Constructing Survey Data

An Interactional Approach

Giampietro Gobo, Sergio Mauceri, Author

• Publisher: SAGE Publications Ltd
• Year: 2013
• Print ISBN: 9781849201773, 9781849201766
• eBook ISBN: 9781473904835
• Edition: 1
• Topic: Social Sciences
• Subtopic: Social Science Research & Methodology
• Index: Social Sciences

Engaging and informative, this book provides students and researchers with a pragmatic, new perspective on the process of collecting survey data. By proposing a post-positivist, interviewee-centred approach, it improves the quality and impact of survey data by emphasising the interaction between interviewer and interviewee. Extending the conventional methodology with contributions from linguistics, anthropology, cognitive studies and ethnomethodology, Gobo and Mauceri analyse the answering process in structured interviews built around questionnaires.  

The following key areas are explored in detail:

• An historical overview of survey research
• The process of preparing the survey and designing data collection
• The methods of detecting bias and improving data quality
• The strategies for combining quantitative and qualitative approaches
• The survey within global and local contexts

Incorporating the work of experts in interpersonal and intercultural relations, this book offers readers an intriguing critical perspective on survey research.

Giampietro Gobo, Ph.D., is Professor of Methodology of Social Research and Evaluation Methods at the Department of Social and Political Studies - University of Milan. He has published over fifty articles in the areas of qualitative and quantitative methods. His books include Doing Ethnography (Sage 2008) and Qualitative Research Practice (Sage 2004, co-edited with C. Seale, J.F. Gubrium and D. Silverman). He is currently engaged in projects in the area of workplace studies.

Sergio Mauceri, Ph.D., is Lecturer in Methodology of Social Sciences and teaches Quantitative and Qualitative Strategies of Social Research at the Department of Communication and Social Research - University of Rome 'La Sapienza'. He has published several books and articles on data quality in survey research, mixed strategies, ethnic prejudice, multicultural cohabitation, delay in the transition to adulthood, worker well-being in call centres and homophobia.

I

The context

ONE

Surveying the survey: back to the past

A look, however brief, at the history of the survey can lead to a better understanding of this methodology and of the technical aspects of the standardized interview. History need not be considered a sort of musty portrait gallery or a boring discipline smelling of mothballs. History helps us deconstruct our mental categories, usually seen and presented to us as natural, normal and obvious. For example the term ‘statistics’ (Staatskunde) was coined in Germany by Ermanno Conring (1606–1681), professor of Public Law, and later by Gottfried Achenwall (1719–1772), jurist and philosopher. It meant ‘science of the state’, or the systematic description of the relevant aspects of a state. Hence university statistics was configured as a description, mostly qualitative, of states. Counter-intuitively, history reveals that statistics was coined in the field of qualitative and descriptive research, without the need of numerical descriptions, which were even considered by the fellows of German university Statistics, to be ‘vulgar’! Only after the merge with political arithmetic does statistics take on the quantitative aspects that characterizes it today.

In addition, an historical viewpoint immunizes us against the ingenuousness (increasingly commonplace among contemporary social scientists) of presenting as novel theories and methods which were proposed seventy or eighty years ago. Knowledge of history saves us from having constantly to reinvent the wheel, to use the well-known expression. For this reason, throughout the book we will always quote the authors who (as far as we know) first introduced an experiment, a hypothesis, a theory. To those who will see many of our sources as outdated we can only reply that we prefer the original to the copy.

The following historical review is meant to open the way to an understanding of the economic, political, social and cultural climate in which the survey developed and spread, the aim being to bring it up to date and free it from its ideological baggage.

1.1

The making of the ‘survey society’: the nineteenth century

Gubrium and Holstein have re-examined the debate conducted in the 1950s by David Riesman, Mark Benney and Everett C. Hughes on the origin of the interview. This reflexive move has had the great merit of reminding us that ‘no longer should we regard the interview as simply an instrument of data gathering technology; it is also an integral part of society – now more decidedly than ever’ (2001: xii).

Although the ancient historian Thucydides was using interviews to gather data on the Peloponnesian Wars as early as the fourth century BC, Gubrium and Holstein maintain that the sociological interview was born at a particular moment in the history of society (the nineteenth century) and embodied certain of its cultural features. The first journalistic interview appeared in a newspaper in 1859, and in 1863 its use became common practice (see Davies 2009).

