This post is based on a talk I gave at the annual British Sociological Association Conference in 2017. Going back to it now, I was surprised to see that it already contains many of the ideas I eventually developed and discuss at greater length in my monograph British Sociology and Statistics 1833-1979 (2020).
I start with a simple human story.
A benign act of theft
This story begins in the first half of the 19th century with a social theorist who’s regarded by many as the founding father of sociology, partly because he gave the subject its name – and for this alone we should be grateful to one Auguste Comte (1798-1857). But like children who’ve grown up and are thankful for all their parents have done for them but who no longer listen so closely to their parents’ advice, sociologists stopped listening to Comte’s ‘advice’ a long time ago.
But even our gratitude to Comte for giving us our name has been somewhat misdirected. If there’s anyone we really have to thank for the fact that, today, we call ourselves sociologists, it’s not Comte, but another, not so well-known contemporary of his. While Comte was busy tolling the bells for the phantom birth of a social science which, according to his grand theory, would take its deserved place above all sciences and study society using the inductive historical (or comparative) method, it was Adolphe Quetelet (1796-1874) who actually ‘gave birth’ to something which was eventually going to have profound consequences for the understanding of social phenomena. Quetelet was the first to collect masses of statistical data and to analyse them using the theory of probability for the purpose of investigating social phenomena like crime, education and poverty. To describe aggregations of data, he invented “the average man” – which is still today, one of the pillars of statistical analysis.
Quetelet borrowed the original term with which Comte described his social science – namely ‘social physics’ – and used it to describe his own statistical work. This apparently innocent act quite infuriated Comte who couldn’t bear to see his term being used for something which, as Comte’ himself put it, was ‘nothing more than simple statistics’, ‘an extravagance’, and ‘a fanciful mathematical theory of chances’. So, to avoid being associated with the statistical study of society, Comte quickly invented another term for his grand science – sociology. If it weren’t for Quetelet’s benign act of ‘theft’, sociologists today would be known as social physicists. ‘Social Physicists’ – personally, I kind of like the sound of it!
Of course, there’s a lot more to the historic development of social science than quarrels over names or labels, but for the purpose of our story, this episode serves to draw your attention to two things that are relevant to current debates.
The first is that even when social science was in its infancy and was very broadly defined, it was almost impossible, ridiculous even, to use one term, whatever this term maybe, to describe both the statistical investigation of social phenomena on the one hand; and theory building and ‘qualitative research’ on the other. Data from interviews I’ve conducted as part of my doctoral research showed that social statisticians, often feel uncomfortable with calling themselves ‘sociologists’; while many sociologists in the UK, in agreement with Comte, don’t recognise statistics as a sociological method.
The second is that this superficial quarrel over terms, which evolved into a long-term failure to reconcile statistics and sociology, was based, on a general lack of understanding, or misunderstanding among British sociologists of what statistics is. Something which, unfortunately, persists to this day.
Like chalk and cheese?
Comte and Quetelet are just two contrasting figures of many in the history of social science but, here in the UK, during the 19th and early 20th centuries, they came to represent two distinct groups. Comte’s sociology became the founding philosophy of the first British Sociological Society, established in 1903; while Quetelet helped establish the centre from which the Royal Statistical Society (RSS) emerged. In the late 19th century, other British investigators like Francis Galton, Karl Pearson and their colleagues developed further Quetelet’s advances, thereby establishing statistics as we know it today.
My archival research on the RSS and the groups of social statisticians around Galton and Pearson, and on the Sociological Society and its later affiliates, suggest that, the relationship between the statisticians and the sociologists was, from the very beginning, weak to non-existent; dim signs of hostility appear occasionally, but the period from the 1830s, when Comte and Quetelet worked, to the late 1930s, was characterized largely by an attitude of indifference on both sides.
Comte and Quetelet themselves can’t be said to have caused this rift between sociology and statistics in Britain. No, careful historical examination of the development of both subjects up to the Second World War tells us that British sociologists have had their particular reasons not to embrace statistics, and these, of course, need to be examined in their particular context. However, different individual decisions have together produced a striking regularity. It is even more striking that the social scientists at the forefront of sociology in the first half of the 20th century, L. T. Hobhouse, Morris Ginsberg and Patrick Geddes, distanced themselves from statistics and criticised it without themselves ever having acquired any statistical skills beyond the simple use of percentages.
British sociology in the post-war period
What about the period after the Second World War? There’s one big event that dominates this era – the government-driven institutional expansion of sociology. It was this sudden external expansion, not sociology’s slow internal growth, that transformed sociology into a popular academic subject. The number of students, degrees, teachers and monographs all grew phenomenally. This expansion posed quite a few challenges, most notably – if we don’t know what sociology is, in what direction do we expand it? Lack of clarity about the subject and methods of sociology is evident in the work of sociologists from the first half of he 20th century. And examination of early post-war papers on the state of sociology reveal, too, that there was still no clear sense of direction regarding its development.
This is most clearly to be seen in the sociology degree syllabuses, particularly those from the LSE – the university that contributed most to the popularisation and spread of sociology. The syllabuses show that from a subject that in the interwar period contained a mishmash of social philosophy, sociology turned in the 1950s-60s into…a mishmash of social philosophy, social theory, social administration and social anthropology. Today we have accepted this lack of coherence and co-ordination as benevolent ‘pluralism’, as a wilful creation, but there’s hardly anything in the archives that suggests that this was an intentional result. Even the early records of the British Sociological Association suggest a lack of clarity about what sociology is.
Methods in British sociology
Did the relationship between sociology and statistics change when sociology became a widely established academic subject? The answer is both yes and no. Every major official report on the state of sociology or social statistics published since 1946 has recommended a more effective and extensive incorporation of statistics into sociology. The same recommendation was repeated over and over again in the last 60 years, but to no effect. Details from the post-war history of academic sociology help to explain why.
Sociology’s expansion created some room for statistics teaching in sociology– general sociology syllabuses and methods courses syllabuses show that statistics was a consistent component in the sociology degree in the 1950s-60s. But many of these courses didn’t, or couldn’t have done enough to, teach statistical skills effectively and, sociology courses more generally, promoted a theoretical approach to studying and acquiring sociological knowledge.
In this period, however, statistics was the only methods course on the LSE syllabus – it wasn’t until the early 70s that other methods courses began to appear and not until the early 80s that sociologists began to pay closer attention to methodology teaching. The fact that statistics teaching had been, for a period, the only methods course on sociology syllabuses, explains partly why, even to this day, many sociologists believe that sociology had gone through ‘a positivist’ or overly statistical past in the 1950s-60s.
To this day among British sociologists, there’s a great deal of misunderstanding and, indeed, mythology surrounding both the historical and present day relationship (what might be termed a gap in understanding) between sociology and statistics; of how this gap came about and why it has persisted for so long. My book, British Sociology and Statistics 1833-1979, attempts to give a clearer, less prejudiced, view of how this divide came about and what bridges can and are being built in order to breech it. At the very least, sociologists aught to recognise that possessing some statistical skills is essential before any informed judgement can be made as to how extravagant, fanciful, or indeed, sociological, statistics actually is.