8 - Not Just Nuisance: Spatialising Social Statistics

Summary

Introduction

This chapter is about treating space seriously within quantitative methods and thinking. It argues that geography is a fundamental characteristic of society, and of the social processes studied in the social sciences. However, the sorts of statistical techniques typically taught to social scientists, and used to inform social policy, either ignore the spatial dimension entirely or regard it as a nuisance. Neither of these acts of geographical short-sightedness is adequate, and both can result in misunderstanding about what is happening, where and why.

A few years ago, concern about these inadequacies might have been dismissed as part of a wider disinterest in (and distrust of) quantitative methods within the social sciences. However, this chapter is written at a time when statistical teaching has returned to prominence, reinvigorated by a resurgent interest in quantitative social science (broadly defined as the application of quantitative data, computation and statistics to probe areas of disciplinary and interdisciplinary interest). To a large degree (and to mutual relief), past antagonisms between quantitative and qualitative approaches have been set aside in favour of a shared interest in evidence-based knowledge, critical enquiry and empirically informed understanding of socio-spatial processes.

There is much to gain from this rapprochement. Reasoning with numbers is a vital skill for any student of the social sciences. Lynch makes the case, after Kant, that although we cannot defend scientific methods (including quantification) as more rational than other methods, we ‘can show that that they are more democratic, more respectful of basic human autonomy’ because ‘observation and logic are strategies that everyone can, at least to some extent, use themselves and employ in their social networks, and that can be made at least a little more effective with training’ (Lynch, 2016: 59–60, original emphasis).

Unfortunately, insofar as that includes statistical training, the tools and methods can act to disregard the very thing that motivates their usage – the study of a geographically varied and spatially embedded social world. If a key role of quantitative social science is to measure, to explain and to offer insight into the socio-spatial differentiations, inequalities and variations that characterise society then to do so is at odds with the sorts of spatially myopic methods that are characteristic of textbook introductions to statistics and applied policy analysis.