The household forms an important category in social science research. It is used to collect data, to classify it and to represent the results. In 2009, for example, 3.48 % of U.K. households were classified as the most influential and wealthiest individuals in the U.K.; at the other end 5.16 % of U.K. households formed the most disadvantaged people. While the households in the first group represented positions of power in the private and public sector and could afford luxury items, the latter had a budget to only cope with the daily necessities.1 Endless examples of household-based statistics or of classifications with associated patterns of spending and behaviour could be added. However, what seems to be a simple listing of facts becomes less clear when a basic question is raised: What is a household? Is it a family living under one roof? Is a roof limited to a house, or does a flat already constitute a household? Do members of a household have to be officially related, meaning married, adopted etc., or even related by blood? And how do households and definitions of households differ over time and space? In 2008, the United Nations distinguished between two definitions of a household: the housekeeping concept and the dwelling concept. While the first referred to arrangements made by persons, individually or in groups, for providing themselves with food and other essentials for living, the latter regarded all persons living in a housing unit as belonging to the same household.2 This sounds like a pragmatic solution with little regard to the social relationships of the actual human beings living in a household. However, there are indeed power relations within a household (e.g. between parents and children). Social scientists also observed these everyday asymmetries and therefore constructed a hierarchy in social classifications when they placed the household in a specific class according to the ›Head of Household‹ or the ›Household Reference Person‹, the ›Chief Wage Earner‹ the ›Householder‹ etc. The different designations of the reference person indicate that it is not an easy task to name this person or to define this person without a normative bias. The principles and recommendations put forward by the United Nations in 2008 give a clue about the normative assumptions that are at the basis of possible definitions: »Even in the many countries where the traditional concept of head of household is still relevant, it is important to recognize that the procedures followed in applying the concept may distort the true picture, particularly with regard to female heads of households. The most common assumption that can distort the facts is that no woman can be the head of any household that also contains an adult male«.3 The United Kingdom is one of these nations with a long tradition of allowing only men (meaning members of the male sex) as ›Heads of Household‹.
By taking the example of Great Britain,4 I want to demonstrate that the definition of the ›Head of Household‹ was a normative category rather than a descriptive one, meaning that it was less able to facilitate analysis of social reality and that it fortified a normative view with the help of statistics. In contrast to the British practice, an open definition – as already in use in other countries in the second half of the 20th century – included self-descriptions or self-assessments of historical actors and changes in society as perceived by members of the society rather than through social scientists’ categories. While feminists and other historical actors in different states already criticised the normative bias of the definition in the 1960s and 1970s, a different question seems to be of equal or even greater importance to the historian: How, when and why did different nations and professions decide to drop the normative in favour of a descriptive definition of the ›Head of Household‹? This leads to a more general question: How did administrators, statisticians and other survey researchers deal with the aim of long-term stability of statistical categories for the sake of comparability, e.g. in a national census, on the one hand, and with adaption to societal change on the other hand? In taking the example of the United Kingdom, the following story combines aspects of a history of knowledge with administrative history.5 After embedding the administrative practice in assumptions from the history of knowledge (I.), the article continues by reconstructing the official definition of the ›Head of Household‹ in British government social research (II.), followed by an analysis of knowledge production about a household in other branches of survey research in the U.K. (III.). Subsequently a summary of the topic in U.S. history will be given (IV.) in order to compare it to the different path taken in the U.K. (V.). This is followed by some conclusions (VI.).
I Knowledge Production and the ›Head of Household‹
By combining the history of knowledge with administrative history, I seek to explore »the dialectic of tension between categories and contexts«, as Simon Szreter, Hania Sholkamy, and A. Dharmalingam have suggested.6 Starting with the assumption that categories like the ›Head of Household‹ are socially constructed as well as »ideologically and politically charged acts of representation and intervention«,7 consideration is also given to the administrative dimension: in order to administrate, it is necessary to build categories for reducing the complexity of social life. This imperative could be illustrated by the notion of drawing a map at a scale of 1:1, which is absurd.8 We have to rely on abstractions and generalisations, and for this reason we are forced to make a choice among several options. While statisticians, survey researchers and administrators in one nation chose to drop the normative category of the ›Head of Household‹, at the same time the same professions in other nations did not. In accepting the use of categories as necessary for reasons of administration and knowledge production, it is also necessary to recognise and reflect on the ›intellectual opportunity costs‹ that come with them and sometimes bring a need to adapt and change categories or their underlying definitions.9
Some assumptions of the history of knowledge are helpful to pursue this objective of a critical assessment. First, there is no differentiation between information understood as real facts versus knowledge as something that is processed; and consequently there is no single knowledge, but »knowledges« in the plural.10 In the case of the ›Head of Household‹ this means that definitions are made by men, and therefore produce specific knowledges in relation to the chosen definition. Other definitions would produce – and in fact did produce, as will be seen below – other knowledges.11 Second, the different cultures of knowledges include practices, methods, assumptions, ways of organising and teaching etc.12 Classifying is one of these practices and, again, classifications may seem natural, but they are the results of ordering systems made by men. The ›Head of Household‹ at the core of this article is one example for a hierarchic order at the basis of classification. This takes us to a third point, inspired by Foucault and his writings on power as well as by postcolonial studies: knowledges are not equal. Or, phrased differently: knowledge regimes are embedded in power relations.13 Even if historical actors know that their assumptions are normative, as in the case of the ›Head of Household‹ it does not mean that those in favour of dropping the normative definition have the power to actually do so. Here the context and a related fourth point come into play: the idea of situated knowledge. It was already brought forward by Karl Mannheim, who belonged to the first wave of the sociology of knowledge.14 In the 1920s and 1930s sociologists thought about social factors that influenced the production of knowledge in a specific context. Mannheim suggested his idea of the social position and existential connectedness (›Seinsverbundenheit‹) of knowledge, meaning that spatial and social processes and settings influence the cognitive process of individual thinking.15 Aspects of this can be recognized in works of later sociologists of knowledge who formed a second wave in the history of knowledge. Amongst them were Pierre Bourdieu, Michel Foucault, postcolonial thinkers and protagonists of a renewed feminism. The story of this article is situated in the changes following ›1968‹, which involved, especially in the case of feminism, the questioning of the academic system and an awareness of gendered ways of knowing.16 In this context the definition of the ›Head of Household‹ came under scrutiny. A fifth and final assumption from the history of knowledge that is important for this story concerns the idea that circulation shapes knowledge. Knowledge circulates between different historical actors with different agendas and thus forms an important element in producing knowledges.17 In the case of the ›Head of Household‹ it was one specific tool that circulated between different historical actors and produced knowledge.
