The Information Technology (I.T.) revolution is having a profound effect upon the way we collect, organise, store and ultimately examine archaeological data. Statistical methodology is related to how we try to understand and formalise the last of these processes. I shall argue in this paper that the norm in archaeology is continually to examine and re-examine the same data (rather than to collect and examine it only in order to test a single hypothesis), a process which the I.T. revolution will serve only to emphasize and enhance. Classical statistical approaches, which underlie all commonly used statistical tests, are quite inappropriate in these circumstances. The Bayesian approach to statistical inference, although widely misunderstood in the past, appears to offer a far more relevant tool. Its closer examination may clarify considerably certain issues in archaeological methodology.