Investigating starting-point bias: a survey of willingness to pay for insecticide-treated nets

The objective of this study was to investigate the existence of starting-point bias in the bidding game contingent valuation elicitation technique when determining the willingness to pay (WTP) for insecticide-treated nets (ITNs) and ITNs re-treatment in rural Nigeria. Of all existing contingent techniques, the bidding game most closely mimics the normal price taking behaviour in local markets in Nigeria. Three different starting-points (low, medium and high) were used to determine WTP for large and small ITNs, and for ITNs re-treatment, respectively. The respondents were randomly assigned to any of the starting-points and a pre-tested interviewer-administered questionnaire used to elicit WTP. Non-parametric tests and the Tobit model were used to analyse the data for evidence of starting-point bias. Plots of respondents’ cumulative density functions by starting-points were also examined to show the pattern of responses. The non-parametric tests showed no statistically significant differences between the three starting points in WTP for large ITNs ( p=0.262) and for ITNs re-treatment ( p=0.412). However, there was a statistical significant difference in WTP for small ITNs ( p=0.045). Nevertheless, in this instance, the high starting point group had a lower mean WTP than the low group, and also had the lowest median WTP amongst the three groups. However, using the conditional WTP (only males), there were no differences among the three starting-points for all goods. The multiple regression analyses using the Tobit model confirmed the results of the non-parametric tests. The plots of cumulative densities were also similar for the three starting-points for the three products. However, the high starting-point group had those more willing to pay higher amounts for large and small nets. There was no conclusive evidence of starting-point bias. Future research is required in order to gain a deeper understanding on factors determining peoples’ valuation of goods and services, reasons for any type of starting-point bias, and how the bidding game can be improved.