This specific tool leads us to the administrative component of the story: knowledge about a household was usually gained through a questionnaire or a form that was either part of an interview or was left to a household to be filled in. This ›little tool of knowledge‹ formed a basic element of the bureaucratic practice meant to facilitate, rationalise and standardise communication between citizens or consumers on the one hand and an authority, such as the census office or opinion pollsters and market researchers, on the other hand.18 Usually the definition of the ›Head of Household‹ was not given on the questionnaire. This missing piece of communication might be one reason for the long normative story of the male ›Head of Household‹ in Britain. For the longest time during the 19th and 20th centuries, it seemed to be tacit knowledge who was considered to be the ›Head of Household‹ i.e. the husband. However, changing gender perceptions had an impact in the last decades of the 20th century, when actors within survey research institutions felt the need to define the term in order to carry on using it. How exactly was the category defined?
II The ›Head of Household‹ in British Government Social Research
In a household containing only husband, wife and children under 16 (and boarders) the husband is always the HOH.
Similarly, when a couple have been recorded as living together/cohabiting the male partner should be treated as the HOH.19
In the year 1991 this definition was printed in the official Handbook for Interviewers published by the British Social Survey Division (SSD). The SSD was part of the Office of Population Censuses and Surveys (OPCS) that performed tasks including taking the census of population and the registration of births, marriages and deaths. The Handbook for Interviewers presented the SSD to its readers as the official social survey organisation within central government responsible for collecting »information about the circumstances, conditions, behaviour and attitudes of members of the population or parts of it«.20 It was first published in 1950 and revised several times,21 nevertheless, the definition that gave precedence to the male ›Head of Household‹ remained. The definition was important for the interviewer at the doorstep when he or (mostly) she22 had to find the correct person to obtain data about. So-called informants could change due to the aim of each survey, but if the ›Head of Household‹ was the major informant, the definition given above was the relevant one. The idea of naming a male ›Head of Household‹ and subordinating all other members to him indicated a clear understanding of the order of families and sexes. As can be seen in a caricature printed in the English magazine Punch, this was already being criticised during the census of 1851 (see fig. 1).
In 1851 the English census enquired about the ›Head of Family‹ for the first time. From the census of 1921 onward the word ›family‹ was substituted by ›household‹, otherwise everything seemed to be almost the same from 1851 until the 1990s. While suffragettes used the census of 1911 to call attention to their missing right to vote and therefore also pointed to the biased census questionnaire,23 neither they nor the second wave of feminism in the 1970s had an immediate impact on redefining the ›Head of Household‹ Up to the 1970s the stability of the definition might have well fit the majority of self-assessments of different members of society. However, there is a need to explain the reasons why the definition was not changed in the 1970s, but kept for the following decades. It was only for the census in 2001 that the normative and biased definition was abandoned. While feminists and actors of other social movements of the 1970s were successful in some of their goals, as can be seen in the Equal Pay Act of 1970 and the Sex Discrimination Act of 1975, it is an open question why the male-oriented definition of the ›Head of Household‹ was so persistent.
The first reason for the persistence of the definition seems to be that it was not easy to find. The Handbook for Interviewers stated in this regard that definitions »will not normally be printed on your questionnaires so you must learn them and apply them whenever you collect classification information«.24 Looking at the different questionnaires of the decennial population censuses of the years 1981 and 1991, it could even be argued that these forms suggested an equal understanding of the sexes. In both censuses the questionnaires were addressed »To the Head or Joint Heads or members of a Household« and a ›1st person‹ could be named to whom all the other members of the household were related (see fig. 2).
While the possibility to name a ›1st person‹ already existed in 1971, the examples in the 1971 questionnaire clearly suggested that the husband should be taken as the ›Head of Household‹: »Write ›HEAD‹ for the ›Head of Household‹ and relationship to the head for each of the other persons: for example ›Wife‹, ›Son‹, Daughter-in-law‹ ›Visitor‹, ›Boarder‹, ›Paying Guest‹.«25 In contrast, the questionnaires of 1981 and 1991 could lead to the conclusion that it was the decision of the different members of a household to name a head (see fig. 2). Despite this impression the Handbook for Interviewers of the year 1991 clarified that in cases where more than one person equally claimed to be HOH, the following rules applied: »1. where they are of the same sex the oldest is HOH; 2. where they are of different sexes the male is HOH«.26 As a justification the Handbook explained: »These rules on deciding who is HOH are necessary because the use of joint heads of household is not practical for analysis purposes. Because of this it is necessary to have consistency in the way in which decisions are made«.27
This leads to the second reason for the persistence of the definition, which is long-term comparability of data. Due to the fact that the husband had always been the ›Head of Household‹, it remained thus: »a husband always takes precedence«.28 However, it was not impossible to adapt the definition to some societal changes: when a couple was not married but lived together, the male partner should be treated as the ›Head of Household‹. So even in cases where couples had made a deliberate decision not to get married with all its consequences (and possible political impacts or rather political statements), nevertheless they were treated as if they were married.29 From the perspective of social scientists this was a logical act because cohabitation could be seen as a new phenomenon that needed to be customised to time-series statistics – and this was a pragmatic way to achieve it.
A third and related but slightly different reason for the persistence is connected to a pragmatic way of methodologically handling the problematic definition. In a paper in the government Survey Methodological Bulletin Jean Martin and Jeremy Barton discussed the possible effects of a change of definition und stated in 1996: »Criticisms of the existing definition do not in themselves make the case for change; any alternative definition must be shown to improve sufficiently on the current one to justify a change, particularly in terms of discontinuities in time series«.30 For this reason, in the 1990s survey researchers in the government social survey and in other branches of survey research started to empirically test the effects of possible changes of definitions. Martin and Barton, for example, came to the result that dropping the biased definition would lead to a change in ›Heads of Household‹ or ›Household Reference Persons‹ in about 15 % of all households.31 While this seems to be a significant number, the authors stated that, before deciding whether to adopt a new definition for the ›Household Reference Person‹, various users of the main government household surveys in commerce, NGOs, public administration and other areas of usage should be consulted. This hesitant handling resulted in a parallel use of different categories and definitions, the ›Household Reference Person‹ and the ›Head of Household‹ being two of them.
