Quantity and Quality in Social Research

Alan Bryman

• Publisher: Routledge
• Year: 2003
• eBook ISBN: 9781134897360
• Edition: 1
• Topic: Social Sciences
• Subtopic: Sociology
• Index: Social Sciences

This book focuses upon the debate about quantitative and qualitative research which took root in the 1960s, although many of the central themes go back centuries. The basic terms of the debate have been felt in many of the disciplines which make up the social sciences, especially sociology, social psychology, education research, organization studies, and evaluation research.

1

Introduction

The chief focus of this book is the debate about quantitative and qualitative research in the social sciences, in particular the relative merits and disadvantages of these two styles of inquiry. This is a controversy in which philosophical issues tend to be interwoven with discussions about the nature and capacities of different methods of research. Quantitative research is typically taken to be exemplified by the social survey and by experimental investigations. Qualitative research tends to be associated with participant observation and unstructured, in-depth interviewing. On the face of it, questions relating to the advantages and capacities of these two approaches and their associated techniques would seem to be technical ones, pertaining to their respective strengths and weaknesses in relation to particular research topics. In fact, philosophical issues figure very strongly and have much to do with a growing interest in the methods associated with a qualitative style of inquiry.

This state of affairs—the entanglement with philosophy and the interest in qualitative research—is of relatively recent origin. There has, of course, always been an awareness of the differences between the nature of, for example, the social survey and participant observation. However, the focus tended to be on the capacity of surveys to provide a framework in which the procedures associated with the scientific method could be followed, and the poor showing of participant observation in this regard. Consequently, qualitative methods such as participant observation tended to be regarded as relatively marginal in the context of the social scientist’s armoury of data collection techniques.

The standard format of social research methods textbooks (particularly those published prior to the mid-1970s) exemplifies these tendencies. The typical methods text began with a number of chapters on the scientific method. This discussion formed a bedrock for examining the procedures associated with the survey and often with experiments. Because of its inability to conform to the canons of scientific method, a technique like participant observation was accorded scant attention. In Goode and Hatt’s (1952) influential text, participant observation is consigned to a fairly cursory treatment in a chapter on ‘Some problems in qualitative and case analysis’ and another on ‘Observation’; this contrasts sharply with the eight chapters dealing with aspects of survey procedures, three on scaling methods alone. The growing interest in the analysis of survey data, following the influence of writers like Lazarsfeld and Blalock, resulted very often in much greater attention being devoted to its associated procedures: Phillips (1966) and Nachmias and Nachmias (1976), for example, both wrote a whole section on data analysis, representing some one hundred pages in each; by contrast, participant observation can be found in a relatively short chapter on observational methods in which it is sandwiched between explications of structured observation (which is in fact typically deployed as a means of generating quantitative data). Even though writers often recognized the potential strengths of participant observation, the tendency was to view it somewhat deprecatingly as simply a procedure for developing hunches and hypotheses to be subsequently corroborated by the more rigorous survey, experiment or whatever.

In general, techniques of participant observation are extremely useful in providing initial insights and hunches that can lead to more careful formulations of the problem and explicit hypotheses. But they are open to the charge that findings may be idiosyncratic and difficult to replicate. Therefore many social scientists prefer to think of participant observation as being useful at a certain stage in the research process rather than being an approach that yields a finished piece of research. (Blalock, 1970, pp. 45–6)

Such a statement carries a clear implication that the role of a qualitative technique such as participant observation is a very restricted one and that it does not possess the solidity of research designed within a framework more obviously redolent of the scientific method.

