment card, open-ended and bidding game produce inverse demand curves. All formats are prone to multiple biases. When choosing an elicitation format, it is important to first consider the purpose and use of the data. Each elicitation method has strengths and weaknesses and can be used for different purposes in technology assessment.
OBJECTIVES: To compare four contingent valuation elicitation methods as a means to estimate the value of a pneumococcal vaccine in Bangladesh and to test if the elicitation methods are subject to bias and if they produce valid responses. METHODS: Three hundred sixty-one households with at least one child under 5 years of age were recruited in Dhaka, Bangladesh. Subjects were cluster-randomized to various elicitation methods: open-ended, dichotomous choice (at one of two asking prices), payment card (one of two cards with differing ranges). The dichotomous choice method was then followed up with a bidding game methodology, with the dichotomous choice price acting as the starting price for the bidding game. Analysis focused on summary statistics, demand curve estimation and multivariate regression analysis to test for validity and bias. RESULTS: Thirty-one households refused to participate, leaving a total of 330 participating households (a 91.4 percent response rate). Willingness to pay estimates varied significantly across the methods (p < .001), with average estimates varying between $2.34 and $18 (US). The open-ended elicitation method was found to produce highly inflated values that were insensitive to construct validity tests. The dichotomous choice method produced quantity (demand) estimates rather than value estimates, and there was some evidence of yea-saying. The payment card elicitation method was found to be affected by range bias. The bidding game elicitation method was found to be less sensitive to starting point bias and yea-saying. CONCLUSIONS: Different elicitation format do give rise to different demand curves; however, this may be partially due to the fact that they do not measure the same outcome. For example, the dichotomous choice format produces a demand curve, while the payment card, open-ended and bidding game produce inverse demand curves. All formats are prone to multiple biases. When choosing an elicitation format, it is important to first consider the purpose and use of the data. Each elicitation method has strengths and weaknesses and can be used for different purposes in technology assessment.
Authors: Adrienne F G Alayli-Goebbels; Job van Exel; André J H A Ament; Nanne K de Vries; Sandra D M Bot; Johan L Severens Journal: Health Expect Date: 2014-08-18 Impact factor: 3.377
Authors: Nancy L Schoenborn; Ellen M Janssen; Cynthia Boyd; John F P Bridges; Antonio C Wolff; Qian-Li Xue; Craig E Pollack Journal: Ann Fam Med Date: 2018-11 Impact factor: 5.166
Authors: Stella Nalukwago Settumba; Marian Shanahan; Georgina M Chambers; Peter Schofield; Tony Butler Journal: BMJ Open Date: 2019-03-23 Impact factor: 2.692
Authors: Shaphan Y Chia; John Macharia; Gracious M Diiro; Menale Kassie; Sunday Ekesi; Joop J A van Loon; Marcel Dicke; Chrysantus M Tanga Journal: PLoS One Date: 2020-03-25 Impact factor: 3.240