The interview was a product of modernity and it reflected ‘the modern temper’ (Riesman and Benney 1956). It consequently reflected a change in social relations whereby it gradually became normal to converse with strangers, to express one’s opinions in public (with the consequent shift in the boundary between public and private opinion) and, more recently, to display one’s feelings and emotions overtly. In the words of French philosopher Michel Foucault, the interview is one of the ‘technologies of the self’, one of the devices for the modern invention of subjectivity, an institutional practice intended to create the idea of the ‘individual’ and to emphasize the personal point of view.

Atkinson and Silverman (1997) have perceptively noted that we live in an ‘interview society’, a society in which interviewing has become a fundamental activity, and interviews seem to have become a crucial means by which people make sense of their lives. On one hand, the interview (in the broad sense) today is one of the most widespread social forms of information gathering: policemen, doctors, shop assistants, judges, social workers, managers, teachers, psychotherapists, priests, journalists, personnel recruiters and call-centre operators all engage in some form of interviewing in their routine work. On the other hand, radio and television constantly and invasively transmit the expression (sometimes the flaunting) of the self into our homes: talk shows, variety shows, documentaries, the sound bites elicited at the end of sports events, the emotions stolen from victims at scenes of tragedies or disasters. One may be tempted to ask whether the pervasiveness of this form of social relation (the interview) is linked to the fact that the survey and in-depth interview, having originated together, predominate in social research as well. According to Silverman (1997: 248), society and the interview are mutually constitutive: on the one hand, to come into being and develop, in-depth and survey interviewing required a particular type of society; on the other, these research methods strengthen the society which has produced them. In accordance with this position, Gubrium and Holstein (2001: xii) argue that the interview is not a simple technique, a neutral instrument of information gathering, but has become an integral part of contemporary society, which has in turn created the social and cultural conditions for its emergence.

1.2

The common roots of the survey and in-depth interview

Some will be surprised at the above juxtaposition of the survey and the in-depth interview. All the textbooks in fact consider them as two distinct methodologies, if not actually in direct contrast to one another. Be that as it may, however, their differentiation is a relatively recent phenomenon, dating to the beginning of the 1940s. Before that time, the survey consisted of a conversation during which both open-ended and fixed-choice questions were put to an interviewee. Thus, there were no clear distinctions between question types. In fact, within both of the two culturally and socially independent traditions (European and American) that gave birth to the survey, information was gathered by much more flexible means than those in use today.

1.2.1

The American way to the survey: straw polls (1824–1900)

It is not easy to date the origin of the survey with certainty. A plausible hypothesis points to the first straw polls carried out in 1824 on the Atlantic coast of the US during that year’s presidential election campaign (Smith 1990). Straw polls (from the metaphor of throwing a piece of straw into the wind to see which way it is blowing) were devised to discover the people’s mood, reveal their intentions and predict the result of the election. The method was very simple and consisted of counting raised hands expressing candidate preferences during public assemblies or holiday celebrations (such as the fourth of July). It was thus a collective rather than individual inquiry – that is, carried out in the presence of a great number of people. This practice was then gradually extended to include much smaller groups such as jurors, crew members and passengers on ships and trains, people waiting to pay their taxes or attending military parades, where the preferences of the soldiers themselves could be gathered as well. Another device was the use of poll books, which could be left for days in public places. The results were then published in the newspapers. Straw polls soon became an important source of information for journalists.

The first straw polls were promoted by a candidate’s campaign committee or by a party faction in revolt against the power of the caucuses that selected, top-down, the names of the candidates. The congressional caucus was a gathering of congressmen; apparently the term comes from the Algonquin language and means ‘counsel’, cau´-cau-as´u, and probably came into American political jargon thanks to the Tammany Society, a charitable organization linked to the Democratic Party in New York that liked to use Native American expressions. The caucus chose the candidate for a presidential election in a non-public meeting (the smoke-filled room) run by a small circle of Washington politicians. For this reason polls were an indication of fresh democratic ferment, a bottom-up desire to participate. And, in fact, the straw polls of 1824 marked the disappearance of congressional caucuses and represented a new way of choosing the candidates. They were a novel instrument for a new form of political participation.

