This study examines whether making question formats better fit the cultural context of markets would improve the construct validity of estimates of willingness to pay (WTP). WTP for insecticide-treated mosquito nets was elicited using the bidding game, binary with follow-up (BWFU), and a novel structured haggling technique (SH) that mimicked price taking in market places in the study area. The results show that different question formats generated different distributions of WTP. Following a comparison of alternative models for each question format, construct validity was compared using the most consistently appropriate model across question formats for the positive WTP values, in this case, ordinary least squares. Three criteria (the number of statistically significant explanatory variables that had the anticipated sign, the value of the adjusted R(2), and the proportion that were statistically significant With the anticipated sign) used to assess the relative performance of each question format indicated that SH performed best and BWFU worst. However, differences in the levels of income, education, and percentage of household heads responding to the different question formats across the samples complicate this conclusion. Hence, the results suggest that the SH technique is worthy of further investigation and use.