A fourth reason for the persistence may be due to the administrative and staff structure of the government social survey branch, which was dominated by male social scientists – at least this seems to be true for the 19th century and the first half of the 20th century. The historian Edward Higgs made this argument as regards 19th-century statistics on women, occupations and work in general:
The figures in the occupational tables are not ›hard facts‹ or ›raw data‹. They were constructed by men (in the specific sense of members of the male gender) who had certain assumptions about the position of women in society. In broad terms, women tended to be defined as dependants, whatever their productive functions, whilst men were classified according to the nature of their labour. As a consequence there are serious problems in reconstructing the role of women in the reproduction of society in the nineteenth century. This has important implications for models of the economy as a whole.32
While Higgs’s point is persuasive for the 19th century, the setting had changed during the course of the 20th century. The Handbook for Interviewers, for example, was edited and written by a woman, and Gillian Theresa Banks was the first woman to hold the position of Registrar-General, in the years 1986 to 1990.33 It seems less productive to ascribe questions of gender equality to the sex of the social scientists than to focus on the basic conflict of long-term comparability versus adaptation to social reality. Although I will come back to the gendered interests later, for now I would like to look more closely at the abolishment of the definition.
The definition of the ›Head of Household‹ was dropped for the population census of 2001. However, this was not the case for all surveys conducted under the aegis of the Office of Population Censuses and Surveys, which became part of the newly created Office for National Statistics in 1995. When I asked the Office for National Statistics in 2012 if the old definition of the ›Head of Household‹ had been dropped or if it was no longer defined in favor of the male sex, they answered with reference to the Labour Force Survey – a longitudinal survey first conducted in 1973. For this survey the ›Household Reference Person‹ was given as the major category that was defined regardless of sex.
The household reference person is the householder, which is the household member who owns the accommodation; or is legally responsible for the rent; or occupies the accommodation as reward of their employment, or through some relationship to its owner who is not a member of the household. If there are joint householders, the one with the highest income is the household reference person. If their income is the same, then the eldest one is the household reference person.34
In addition, it was stated that the ›Household Reference Person‹ was introduced into the Labour Force Survey in 2001, in line with other ONS household surveys, to replace the ›Head of Household‹35 Nevertheless, the definition of the ›Head of Household‹ still existed in the Labour Force Survey Guide of 2008:
If there are two adults of the opposite sex living together as a married or cohabiting couple, the husband/male partner is the HoH. Otherwise, the oldest male householder, or the husband/male partner of the oldest female householder, is the HoH. Otherwise, the oldest female householder is the HoH.36
The Labour Force Survey Guide stressed that information about the ›Head of Household‹ was still collected, for example in the Labour Force Survey, but that information about the ›Household Reference Person‹ was also collected and mainly used instead. The authors of the Labour Force Survey Guide seemed to feel a need to justify the ongoing collection of data about the ›Head of Household‹ They argued that in most cases the male adult in a household would be given as the ›Head of Household‹ anyway: »Most HoHs are male, because in households comprising either a mixed-sex couple or joint householders of the opposite sex, the male partner/householder is classified as the HoH, regardless of their income and age.«37 The change in definitions led to about 10 per cent of households being classified differently.38 This result indicates that different definitions produce different kinds of knowledges. However, were 10 % to 15 % (as found by Martin and Barton) of households classified differently due to the new definition significant enough to incorporate the change? The assessment of this question did not remain within the inner circle of government social scientists. It could not remain there because citizens had to answer the questions of an interviewer or to fill in the questionnaire. To do so interviewees had to translate aspects of their own lives into the given form of a questionnaire. If their perceptions differed too much from the prescribed categories they did not fill in the forms. Thus, interviewees and interviewers were part of a network of actors who together produced knowledge. This knowledge was the result of a circulation process between various actors and in this process the definition was increasingly criticised at the doorstep in the 1990s. This point was especially stressed by survey researchers outside the government, who also were part of the circulating production of knowledge. They were more dependent on the cooperation of citizens and consumers in conducting their surveys and satisfying their clients’ needs. This seemed to be especially important for the 1990s, when declines in survey participation rates, the growth of alternative modes of data collection, and the increase of data from digital systems (especially from the Internet) became significant.39
III The ›Head of Household‹ in Other Branches of Survey Research in the U.K
Since at least the 1970s survey researchers outside the Office of Population Censuses and Surveys (OPCS) were well aware of the problematic definition of the ›Head of Household‹. In their basic classification system, the so-called Social Grades, households were allocated to a specific social class or grade depending on the data taken about the ›Head of Household‹.40 For the purpose of the Social Grades, up to the 1990s the ›Head of Household‹ was defined as follows:
The head of household is that member of the household who either owns the accommodation or is responsible for the rent, or, if the accommodation is rent free, the person who is responsible for the household having it rent free. If this person is a married woman whose husband is a member of the household, then the husband is counted as the ›head of household‹.41
In exactly the same way as in the OPCS (or later the ONS), survey researchers in the private sector also put forward a gendered hierarchy with the husband as the head of the household. In doing so, they knew that they were not in line with some historical actors, but they did it anyway. This conclusion could be drawn from an information brochure published by the major professional organisation in Britain, The Market Research Society, in 1974:
Here the usual convention is to identify the ›chief wage earner‹, or the person responsible for the accommodation (legal owner, or person paying the rent), and where there are two equal candidates, to ignore women’s lib‹ [liberation, KB] and take the male as head of household, and the older of two people if both are of the same sex.42
Although there were different labels in use, the governmental and the non-governmental social surveys used the same biased definitions and procedures of identifying the ›Head of Household‹. This did not change within the next decade. In 1984, a revised version of the information brochure pointed out that the majority of women of working age now held a job, and therefore the ›Head of Household‹ might be less unambiguous to name. However, they still stuck to the old definition. The paragraph in the brochure dealing with this issue stated: »[…] and where there are two equal candidates, to ignore ›feminism‹ and take the male as head of household.«43 Similar statements could be found in handbooks for survey researchers of the time. For example, Gerald Hoinville and Roger Jowell of the research institute Social and Community Planning Research (SCPR) wrote in 1982: »Some of these rules are increasingly under attack on the grounds of their male chauvinistic bias and the fact that they are anachronistic. […] Nevertheless, current practice is still to define the head of household as the husband.«44 Two main arguments for keeping the old-fashioned definition were invoked: first, this system had been in use since the 1950s, and second, no better method had been found yet.45 Similar to the rationale in the government sector, the long-term comparability and especially the well-established practices were brought forward in favour of this definition.