In the earlier generations of textbook there was a clear awareness of a difference between quantitative and qualitative research. This awareness can also be discerned in occasional attempts by researchers to compare and contrast the virtues and vices of participant observation and survey methods (e.g. Vidich and Shapiro, 1955; Becker and Geer, 1957). However, all of these discussions operated almost exclusively at the level of the technical adequacy of the techniques as such. What distinguishes the debate that gained ground in the 1970s was the systematic and self-conscious intrusion of broader philosophical issues into discussions about methods of research. The pivotal point for much of the controversy was the appropriateness of a natural science model to the social sciences. Whereas the writers of the earlier methods textbooks almost took for granted the desirability of following natural science procedures, the proponents of qualitative research argued that this was an inappropriate model for studying people. Much of the argument levelled against the orthodoxy of quantitative research derived from the growing awareness and influence of phenomenological ideas which gained a considerable following in the 1960s. It was argued that the application of a ‘scientific’ approach—in the form of surveys and experiments—fails to take into account the differences between people and the objects of the natural sciences. Research methods were required which reflected and capitalized upon the special character of people as objects of inquiry. A qualitative research strategy, in which participant observation and unstructured interviewing were seen as the central data gathering planks, was proposed since its practitioners would be able to get closer to the people they were investigating and be less inclined to impose inappropriate conceptual frameworks on them.

In other words, philosophical ideas gained prominence because a key ingredient is the question of the appropriateness of the canons of scientific method to the study of people. As indicated above, the growing interest in qualitative research (and the formation of a philosophical rationale for it) was a major impetus to this development. Increasingly, the terms ‘quantitative research’ and ‘qualitative research’ came to signify much more than ways of gathering data; they came to denote divergent assumptions about the nature and purposes of research in the social sciences. The fact that the terminology seems to imply that ‘quantification’ or its absence is the central issue is highly unfortunate, since the issues span much more widely than this implies. Indeed, a number of writers have proposed alternative terms. For example, Cuba and Lincoln (1982) propose a contrast between rationalistic (i.e. quantitative) and naturalistic (i.e. qualitative) paradigms, while Evered and Louis (1981) use a contrast between ‘inquiry from the outside’ and ‘inquiry from the inside’. Magoon (1977) and J.K.Smith (1983) refer to ‘constructivist’ and ‘interpretive’ approaches respectively in place of ‘qualitative’. However, such alternative terms have not achieved a wide currency and the quantitative/qualitative divide tends to be the main focus.

It is difficult to say precisely at what point the debate became prominent. In the 1960s the discussion of the nature of participant observation by Bruyn (1966) and the delineation of an atttractive logic to the connection between theory and qualitative data by Glaser and Strauss (1967) did much to lay the groundwork. A volume edited by Filstead (1970) figured early in the foray. Although it contained previously published papers, the fact that it was exclusively concerned with ‘qualitative methodology’ and that its introduction appeared to be pointing to a different approach to studying social life from that advocated by the practitioners of a scientific approach, heightened its impact. Hot on its heels came Lofland’s (1971) little textbook on qualitative methods. By the end of the decade Schatzman and Strauss (1973), Fletcher (1974), Bogdan and Taylor (1975), Douglas (1976), Schwartz and Jacobs (1979) and others had written textbooks in which qualitative research figured strongly or exclusively. Journals devoted to publishing articles based on qualitative research began to appear and in 1979 the Administrative Science Quarterly—a bastion of quantitative research—published a special number devoted to qualitative research. Further, the debate has made incursions into a variety of territories: evaluation research, educational studies, organizational studies, social psychology, and other fields.

The interest in the debate may in part be attributed to the growing interest in T.S.Kuhn’s (1970) work on the history of science. One aspect of this influential book is particularly pertinent, namely the idea of a ‘paradigm’—a cluster of beliefs and dictates which for scientists in a particular discipline influence what should be studied, how research should be done, how results should be interpreted, and so on. The various social science disciplines were deemed to be ‘preparadigmatic’, meaning that there was no one overarching paradigm pertaining to each discipline; instead, it was suggested, there are a number of ‘pre-paradigms’ which compete for paradigmatic status. These ideas seemed to contribute to a greater sensitivity to the assumptions and methods associated with competing approaches to the social sciences. Further, many writers on the debate about quantitative and qualitative research refer to the two approaches as paradigms, while one can often detect among other writers a tendency to think in terms of the idea of paradigm even though the term itself is not used. However, the introduction of such issues into the language of the social sciences carries with it certain dangers, for, as one notable philosopher of science has remarked,

Never before has the literature on the philosophy of science been invaded by so many creeps and incompetents. Kuhn encourages people who have no idea why a stone falls to the ground to talk with assurance about scientific method. (Feyerabend, 1975, p. 6)

On the face of it, the incursion of broader philosophical issues into the study of methods was a breath of fresh air. It implied that methodology is not an arid discipline replete solely with technical issues such as when to use a postal questionnaire, the structure of a Solomon Four-Group Design, or recognizing the dire effects of failure to control for suppressor variables. It could entertain the consideration of bigger issues.

