The contingent valuation method was used to estimate the willingness to pay of the inhabitants of the town of Neuchâtel (Switzerland) to reduce road traffic noise. In a survey representative of the local population, respondents were asked how much they would pay to halve the traffic noise exposure level. During the regression analysis of the survey results special attention was paid to cope with two major bias sources: the impact of large bids and the problem of bid refusal resulting from strategic behaviour. The variation of WTPs can be explained by a range of variables in accordance with those suggested by theory. The average reported monthly bid is 70 Swiss francs, while the average WTP estimated varies from Sfr. 67 to Sfr. 56 depending on the model used. These results are similar to those of hedonic and contingent studies carried out previously in Switzerland, providing evidence of the external validity of the CVM estimations.