An increasing number of stated preference surveys are now carried out on the Internet, yet only few studies have compared the results from web-based surveys with results from other survey modes. The main objective of the present paper is to examine the use of Internet as a survey mode for a CVM study aimed at valuing a gain in life expectancy in the context of air pollution. Convergent validity of the web-based survey mode is examined against face-to-face interviews with respect to differences in socio-demographic characteristics, non-response bias, and differences in willingness-to-pay-related parameters. The general result is that the mean and median willingness-to-pay estimates are statistically indistinguishable across the two survey modes while differences in other willingness-to-pay related parameters exist across the two survey samples.