In 1992 this practice ended. Asked about the cause, the director of a prominent market research institute answered: »We found that wives were getting upset«.46 Surely, this was only one of several reasons, nevertheless an important one. Due to the circulating process of knowledge production, market researchers needed men and women to answer their questionnaires, which proved to be difficult using a biased definition. After a period of continuous growth of survey research, the new developments of the 1990s led survey researchers to revise their definition and to do so more quickly than their colleagues in the government sector who – with a mandatory census every ten years – had other ways of persuading citizens to participate. Moreover, with diverse clients such as supermarkets, manufacturers, political parties and so on, survey researchers in the private sector had to respond more quickly to sinking participation rates in order to be less »out of touch with modern life and marketing needs«.47
At the same time they had to rethink another major category of consumer research: the ›Housewife‹. For most of the 20th century, the ›Housewife‹ was a knowable social type upon whom manufacturers and advertisers could act.48 In 1931, Paul Redmayne and Hugh Week published the first British book solely devoted to market research.49 As regards the place and method of purchase they were very sure that »for quite a large range of products no doubt arises at all as to the type of shop where it is bought«.50 Where there was doubt about the basic question of who consumes the product at what times and how, according to Redmayne and Week, »the investigator will naturally interview the housewife who both buys and uses the products«.51 The ›Housewife‹ as a category was introduced to target buyers of fast-moving consumer goods. It also came to be seen as old-fashioned by the 1990s. Parallel to the man-based definition of the ›Head of Household‹ commonly the ›Housewife‹ was restricted to women. While the ›Head of Household‹ was replaced by the ›Chief Income Earner‹ in the early 1990s, the ›House-wife‹ was replaced by the ›Shopper‹.52 In 1994, when Erhard Meier of Research Services Ltd. analysed the effects of the new definitions, he found that the ›Shopper‹ definition led more men to categorise themselves as shoppers, but overall the results remained very close to those of the old ›Housewife‹. As regards the changing definition of the ›Head of Household‹ or ›Chief Income Earners‹, Meier found that 90 % of all households remained in the same Social Grade.53 He also came to the conclusion that the effect of the new definition was particularly strong amongst women: About 15 % of women were now classified as ›Chief Income Earner‹ who would have been classified as ›Wife of Head of Household‹ under the old definition.54 The newspaper The Independent started its report about the changes with the intriguing example of a female chief executive of a major chemical company who was classified as part of a working-class household because her live-in boyfriend worked as a garage mechanic.55 This short digression on the ›Housewife‹ as an important figure in survey research can be linked to one of the assumptions of the history of knowledge, that is, the relation between power and knowledge. While the ›Head of Household‹ seems to be a desirable powerful position, being classified as a ›Housewife‹ – although important to survey clients, e.g. in consumer research – seemed to be less desirable. In social classification, being a ›Housewife‹ has no representation in the social hierarchy on its own, but only by the husband. From the assumption that women’s work in all its facets (including unpaid work) should also be represented in social classifications came one line of arguments that let to feminist mobilisation against the male-biased ›Head of Household‹.
IV Feminist Mobilisation in the 1970s: The Case of the United States
Different nations took different paths in dealing with the household as a statistical unit. The United States of America and the United Kingdom were amongst those countries that relied on the category of the ›Head of Household‹ to denote relationships within a household. Compared to the U.S. history, the U.K. was relatively slow in abandoning the male-biased definition. Up to the U.S. census of 1970, the ›Head of Household‹ was defined much like in Britain, that is, a married woman could not be named as ›Head of Household‹. With the resurgence of feminism in the 1960s and 1970s, the category was considered to be offensive to many people. In the late 1990s, the sociologist and demographer Harriet B. Presser recalled her version of abandoning the male-dominated definition.56 According to her account, one of the first questions she was confronted with when joining the faculty at the University of Maryland in 1976 seemed to be a simple one: »What do demographers mean by the term ›head of household‹?« Barbara Bergmann, then Professor of Economics at Maryland and a member of the Census Advisory Committee of the American Economic Association, brought this question to her with a hidden agenda that Presser subsequently supported: the »decapitation« – as Presser called it – of the ›Head of Household‹ in the U.S. census.
Already in 1790 with the first U.S. decennial census, the designation of a family head was established.57 The practice continued through 1840. In the censuses of 1850 to 1870 no head was asked for, but names of everyone living in a household were noted. From 1880 to 1930 all family members were listed in the census and their relationship to the ›Head of Family‹ was also required. From 1940 through 1970 the term ›Head of Family‹ was replaced by ›Head of Household‹. The story so far seems to be comparable to the British history, but what follows is different. In the U.S., a small group of social scientists, including Presser and Bergmann, campaigned successfully to stop the Census Bureau from using the category in the future. In the context of the resurgence of feminism in the early and mid-1970s, feminists had begun to express their dismay over the use of the ›Head of Household‹ definition by the Census Bureau. Similar to the British context, the U.S. Census Bureau did not provide a formal definition for the ›Head of Household‹ in the questionnaire, rather it asked to put down »the person who is regarded as the head by the members of the household«.58 However, this was only seemingly an open definition: if a married woman or her husband were to report the woman as ›Head of Household‹, the Census Bureau in its administrative process changed it to the husband. Against this background, feminists objected to the definition because it implied an authority structure for households and families that was imputed by the Bureau rather than measured – and feminists found that it was not the business of a federal census to do either.59 The Census Bureau responded to this accusation by conducting tests about people’s views as to who they thought was ›Head of Household‹, While the Bureau found the definition to be justified, feminists did not. The first meeting between Bergmann, Presser and some other demographers, economists, psychologists and sociologists took place in early November 1976. All of them wanted to clarify some kind of a role conflict: »As feminists, we were sympathetic to the objections raised about its [the male-based ›Head of Household‹ definition, KB] continued use; as researchers, we were concerned as to whether there would be an analytic ›cost‹ to dropping the concept«.60 In their first meeting the group decided to formally organise as Social Scientists in Population Research (SSPR) and to arrange a meeting with Census Bureau professionals.61 From this exchange it was learned that not all countries used a ›Head of Household‹ designation in their censuses and that some were in the process of change. This change could involve abandoning the definition completely (e.g. to name everyone in a household without a relationship to a designated head), or revising it to an unbiased version. The Census Bureau argued that some of their users were strongly in favour of a head – although without the sexist bias.62 An unbiased version was already recommended by the United Nations for the 1970 Population Censuses: »The head of the household is that person in the household who is acknowledged as such by the other household members«.63 In addition, Sweden was given as an example of a country that had never used the term ›Head of Household‹, simply asking for a reference person, while Canada was about to change from ›household head‹ to ›person 1‹.64 After intensive discussion the members of the SSPR came to the conclusion that they were in favour of abolishing the ›Head of Household‹ definition altogether. In their subsequent lobbying efforts, they were supported by the Census Advisory Committee of the American Economic Association, of which Barbara Bergmann was a member. The Committee passed a resolution on December 3, 1976, stating:
