“As you know, many people have been killing their wives these days. Do you happen to have killed yours?” is an example of the “everybody” approach to the problem of asking embarrassing the question in surveys. In the “numbered card” approach, a respondent would be asked “what has become of your wife?” and handed a card with the options “1) Natural death; 2) I killed her; 3) Other (what?),” from which they could state their response by number. The recurring problems of which questions should be asked, how they should be put to respondents, and how to make order out of the answers lie at the heart of Kerstin Brückweh's book. The hypothetical “wife-killing” example, drawn originally from a 1950s article on public opinion research, is one of many illuminating and occasionally idiosyncratic materials she brings to bear in her engaging study of census and survey research in Britain from 1801 to the present.

The time period alone makes the book unusual within the modestly sized field of sociological and historical research on statistics and statisticians. The long time frame starts with the first British census of 1801, little more than a headcount—conducted, as in most nations at the time, through the intermediary of parish worthies of one sort or another. The study continues where many historians fear to tread, right up to the present. It covers the most recent census of 2011, which was completed by individual respondents, some online, who answered the greatest number of questions ever—and carried out just as the future existence of a decennial census was coming under more serious threat than ever before.

Within the specific literature on statistics in Britain the study’s extensive coverage of the later 20^th century is especially welcome. The part played by statisticians and statistical data in the Victorian enthusiasm for public health and social reform has been the subject of several important works, as have the statistical ideas developed from the turn of the 20^th century by British or British-connected statisticians, some of whose names—Pearson, Fisher—will be familiar from undergraduate quantitative methods courses. Little has been written, however, on statistical practice in Britain since 1945, the era of the modern welfare state, the end of empire and Commonwealth immigration and, from 1979, of Thatcherism. Existing monographs covering this period were mostly written by professional statisticians, and many are institutional Festschriften. Though these include much useful historical data on institutions and enquiries, they leave largely unexamined the self-understandings of official statistics and the mechanics of the production of statistical knowledge.

Brückweh, by contrast, is centrally interested in the “shop-floor of social fact-making,” a phrase she takes from Lutz Raphael’s influential paper on the scientisation (Verwissenschaftlichung) of the social. She makes judicious use of the existing histories just described, but also presents extensive original research in professional and commercial periodicals, institutional repositories and the National Archives, especially the latter’s holdings from the General Register Office. The object of her research is Umfrageforschung, “survey research,” although the term has a considerably broader meaning than its English translation. The book investigates two “fields of application” of survey research, population censuses and commercial market research, and in so doing yokes together a rather unlikely pair.

A population census represents official statistics in their most state-ly, universal and obligatory, general and ponderous, while for market research speed, specificity and expediency are the rule. The two face common problems: how to ask embarrassing questions, for example, or how to recruit and train a field staff of interviewers. Census data, with their inherent models and categories, were an essential source in market research, as in many other fields of applied knowledge-production; this is shown most clearly in the book’s discussion of social-class classifications. Brückweh argues that census and market research have increasingly diverged since the 1970s, but beyond the fairly unsurprising commonalities there is not much that suggests lively interchange between the two fields before then. Official sample surveys investigating specific social problems and circumstances, to which governments after 1945 turned with enthusiasm, represent the most plausible middle-ground for exchange between census and consumer research, but the book touches on these only in passing. At the end of the book one is left with the impression not of two aspects of a common field but of two activities so different in purpose and organisation that they have not all that much to do with one another.

This, as well as the fact that the book is thematically, rather than chronologically, organised, makes for the occasional surprising switch in narrative focus. It does not detract, however, from the quality and interest of the material presented. The first part looks at the means necessary for the production of a census—personnel and questionnaires—and conveys better than any other study of census-taking I have seen the sheer scale of such operations. The chapter draws out the complex division of labour, and the importance of specialist knowledge about, for example, the formulation of questions, the design of questionnaires and the conduct of interviews, which were formalised as scientific methodology much later than mathematical methods of statistical analysis. This section also demonstrates how much the expanded scope of the British census is a recent phenomenon, made possible in large part by the progressive computerisation of processing, production and latterly collection.