It was only later that they became a form of electoral propaganda, used by party publications or politically aligned newspapers to influence the electorate.

Methodologically…

From a methodological point of view the straw poll left much to be desired – it was anything but representative, blatantly biased or partisan and beset by outright fraud in vote counting. Towards the end of the nineteenth century there was an increase in the territorial coverage of polling, and the gathering of voter intentions was more systematic. In the early 1900s, some 75 years after they first appeared, straw polls began to enjoy a new era of prominence. No longer were they organized and promoted by militants, but rather by publications such as the New York Herald, Literary Digest, Farm Journal, etc., which put a variety of methods into practice, including face-to-face interviews, telephone interviews, postcards and newspaper coupons (Robinson 1932). In the 1920s (when the Chicago School was culturally dominant in sociology), dozens of pollsters criss-crossed the city with portable ballots and boxes to collect votes in offices, factories, theatres, clubs, hotels, residences, and on trolley cars and street corners.

Sometimes the newspapers would organize small fleets of trucks that toured the states, stationing pollsters in the main streets of cities and in front of factories at closing time (Robinson 1932).

The success of these efforts was extraordinary, especially in the period between the world wars. What is particularly amazing is that these crude tools, full of methodological omissions and beset by systematic errors, managed for decades to predict the winning candidate. As a result, the polls became so famous that they were able to stake out a position as marketing tools for the newspapers. It was by means of marketing polls that the papers sought to increase their influence and circulation and become more credible and authoritative. People who were interviewed, or sent the coupons or postcards, became potential subscribers.

Of all the newspapers involved, the Columbus Dispatch (Ohio) apparently holds the distinction of being the first, beginning at the end of the nineteenth century, to organize a local poll with properly trained interviewers, careful controls and data collection criteria based on socio-demographic variables similar to those used today in quota samples, criteria that would be codified in survey methodology and statistical sampling theory only many years later (Teer and Spence 1973).

1.2.2

The European origins of the survey: mixed methods

While it was from straw polls that the survey arose in the US, it was developed in Europe (at about the same time) on other grounds – as a mixed method.

The first actual interviews happened in the 1820s–30s: in England, conducted by factory inspectors; in France, first by medical hygienists (among whom the most important was Louis-René Villermé) and then by Flora Tristan (French socialist writer and feminist, and grandmother of the painter Gauguin) and especially Antoine-Eugène Buret, a political economist and utopian socialist (a pupil of Sismondi) whose work was an important source for the young Marx (see Fabiano 1979).

The mining engineer and (later) sociologist Frédéric Le Play (1806–1882) probably invented (in the late 1840s) what can be considered the first prototype of the questionnaire. This was used to collect information about family budgets in a diary of earnings and expenditures which expressed the family’s life in figures (Seger 1970: 181). Moreover, Le Play was probably the first to use the ‘monographic’ method (or case study), the detailed investigation of ‘typical’ cases consisting of groups of working-class families representing a country, a region or an industrial sector. He was also the first to use the methodology of participant observation whereby the researcher lived with the family under study for the time required to collect the necessary documentation (Seger 1970: 184). Le Play also pioneered the idea that researchers should collect original information – that is, information obtained directly from the subjects studied. Although this practice was certainly innovative, it was not subsequently taken up by researchers for several decades. Indeed, even at the end of the 1800s the industrialist and philanthropist Charles Booth (1840–1916) conducted his survey on poverty in London by consulting 2500 school inspectors rather than directly interviewing the subjects of his research – probably because of the widespread prejudice that the poor could not be trusted to give reliable information (Thompson 1973: 43; Marsh 1982: 18). This prejudice was also exhibited by Marx and Engels (1848) in their opinion of the lumpenproletariat: thieves, rogues, beggars, the unemployed, the hungry, the miserable and delinquent, who were by...

Table of contents

• Cover Page
• Half Title
• Title
• Copyright
• Dedication
• Brief Contents
• Contents
• About The Authors
• Acknowledgements
• Introduction: Rescuing the survey from the surveyists
• Part I The context
• Part II From questions to answers
• Part III Constructing answer comparability
• Part IV Designing data quality through mixed strategies
• Part V Envisioning the future
• References
• Index