We believe the term ›head of household‹ in the questionnaire and in the public tabulations is ambiguous, not currently descriptive of many households, and offends numbers of people. Elicitation of information needed by users of the Census is feasible through other methods. We urge that the Census Bureau make the change in time for the 1980 Census.65
The SSPR used this resolution for their own lobbying efforts which included disseminating their view through several professional bodies such as the American Psychological Association, the American Public Health Association, the American Sociological Association, the American Statistical Association etc. They also warned that if the Census Bureau did not demonstrate a serious commitment to assessing alternatives and making appropriate changes, this would lead to organised noncompliance with the 1980 Census.66 With the help of a letter-writing campaign, the support of congresswoman Patricia Schroeder, who was at the time Chair of the U.S. House Subcommittee on Population and the Census, and the assistance of prominent voices such as Arthur S. Flemming, Chair of the U.S. Commission on Civil Rights, they were successful in their efforts. According to Harriet B. Presser’s version of the story, Schroeder’s visit with Census Bureau officials and her insistence that they drop the definition of the household head in the 1980 Census gave the final impulse.67 For the 1980 Census the Census Bureau chose to use the ›Householder‹ as the reference person to whom to relate other household members; they defined the ›Householder‹ as »the household member (or one of the household members) in whose name this house or apartment is owned, being bought, or rented. If there is no such person, start on line 1 with any adult household member«.68 With this new definition either spouse of a married-couple family could now be the ›Householder‹.
The U.S. case tells the story of successful mobilisation.69 The U.S. female sociologists organised to advocate their agenda from an early stage and were able to fruitfully connect to other (male) voices and politicians in their lobbying, as well as to mobilise others for their agenda (e.g. by letter-writing campaigns). British social scientists took a different path.
V Battles in Segmented Fields of Survey Research: The Case of Britain
The climate surrounding ›1968‹ also had an impact on feminist criticism of male-biased social research in 1960s and 1970s Britain. However, I argue that criticism of the definition of the ›Head of Household‹ coincided with an increased segmentation of the field of survey research in Britain. Survey research in Britain can be traced back to the 19th century.70 It expanded to different fields of application (amongst them market research) in the 1920s and continued to be an innovative methodological area until the 1950s. Researchers from different fields of application (including the Wartime Social Survey) worked together on discussing and improving their methods. They were also able to find employment in the different areas of application in the private and public sectors as well as in academia. The foundation of the Market Research Society in 1946 as the corporate body of all persons interested in »the furtherance of the profession of marketing and social research« can be seen as the organised effort to collaboratively promote survey methods.71 From the 1960s onwards, survey researchers in different areas became less interested in each other’s work. As a consequence, survey research in the government sector, in the private sector and in academia became less connected.72 Research in the private sector with a focus on opinion polling, media and the market had to survive in a competitive market, and thus tried to use data and methods found elsewhere. The public sector also increasingly became a market for them when cuts in government funding for social research initiated outsourcing of survey research in the Thatcher era.73 However, market researchers became less interested in discussing the theoretical presuppositions of their work: They took a pragmatic approach, making use of what was there and had worked so far. When they adopted new developments, e.g. the emerging geodemographics in the late 1970s, they treated their methods as business secrets.74 At the same time they were quite self-aware and self-critical of their pragmatic views and the possible consequences. John Samuels of the British Market Research Bureau, for example, wrote in 1988: »Market and media researchers are notoriously bad at ›reading the literature‹. We tend not to learn from each other’s experience, and each generation reinvents the wheel«.75 As regards different systems of classification for which the ›Head of Household‹ was considered to be important, Samuels argued:
Whilst our social grading system has no real theoretical base, there are now in existence systems which are based in theory, and where the allocation of individual occupations to groupings is both explicit and replicable. But virtually all market and media researchers are as ignorant of them as I was a few weeks ago. For reference, you might care to investigate the Goldthorpe/Casim scale and the Cambridge Stratification Scale. I say you might care, but you won’t.76
Recognising the segmentation of the different fields of application is essential for understanding the British story of the ›Head of Household‹, because it was precisely the academic sector surrounding John Goldthorpe that became the focus of feminist criticism. The academic journal Sociology served as the major battlefield in the 1980s.77 Feminist criticism had already started in the 1960s and had targeted theoretical ideas relating to social class and classification in sociology. It was criticised that social stratification studies took class as a major category, whereby classes were based on (male) occupation. The assumption that the class of a household was taken from the male ›Head of Household‹ as the supposed main breadwinner came under attack. This led the prominent sociologist, Rosemary Crompton, to remember in 1993 that »without exception, therefore, in Britain all of the major surveys in the area of class and stratification had, until the 1970s, drawn upon men-only samples«.78 The criticism was twofold: it was directed at the exclusion of women from empirical investigations and at the underlying assumptions upon which the identification of a class structure was itself gendered. The result was that effects of class and gender could not be disentangled within the structure of employment. Early critics of this amongst American and European sociologists used harsh words when they accused the existing methods of »intellectual sexism because it had neglected women, rested on sexist assumptions in so doing, was out-of-touch with the changes in women’s marital and employment statuses, had used male heads of households to represent family’s status and ignored sexual inequalities within the marriage«.79 The debate stretched across the 1980s, with criticism rebutted with new empirical evidence and vice versa.80 As regards the definition of the ›Head of Household‹, the most important outcome might be that no concordant voice or lobbying effort for the abolishment of the biased definition was formed. Unlike in the U.S., in the U.K. no univocal voice was formed among female academics; rather, there were multiple, separate controversial debates within survey research and within academia, with female academics taking different sides. In the late 1990s Catherine Hakim wrote about the treatment of one of her articles following a discussion between her and Rosemary Crompton and Fiona Harris: »This article was an attack on the victim feminism that is fashionable in academic circles and is reiterated by Crompton and Harris«.81 In the U.S. a group of female scientists took a feminist standpoint, formulated a common goal (»Decapitating the U.S. Census Bureau’s ›Head of Household‹«), chose a specific strategy when they named their professional organisation ›Social Scientists in Population Research‹ (instead of ›Feminist Social Scientists‹ or something similar) and were thus successful in lobbying their agenda within different professional organisations. In Britain, female scientists also had a place82 in sociology and the various ideas of feminism had an impact on British sociology, but – unlike in the U.S. – the abolishment of the male-biased ›Head of Household‹ did not seem to be a major unifying topic for them. They did not show an interest in forming a network of actors and lobbying the end of the male-biased definition.
