On the History of a troubled Relationship: Sociology and Statistics in Britain Part I
Published: 10 November 2019
In this two-part blog entry Dr. Plamena Panayotova investigates the roots of the peculiarly distanced relationship of British Sociology and Statistics. In Part I, Plamena questions attitudes within Sociology that deem historical enquiry as superfluous. In fact, this anti-historicity is often a peculiar commonality of those disciplines which, otherwise, seem to have little in common, at least in Britain. But how, then does one study the historical genesis of this relationship?
Why knowing the history of your subject will help you be a better teacher
In this post I would like to tell you a little about the lessons to be learned about the teaching of statistics through the study of the historical relationship between statistics and sociology in the UK.
People often ask me ‘Why study history?’ Surely, we can do a lot to improve the teaching of statistics by focusing on pedagogical strategies, re-organising the curriculum, making better use of software. We don’t need knowledge of history to do this.
My doctoral research on the history of statistics and sociology in Britain aimed to prove how wrong this assumption is. We do need history. To explain why I often quote one historian of science who argued that we need knowledge of history because ‘if trained professionals do not provide an accurate picture of the past, then the terrible simplifiers will move in and provide distorted accounts to be used to support their own schemes and propagandas’ (Cannon, 1978).
Cannon, however, does not say anything distinct about history – one can give the same answer to questions regarding the study of statistics or economics or any other field dealing with human relations in which there is wide scope for interpretation. And the field of human knowledge is so vast that we have to rely on simplified information all the time with regard to things we’re not ‘experts’ on.
Despite its flaws, Cannon’s defence of history is useful in reminding us that the problem of ‘terrible simplification’ is fundamental and widely spread. We need knowledge of history to avoid and challenge simplified and distorted views and to build a detailed understanding of the subjects we teach.
Neither sociology, nor statistics have avoided the traps of ‘the terrible simplifiers’.
Colleagues in sociology might have come across the novel The History Man by Malcolm Bradbury which was popular in Britain in the 70s and 80s. Its main character, Howard Kirk, is a simplified model of British sociology: Kirk was a renowned radical and a complacent socialite, indulging in abstract theory, corporeal passions and intrigue. He abhorred ‘facts’.
Colleagues in statistics might recognise the character of Thomas Gradgrind from Charles Dickens’ Hard Times which is statistics simplified - Gradgrind is a hard-headed and emotionless fact-cruncher; for him people are little more than calculating machines.
These two characters couldn’t have less in common. Put the two together and you have a simplified version of the relationship between sociology and statistics in the UK, as many nowadays see it.
It turns out that this simplified model of the relationship between sociology and statistics isn’t as far from the truth as you might expect when it comes to the historical relationship between sociology and statistics in Britain – there is plenty of evidence, which I analyse in my work, testifying to the lack of co-operation and the unproductive relationship between the two subjects. This started with the reluctance of the first sociological institutions to embrace the nineteenth century tradition of statistical enquiry and continued very much unchanged as sociology expanded in the post-WWII period.
However, despite the fact that it points in the right direction, this model is not helpful for understanding the relationship between sociology and statistics, and how teaching practices affected or were affected by this relationship. To draw any lessons about the teaching of statistics to social science students, we need to go beyond these models. Analysing data usually helps.
In my study, I analysed data on the kind of skills, including statistical skills, that were taught in sociology at the London School of Economics and Political Science (LSE) in the twentieth century (the LSE was the only place up to 1945 offering a sociology degree and continued to exert a powerful influence in the post-war period on the formation of sociology degrees elsewhere). The data showed that teaching theoretical, not statistical, skills was the priority in British sociology. But this is not the whole story.
Two collections of sociology methods courses by Peel (1968) and Wakeford (1979) allowed me to compare the teaching of skills within methods courses right across universities in the UK between 1967-78 – a formative period in British sociology when, in the aftermath of a great academic expansion, sociology’s character took up its present shape. The analysis showed three important things. First, more than a half of sociology methods courses taught some statistics; and about a quarter were devoted especially to stats. Second, advanced statistical training was also present – nearly half of those courses teaching stats, offered advanced training. Use of computing also increased. Third, nearly all universities teaching sociology taught surveys; about a half taught some advanced elements of statistical data collection/analysis and more than two thirds taught stats at more elementary level. So the very least we can say based on these data is that the teaching of quantitative methods and surveys was consistently present in this crucial period (for more details, see Panayotova, 2019).
Does this mean then that the image of sociology as a generally non-statistical subject does not hold? If not, where did this image come from? When we examine the intellectual context in which statistics teaching took place, especially in the post-war period, it becomes clear that there are several important factors that compromised the effectiveness of this teaching.
References
Cannon, S.F. (1978). Science in Culture: The Early Victorian Period. New York: Dawson, Folkestone and Science History Publications.
Panayotova, P. 2019. ‘Realities and perceptions of methodological teaching and debates in post-war British sociology: new evidence from Peel (1968) and Wakeford (1979)’, Sociology 53(5): 826-842.
Peel, J. 1968. Courses Mainly Concerned with Sociological Theory and Methods in 29 Universities. London: British Sociological Association, mimeo.
Wakeford, J. 1979. Research Methods Syllabuses in Sociology Departments in the United Kingdom (undergraduate courses). Lancaster: Department of Sociology, The University of Lancaster, mimeo.
First published: 10 November 2019