Conditional analysis techniques have been applied to observed and modelled concentrations from a 13-month field study where monitors were deployed at different distances from an interstate highway in Las Vegas, USA. The ADMS Urban dispersion model reproduces the general decline of concentrations with distance from the roadway but tends to over-predict the decline under very stable conditions. The model reproduces the general characteristics of observed wake effects at 10m, but does not reproduce an apparent continuation of observed wake effects to greater distances. Overall, the Las Vegas data set makes a significant contribution to information and understanding of near road dispersion and prediction, and it warrants further studies to test model predictions in more detail and to explore uncertainties. The study has also demonstrated the value of conditional analysis when summarising observational data for purposes of model evaluation