Beware of the ›Head of Household‹. This seems to be one conclusion to be drawn from the story told in this article. Whoever works with statistics and surveys needs to closely scrutinise the basic definitions. While the male bias of the specific definition of the ›Head of Household‹ has been the focus of this article, it should also be mentioned that the definition had other problems and biases as well: for example, the various ideas of interviewees about what constituted a head of household (being in charge of household affairs, of the children, of the income etc.) or the predominance given to older people in a household (which was another normative assumption). But even if historical actors judged this definition as biased and not representative of changes in society, this did not result in a redefinition or an updated practice. Knowledge in theoretical discussions and knowledge in the administrative practice could differ greatly. ›Little tools of knowledge‹ like the questionnaire were necessary to collect so-called information about a household, and thus shaped practices and knowledges.
Knowledge production may also differ in time and space, as the different paths of the U.S. and the U.K. indicate. In the U.S. female scientists played an important role in opinion-making, but did not choose to act as an organisation that spoke for women. Instead they chose a specific goal, formed an organisation with a neutral name, and were thus able to mobilise politicians and other professional organisations for the abolishment of the biased definition in the 1970s. In the U.K., a missing univocal goal arguing for the end of the normative definition, as well as the segmented fields of survey research, led to a different history. For market researchers the pragmatic approach was most important: as soon as response rates were under threat and customers were no longer satisfied with categories, market researchers changed the definition in the 1990s. For the government social survey, it seemed to be a bundle of factors (amongst them transnational discussions of the topic and supranational definitions, e.g. by the UN) that finally led them to drop the biased definition in the 1990s. International discussions and practices had an impact,83 yet another phenomenon that might be best labelled »methodological nationalism«84 can also be observed: survey researchers were employed by international companies, travelled to international conferences and so on, but the major frame for their analysis was their own nation. Sarah E. Igo made a similar observation and related this to statistics as the science of the state as one root of social surveys: »One intriguing feature of most major early survey operations, even if not state– run, was that they imagined their scope to be national, rather than sub-national or international«.85 Transferring this to the history of the ›Head of Household‹ may be another clue for the different paths of the U.S. and the U.K. Feminist mobilisation against the male ›Head of Household‹ in the U.S. took place earlier than in the U.K., that is, before postmodernist and poststructuralist thinking became pre-eminent. In the 1980s, when sociologists in British academia discussed the issue intensively, they had to compete with other theoretical concepts and, in addition, with a »crisis which sociology faced as an academic discipline in Britain, as departments were ›rationalized‹ and subject to increasing economic pressures«, which, as Rosemary Crompton described it, led to a phase when »sociologists themselves underwent the (often painful) process of adapting to ›new times‹ «.86
The idea of ›methodological nationalism‹ also points to a different line of argument. In the U.K., gender issues were considered to be less important than class issues. With Marxism having a strong impact on sociological thinking and social work, gender was part of the larger issue of oppression, or, as the prominent British sociologist Jennifer Platt put it, »I had led a very sheltered middle-class life, and the glimpses this gave of how other people lived were a revelation to me, making class a live issue in a way it had not been to me before«.87 While in the United States research focussed on separate women’s culture, female institutions, the family and sexuality, in Britain labour history was much stronger and female sociologists also had a background of socialist politics with an emphasis on wage work, trade union organisation, and labour politics.88 This may explain why no unifying voice against the male ›Head of Household‹ was to be heard in the 1970s. However, class as a unifying category of all segments of British social survey also came under scrutiny in the 1980s and 1990s, when the arrival of geodemographics shifted the attention to neighbourhoods as another important category of survey research.89 Thus, only when times were changing so significantly that actors at all levels of the process of knowledge production, including interviewers and interviewees, became critical about the biased category, was it finally changed in the 1990s, but it remained, a well-established artefact, in the administrative realm for at least another decade.
