Introduction

STATISTICAL METHODS ARE NOW an accepted, indeed expected, part of a geographer’s training, and most higher education programmes will include some elements of this important area of study. Regrettably, for many students this can be a painful and sometimes unrewarding encounter. Problems may arise from a lack of any comprehensive background in numerical analysis; a problem too often compounded by a seemingly bewildering notational system of unfamiliar Greek symbols. Further problems arise from the disillusion that many undergraduates find when courses and texts venture no further than univariate or bivariate methods, yet it is all too apparent that many geographical issues are multivariate in character, with the result that statistical training may not match geographical insight or analytical ambition.