What are ‘quantitative research’ and ‘qualitative research’? In some treatments they are viewed as competing views about the ways in which social reality ought to be studied, and as such they are essentially divergent clusters of epistemological assumptions, that is, of what should pass as warrantable knowledge about the social world. For other writers, quantitative and qualitative research are simply denotations of different ways of conducting social investigations and which may be conceived of as being appropriate to different kinds of research question and even as capable of being integrated. When this second view is taken, they are more or less simply different approaches to data collection, so that preferences for one or the other or some hybrid approach are based on technical issues. In this view, the prime consideration is that of dovetailing the appropriate technique to a particular research question. Many writers, as the later chapters will reveal, vacillate between these two levels of analysis. To a very large extent, these two research traditions (be they indicative of epistemological or technical positions) can be thought of as divergent genres, especially in regard to their modes of presenting research findings and programmatic statements. Of course, they are more than merely literary devices; but it is difficult not to be struck by the different styles of exposition that practitioners of the two traditions espouse. The employment of a scientistic rhetoric—experiment, variables, control, etc.—in quantitative research imposes expectations on the reader about the sort of framework that is about to be encountered, what sorts of criteria of valid knowledge the author endorses, and so on. In short, such linguistic devices act as signals which forewarn the reader about the material to come. By contrast, the self-conscious endorsement by many qualitative researchers of styles of presentation and literary devices which entail a rejection of a scientific rhetoric can be seen as a countervailing genre. Through their rejection of a scientific idiom and their recourse to the style of qualitative research they signal their adoption of a different framework and expect their work to be read and judged within the confines of that framework.

A Comparison of Two Studies

Many of the points adumbrated thus far can be usefully illustrated by reference to two studies which exemplify the contrasting orientations which lie behind the quantitative and qualitative traditions in social research. Of course, the choice of studies is bound to be arbitrary, in that many other examples of reported pieces of research could have been selected as alternatives. The chosen studies are Hirschi’s (1969) investigation of delinquency and Adler’s (1985) research on drug dealers; these monographs may be taken as reasonably representative of the quantitative and qualitative traditions of social research respectively.

Hirschi’s Study of the Causes of Delinquency

Hirschi’s examination of delinquency fits squarely with what is usually taken to be a natural science approach to the study of social reality. This predilection is evident in an earlier work, in which he expressed his preference for quantitative research: ‘because quantitative data can be analyzed statistically, it is possible to examine complicated theoretical problems, such as the relative importance of many causes of delinquency, far more powerfully than with the verbal analysis of qualitative data’ (Hirschi and Selvin, [1967] 1973, p. xii). In Causes of Delinquency, Hirschi (1969) was concerned to test the relative validity of three contrasting theories of the etiology of delinquency; he was particularly interested in how well his own ‘social control’ theory—which posits that delinquent acts occur when ‘an individual s bond to society is weak or broken’ (p. 16)—held up to empirical testing. He used a social survey in order to achieve his aims. As ‘subjects’ he randomly selected a sample of 5,545 school children in an area of California near San Francisco. Great care was taken in the selection of the sample to ensure that it adequately represented the range of schools in the area, as well as the gender and race distribution of the children in the population. The bulk of the data was collected by a self-administered questionnaire which was completed by the students. In addition to questions relating to social background, the questionnaire comprised a great many questions designed to tap the extent to which children were committed or attached to the school, to the family, and to conventional lines of action, in order to test the social control theory which had been formulated. The questionnaire also contained questions designed to gauge the extent of each child’s involvement in delinquent activities. Further data were gleaned from other sources, such as information on each child’s performance in connection with academic achievement tests from school records.