In the second part, the book turns to the development of social-class and spatial classifications in Umfrageforschung in the later 20^th century. The historicisation of statistical classifications is somewhat of a staple in the historical sociology of statistics. The doyen and best-known exponent of the field, Alain Desrosières, whose final posthumous collection Prouver et Gouverner was reviewed in this journal in 2015, made his name in part through his work on Les Catégories socio-professionnelles. Brückweh makes a worthy contribution in this tradition, arguing persuasively that these two axes of classification have played a dominant role in the picture of society produced by British census and consumer research in the 20^th century. It is in this part that the coupling of subject matter proves most effective, drawing out the contrast between consumer researchers’ interests in classification—crudely, whatever identifies people to whom things can be sold—and the orientation of successive census classifications to sociological conceptions of employment relations and conditions of occupations, which came directly from Goldthorpe’s work. She further shows how the two lines of classification in consumer research have been not so much orthogonal but rather mutually substitutable, with geodemographics playing an increasingly central role in the segmentation of markets. She is, I think, on the side of the majority of social scientists in criticising commercial geodemographics for its theoretical fallacies, methodological opacity and its wearingly alliterative, value-laden epithets (“prudent pensioners”). Her criticism is given extra sharpness here by her having encountered first-hand the secrecy of the industry in the course of her research.

The third part of the book contains two studies of the controversies surrounding the inclusion and form of questions on specific topics in the census. The first looks at “disability”, and the inclusion of questions about deafness, blindness and later lunacy and idiocy from 1851 to 1911. Of particular interest is the finding that inclusion and continuation of these questions was justified explicitly in reference to best practices on the continent and the recommendations of the international Statistical Congresses. The latter have often been seen as important to the construction of disciplinary and professional identity, but having only little bearing on the actual practice of national statistical bureaux in the later 19^th century. The second case study, of the development of questions about place of birth, race and ethnicity, like the first shows how diverse actors, from scientists to interest groups, have engaged in sometimes violent disagreements about what questions should be asked. Brückweh handles this “new” political history of decisions about censuses as deftly as she does the more sociological history of the earlier parts of her study.

In her concluding remarks, Brückweh uses her findings as a “history of present-day problems” in reflecting on the current predicament of the British census. Ministers of the incoming government in 2010 openly voiced their scepticism about the need for a full decennial census, and parliamentary committees and the Office for National Statistics have been duly assessing alternatives such as the use of registration data. Brückweh argues that the capacities of the 21^st century state have met their limits, in the form of private data banks and data-protection legislation. Official statistics are an essential part of the state’s monopoly of information about persons and things in society, and part of its power from at least the early modern era. This monopoly, though never absolute, is rivalled today as never before by the vast trans-national databases accumulated by a small number of internet firms. These firms are relatively unhindered in their ability to link data with one another, whereas statistical offices face the hurdles of law, administration and popular consent.

I am less certain about the limiting effect of data-protection in Britain, as compared to Germany, to which the book’s conclusions sketch a comparison. Germany has a more developed administrative framework for data-protection and a sophisticated public sensibility to the issue, as evidenced in the mid-1980s controversy in West Germany over the census, in which official statistics faced an existential threat to their constitutional basis and popular consent. If this German context does not apply unconditionally in Britain, it is perhaps not so much because of the limits of state capacity as because of the limits of the state’s ambition. The will to collect information presupposes the will to act upon it. This is apparent, for example, in the fact that British official statistics have created neighbourhood classifications, not wholly unlike the geodemographics that Brückweh describes, during sporadic periods of enthusiasm for spatially targeted policy interventions. However, since 1980 state strategies for managing society and the economy have gradually turned away from planning and intervention. This shift, which occurred more abruptly in Britain than in many other European states, has altered the demand for statistical information. Even the near future, as Brückweh concludes, seems open and uncertain, and her book will be not only of interest to scholars of statistics and the history of sociography but also to those engaged in debates on current statistical practice, on which it provides a valuable historical perspective.