ABSTRACTThe measurement of service quality has received significant attention from researchers in recent years. SERVQUAL, a 22 item scale proposed by Parasuraman, Zeithaml, and Berry (1988) has been considered to have limitations. As a consequence, alternative models have been proposed in the literature. This paper reports the findings of a study conducted to test a three component measurement model proposed by Carman (1990). The results show that inclusion of perceived importance ratings does not improve the explanatory power of the model substantially. The weak results obtained might be attributable to measurement problems regarding importance weights. Given the intuitive appeal of the model, the paper calls for further investigation in this area.