ABSTRACT
The need to reassess quantitative geography has become only too apparent in recent years as the range of techniques and strategies to be covered has grown. The days of courses consisting mainly of descriptive measures, graphical devices, a number of non-parametric techniques for under-measured data, and culminating in correlation and regression have long past. New avenues of work have been established which have significant bearing on quantitative teaching. One of these is the modification of existing statistical techniques and procedures to spatial data. A second is the incorporation of categorical models given that geography is overflowing with such data. A third is the need to introduce the idea of robustness so that students are sensitive to the fragility of many statistical techniques. A fourth is the need to computerise quantitative training so that students have the technical apparatus and experience to work with the large data holdings stored in geographical information systems and continent-wide data bases.
