Two techniques are currently used to evaluate the humoral immune responses to rabies vaccination: ELISA, which detects binding antibodies to viral antigens and the WHO reference rapid fluorescent focus inhibition test (RFFIT), which assays in vitro virus-neutralizing antibodies. In this study, we have comparatively evaluated antibody responses of dogs reared either in an experimental kennel or living in field conditions after vaccination with a cell culture-derived rabies vaccine. In experimental conditions, both ELISA and RFFIT techniques were well correlated. However, in field conditions, they yielded discrepant results particularly in evaluating the residual rabies immunity before vaccine administration and in identifying seroconverted dogs. After rabies vaccination in field conditions, while similar antibody titres and seroconversion rates were obtained using either technique, the discrimination of a given dog according to the seroconversion threshold depended on the assay. We concluded, that whereas in experimental conditions, ELISA and RFFIT were well correlated, in field conditions ELISA yielded upper estimates. Consequently, RFFIT, although a cumbersome test, should continue to be considered as the reference rabies antibody assay technique. A seroconversion threshold of 0·5 IU/ml should be cautiously considered and a higher threshold (1 IU/ml) could be more appropriate in the evaluation of rabies immunity in the field in order to marginalize the interfering factors.