ACKER, Joan: »Women and Social Stratification. A Case of Intellectual Sexism«, in: American Journal of Sociology 78 (1973), p. 936–945. ALLEN, Sheila: »Gender Inequality and Class Formation«, in: Anthony Giddens, Gavin Mackenzie (eds.), Social Class and the Division of Labour, Cambridge 1982, p. 137–147. ANDERSON, Margo J.: The American Census. A Social History, New Haven, London 1988. ATKINSON, Jean: A Handbook for Interviewers. A Manual for Government Social Survey Interviewing Staff, Describing Practice and Procedures on Structured Interviewing, London 21971. BECKER, Peter (ed.): Sprachvollzug im Amt. Kommunikation und Verwaltung im Europa des 19. und 20. Jahrhunderts, Bielefeld 2011. • — / CLARK, William (eds.): Little Tools of Knowledge. Historical Essays on Academic and Bureaucratic Practices, Ann Arbor 2001. • »Überlegungen zu einer Kulturgeschichte der Verwaltung«, in: Jahrbuch für europäische Verwaltungsgeschichte 15 (2003), p. 311–336. • »Formulare als ›Fließband‹ der Verwaltung? Zur Rationalisierung und Standardisierung von Kommunikationsbeziehungen«, in: Peter Collin, Klaus-Gert Lutterbeck (eds.), Eine intelligente Maschine? Handlungsorientierungen moderner Verwaltung (19./20. Jahrhundert), Baden-Baden 2009, p.281–298. BERGER, Peter L. / LUCKMANN, Thomas: Die gesellschaftliche Konstruktion der Wirklichkeit. Eine Theorie der Wissenssoziologie, Frankfurt am Main 1997. BLYTHE, Ian: The Making of an Industry. The Market Research Society 1946–1986. A History of Growing Achievement, London 1988. BRÜCKWEH, Kerstin: Menschen zählen. Wissensproduktion durch britische Volkszählungen und Umfragen vom 19. Jahrhundert bis ins digitale Zeitalter, Berlin, Boston 2015. • »Das Eigenleben der Methoden. Eine Wissensgeschichte britischer Konsumentenklassifikationen im 20. Jahrhundert«, in: Geschichte und Gesellschaft 42 (2016), p. 86–112. BURKE, Peter: What is the History of Knowledge, Cambridge, Malden (MA) 2016. COOPER, Andrew: »Sex, Gender and Society«, in: Ivan Reid, Eileen Wormald (eds.), Sex Differences in Britain, London 1982. CROMPTON, Rosemary: Class and Stratification. An Introduction to Current Debates, Cambridge 1993. • — / HARRIS, Fiona: »A Reply to Hakim«, in: The British Journal of Sociology 49 (1998), p. 144–149. CURTIS, Bruce, The Politics of Population. State Formation, Statistics and the Census of Canada 1840–1875, Toronto 2001. DELPHY, Christine: »Women in Stratification Studies«, in: Helen Roberts (ed.), Doing Feminist Research, London 1981, p. 114–128. DEX, Shirley: »Goldthorpe on Class and Gender. The Case Against«, in: Jon Clark, Celia Modgil, Sohan Modgil (eds.), John H. Goldthorpe. Consensus and Controversy, London, New York 1990, p. 135–152. DOWNHAM, John: »How Did the MRS Journal Start?«, in: International Journal of Market Research 50 (2008), p. 7–9. EICHLER, Margrit: The Double Standard. A Feminist Critique of Feminist Social Science, London 1980. ERIKSON, Robert: »Social Class of Men, Women and Families«, in: Sociology 18 (1984), p. 500–514. Experian, Mosaic United Kingdom, The Consumer Classification of the United Kingdom, Nottingham 2009. FOUCAULT, Michel: Power/ Knowledge. Selected Interviews and Other Writings, 1972–1977, Brighton 1980. GEPPERT, Alexander C.T. / JENSEN, Uffa / WEINHOLD, Jörn: »Verräumlichung. Kommunikative Praktiken in historischer Perspektive 1840-1930«, in: Alexander C. T. Geppert, Uffa Jensen, Jörn Weinhold (eds.), Ortsgespräche. Raum und Kommunikation im 19. und 20. Jahrhundert, Bielefeld 2005, p.15–49. GOLDTHORPE, John H.: »Women and Class Analysis. In Defence of the Conventional View«, in: Sociology 17 (1983), p.465-488. • »Women and Class Analysis. A Reply to the Replies«, in: Sociology 18 (1984), p.491-499. GROVES, Robert M.: »Three Eras of Survey Research«, in: Public Opinion Quarterly 75 (2011), p. 861–871. HAKIM, Catherine: »Developing a Sociology for the Twenty-first Century. Preference Theory«, in: The British Journal of Sociology 49 (1998), p. 137–143. HALSEY, Albert H.: A History of Sociology in Britain. Science, Literature, and Society, Oxford 2004. HANNAM, June: »Women’s History, Feminist History«, in: Making History. The Changing Face of the Profession in Britain, http://www.history.ac.uk/makinghistory/resources/articles/womens_history.html (date: 07.06.2016). HARAWAY, Donna: »Situated Knowledges. The Science Question in Feminism and the Privilege of Partial Perspective«, in: Feminist Studies 14 (1988), p. 575–599. HARRIS, Muriel: A Handbook for Interviewers, London 1950. HEATH, Anthony / BRITTEN, Nicky: »Women’s Jobs Do Make a Difference. A Reply to Goldthorpe«, in: Sociology 18 (1984), p. 475–490. HIGGS, Edward: »Women, Occupations and Work in the Nineteenth-century Censuses«, in: History Workshop Journal 23 (1987), p. 59–80. • The Information State in England. The Central Collection of Information on Citizens 1500–2000, Basingstoke 2003. HOINVILLE, Gerald / JOWELL, Roger: Survey Research Practice, London 1982 . IGO, Sarah E.: »Hearing the Masses. The Modern Science of Opinion in the United States«, in: Kerstin Brückweh et al. (eds.), Engineering Society. The Role of the Human and Social Sciences in Modern Societies 1880–1980, Basingstoke 2012, p. 215–233. KNÖBL, Wolfgang: »Das Problem ›Europa‹. Grenzen und Reichweite sozialtheoretischer Deutungsansprüche im 20. Jahrhundert«, in: Lutz Raphael (ed.), Theorien und Experimente der Moderne. Europas Gesellschaften im 20. Jahrhundert, Köln 2012, p. 253–286. OFFICE FOR NATIONAL STATISTICS: Labour Force Survey - User Guide, in: webarchive.nationalarchives.gov.uk/20160105160709/http://www.ons.gov.uk/ons/guide-method/method-quality/specific/labour-market/labour-market-statistics/index.html (date: 06.06.2016). • Labour Force Survey User Guide, vol. 8: Household and Family Data, 2008, in: http://www.ons.gov.uk/ons/guide-method/method-quality/specific/labour-market/labour-marketstatistics/volume-8---2008.pdf (date: 06.06.2016). LANDWEHR, Achim (ed.): Geschichte(n) der Wirklichkeit. Beiträge zur Sozial- und Kulturgeschichte des Wissens, Augsburg 2002. LATOUR, Bruno: Science in Action. How to