Hirschi’s basic orientation to the research process is clear: one needs to formulate some explicit propositions about the topic to be investigated and design the research in advance specifically in order to answer the research problem. There is a clear concern to be able to demonstrate that the sample is representative of a wider population of school children, though the question of the representativeness of the region in which the research is located is given scant attention. The questionnaire is taken to comprise a battery of questions which ‘measure’ the main concepts involved (e.g. attachment to society); each question (either on its own or in conjunction with other questions to form an index) is treated as a variable which can be related to other questions/variables in order to estimate relationships among the variables which are relevant to the theories being tested. For example, Hirschi presents a contingency table which shows a clear inverse relationship between an index of ‘intimacy of communication between parent and child’ (derived from answers to two questions) and the number of self-reported acts of delinquency (p. 91). But Hirschi is rarely content to leave his data analysis simply at the level of estimates of co-variation or correlation among the variables concerned. Much of the time he is concerned to extricate the causal relationships among his variables. Thus, at the end of a chapter on attachments to school, he writes: ‘The causal chain runs from academic incompetence to poor school performance to disliking of school to rejection of the school’s authority to the commission of delinquent acts’ (p. 132). These causal paths are winkled out by multivariate analysis which allows the analyst to sort out the direct and indirect effects by controlling for intervening variables and the like.

In the end, Hirschi finds that none of the three theories of delinquency emerges totally unscathed from the empirical interrogation to which they were submitted. For example, the control theory seems to neglect the role of delinquent friends which his data suggest has considerable importance. Other writers have attempted to replicate aspects of Hirschi’s research (e.g. Hindelgang, 1973).

Adler’s Study of Upper-Level Drug Dealers

Adler (1985) and her husband took up residence in California in order to attend graduate school in sociology. They soon made friends with a close neighbour (Dave, a pseudonym), who, it transpired, was a drug dealer. He was not a small ‘pusher’ of drugs who was trying to provide funds for his own habit, but someone who dealt in vast quantities and who received huge sums of money in exchange, that is an ‘upper-level’ drug dealer. They were encouraged by their supervisor, Jack Douglas, a prominent contributor to qualitative research on deviance (Douglas, 1972; 1976), to infiltrate Dave’s group of associates in order to carry out a study of such dealers, who are normally highly inaccessible. The nature of Adler’s approach to data collection can be gleaned from the following passage:

With my husband as a research partner, I spent six years in the field (from 1974 to 1980) engaged in daily participant observation with members of this dealing and smuggling community. Although I did not deal, myself, I participated in many of their activities, partying with them, attending social gatherings, traveling with them, and watching them plan and execute their business activities… In addition to observing and conversing casually with these dealers and smugglers, I conducted in-depth, taped interviews, and cross-checked my observations and their accounts against further sources of data whenever possible. After leaving the field, I continued to conduct follow-up interviews during periodic visits to the community until 1983. (Adler, 1985, pp. 1–2)

Adler’s broad orientation is to focus on the ‘subjective understanding of how people live, feel, think, and act’ (p. 2) and so ‘to understand the world from their perspectives’ (p. 11). She sees her work as ‘an ethnographic description and analysis of a deviant social scene’ (p. 2).

Adler’s adoption of a perspective which emphasizes the way in which the people being studied understand and interpret their social reality is one of the most central motifs of the qualitative approach. Through this perspective Adler shows that the views of drug dealing that are often presented in the literature do not fully correspond to the dealers’ own perceptions. For exampl...

Table of contents

• Front Cover
• Half Title
• Social Research Today edited by Martin Bulmer
• Title Page
• Copyright
• Contents
• Dedication
• Preface
• 1 Introduction
• 2 The Nature of Quantitative Research
• 3 The Nature of Qualitative Research
• 4 Problems in Qualitative Research
• 5 The Debate about Quantitative and Qualitative Research
• 6 Combining Quantitative and Qualitative Research
• 7 Comparing Quantitative and Qualitative Research
• 8 Conclusion
• Bibliography and Author Index
• Subject Index