Exploiting the differences in supported transmission rates between mobile users, opportunistic scheduling promises a substantial increase of the aggregate throughput of wireless networks. In this paper, we present a Markov model to study the trade-off between fairness and wireless efficiency of opportunistic scheduling at an access point which serves multiple mobile users. The Markov process at hand tracks the queues of outstanding packets for the different mobile users as well as the states of the wireless channels for these users. Because the size of the state space of the Markov model prevents a direct solution, we develop a numerical analysis technique based on Maclaurin series expansions to solve the system in light traffic and in overload. By numerical examples we illustrate the accuracy of our approach and compare a set of performance metrics of various schedulers. In particular, we study how cross-channel correlation affects the performance of these schedulers.