Follow Scientists and Engineers through Society, Cambridge (MA) 1987. LIDDINGTON, Jill / CRA-WFORD, Elizabeth: »›Women Do Not Count, Neither Shall They be Counted‹. Suffrage, Citizenship and the Battle for the 1911 Census«, in: History Workshop Journal 71 (2011), p. 98–127. MANNHEIM, Karl: Ideology and Utopia. An Introduction to the Sociology of Knowledge, New York, London 1936. MARTIN, Jean / BARTON, Jeremy: »The Effect of Changes in the Definition of the Household Reference Person«, in: Survey Methodology Bulletin 381 (1996), p. 1–8. MCCROSSAN, Liz: A Handbook for Interviewers. A Manual of Social Survey Practice and Procedures on Structured Interviewing, London 1984. • A Handbook for Interviewers. A Manual of Social Survey Practice and Procedures on Structured Interviewing, London 31991 . MEIER, Erhard: »The New NRS Classification Measurements. The Differences Between Chief Income Earner, Head of Household, Housewife and Shopper«, in: Conference Papers, The Market Research Society, 37th Annual Conference, 16–18 March 1994, International Convention Centre, Birmingham 1994, p. 139–145. MILLER, Peter / ROSE, Nikolas: Governing the Present. Administering Economic, Social and Personal Life, Cambridge 2008. MONK, Donald: Social Grading on the National Readership Survey, London 41978; 51985. NIXON, Sean: »Mrs. Housewife and the Ad Men. Advertising, Market Research, and Mass Consumption in Postwar Britain«, in: Hartmut Berghoff, Phillip Scranton, Uwe Spiekermann (eds.), The Rise of Marketing and Market Research, New York 2012, p. 193–213. OAKLEY, Ann: Subject Women, London 1981. PLATT, Jennifer: »Women in the British Sociological Labour Market 1960-1995", in: Sociological Research Online 4 (2000), www.socresonline.org.uk/4/4/platt.html (date: 07.06.2016). • »Biographical Journey in Sociology«, in: The British Sociological Association, britsoc.co.uk/what-is-sociology/biographical-journeys.aspx (date: 07.06.2016). PORTER, Theodore M.: »Statistics and Statistical Methods«, in: Theodore M. Porter, Dorothy Ross (eds.), The Cambridge History of Science, vol. 7: The Modern Social Sciences, Cambridge 2003, p.238–250. PRESSER, Harriet B.: »Decapitating the U.S. Census Bureau’s ›Head of Household‹. Feminist Mobilization in the 1970s«, in: Feminist Economics 4 (1998), p. 145–158. RAPHAEL, Lutz: »Ordnungsmuster und Selbstbeschreibungen europäischer Gesellschaften im 20. Jahrhundert«, in: Lutz Raphael (ed.), Theorien und Experimente der Moderne. Europas Gesellschaften im 20. Jahrhundert, Köln 2012, p. 9–20. REDMAYNE, Paul / WEEK, Hugh: Market Research, London 1931. ROSENEIL, Sasha: »The Coming of Age of Feminist Sociology. Some Issues of Practice and Theory for the Next Twenty Years«, in: The British Journal of Sociology 46 (1995), p. 191–205. SAID, Edward: Orientalism, London 1978. SAMETZ, R. M. A. / PLEIZIER, C.: Household Reference Person in the Census of Canada. Some Alternatives and Their Implications, Ottawa 1980. SAMUELS, John, »Social Class in the Future«, in: Admap Magazine (July 1988). SARASIN, Philipp: »Was ist Wissensgeschichte?«, in: Internationales Archiv für Sozialgeschichte der deutschen Literatur 36 (2011), p. 159–172. SAVAGE, Mike: Identities and Social Change in Britain since the 1940s. The Politics of Method, Oxford 2010. SPEICH CHASSÉ, Daniel / GUGERLI, David: »Wissensgeschichte. Eine Standortbestimmung«, in: Traverse 1 (2012), p. 85–100. STANWORTH, Michelle: »Women and Class Analysis. A Reply to John Goldthorpe«, in: Sociology 18 (1984), p. 159–170. STATISTICS CANADA: Census Technical Reports. Families, Catalogue 92–328E, Ottawa 1994. SZRETER, Simon / SHOLKAMY, Hania / DHARMALINGAM, Arunachalam: »Contextualizing Categories: Towards a Critical Reflexive Demography«, in: Simon Szreter, Hania Sholkamy, Arunachalam Dharmalingam (eds.), Categories and Contexts. Anthropological and Historical Studies in Critical Demography, Oxford 2004, p.3–32. • »The Genesis of the Registrar-General’s Social Classification of Occupations«, in: British Journal of Sociology 35 (1984), p. 522–546; Szreter, Simon, »The Official Representation of Social Classes in Britain, the United States, and France. The Professional model and ‘Les Cadres’«, in: Comparative Studies in Society and History. An International Quarterly 35 (1993), p.285–317. TREDRE, Roger: »Women ›Can Be Head of House‹. Sex Bias Removed From Questionnaires«, in: The Independent (March 30, 1993), http://www.independent.co.uk/news/uk/woman-can-be-head-of-house-sex-bias-removed-from-questionnaires-1500789.html (date: 31.08.2009). UNITED NATIONS, DEPARTMENT OF ECONOMIC AND SOCIAL AFFAIRS, STATISTICAL OFFICE OF THE UNITED NATIONS: Principles and Recommendations for the 1970 Population Censuses, New York 1969. • Sex-Based Stereotypes, Sex Biases and National Data Systems, New York 1980. • Principles and Recommendations for Population and Housing Censuses, New York 2008. U.S. BUREAU OF THE CENSUS: 1970 Census of Population. Subject Reports: Family Composition, Washington DC 1973. • Minutes. Census Advisory Committee of the American Economic Association, December 3, 1976. • 1990 Census of Population and Housing. Public Use Microdata Samples. United States, Technical Information, Washington DC 1993. VOGEL, Jakob: »Von der Wissenschafts-zur Wissensgeschichte. Für eine Historisierung der ›Wissensgesellschaft‹«, in: Geschichte und Gesellschaft 30 (2004), p. 639–660. WATSON, Walter B. / BARTH, Ernest A. T.: »Questionable Assumptions in the Theory of Social Stratification«, in: Pacific Sociological Review 7 (1964), p. 10–16. WHITE, Ian: »No Vote – No Census. An Account of Some of the Events of 1910–1911«, in: Population Trends 142 (2010), p. 1–19. WOLFE, Alan R. (ed.): Standardised Questions. A Review for Market Research Executives. A Report by the Research and Development Committee, London 1973–74; 21984. ZIEMANN, Benjamin: »Die Metaphorik des Sozialen. Soziologische Selbstbeschreibungen westeuropäischer Gesellschaften im 20. Jahrhundert«, in: Lutz Raphael (ed.), Theorien und Experimente der Moderne. Europas Gesellschaften im 20. Jahrhundert, Köln 2012, p. 193–227.