66    QUANTITY AND QUALITY IN SOCIAL RESEARCH
ccrned with Ihc process of implementation rather tliiin solely with its outputs (Finch, 1986). Thus ;i qualitative research approach would emphasize the various responses of both those who implement and those who are affected, the interpretations they invoke of the policy initiative, how they respond to each other's views, how perspective! change, and so on.
Similarly, in his participant observation study of n comprehensive school, Ball (1981) was interested in the way in which a major innovation the introduction of mixed-ability groupings - was implemented. Hall documents not only the sequence of events that this innovation comprised, but also the variety of responses of the teachers. He found that
In the absence of an agreed or imposed 'mandate* for change, the teachers at Beachside were free to attribute their own categories of meaning to the innovation . . . Furthermore, the absence of a mandate for change also meant that the teachers were not obliged to change their teaching methods. Indeed, in some respects constraints inherent in the culture and ethos of the school militated against drastic changes in the organization of learning in the classroom. (Mall, 1981, p. 237)
Thus Ball concludes from his investigation of the meanings attributed to the change to mixed ability groupings that there was substantial variation in teachers' perspectives on it and that the extent of the impact of the change on their leaching practices may have been less pronounced than might otherwise have been anticipated. From the teacher's point of view. Ball shows, change is not a radical departure from a pre-existing state, but a gradual drifting away, and it is from this latter stance that he or she sees change as having occurred. Ball's analysis of the innovation entails treating it as a process whereby the change is introduced, then interpreted by teachers, and the implications of these interpretations for teaching practices are then examined. The final sentence in the above quotation also serves as a reminder of the qualitative researcher's inclination towards a contextual understanding, whereby the teachers' Interpretations of the change are grounded in the context of the school's ethos.
Flexibility and Lack of Strw'turv
Qualitative researchers' adherence lo viewing social phenomena through Ihc eyes of their subjects has led to a wariness regarding the imposition of prior and possibly inappropriate frames of reference on the people they study. Consequently, they tend to favour a research strategy which is relatively open and unstruc-
tured, rather than one which has decided in advance precisely what ought to be investigated and how it should be done. It is also often argued that an open research strategy enhances the opportunity of coming across entirely unexpected issues which may Ik* of interest to the ethnographer.  Participant  observation  particularly lends
itself to this orientation because ihc researcher is i.....ieisc<l in a
social context and can defer analysis until fully acquainted with it.
The foregoing discussion would seem to imply that qualitative researchers do not even have a •problem' that they seek to investigate at the outset of their investigations. In fact, they vary quite considerably in this respect. Some researchers seem to have very loose notions of what they arc intending to investigate once they have negotiated access lo a research site. For example, writing about his ethnographic study of 'fiddling' in a bakery, Ditton (1977, p. 11) affirms that his research 'was not set up to answer any empirical questions'. His decision to concentrate on fiddling was not made until a considerable proportion of the research had already been conducted. Adler (1985) conceived of an ethnographic study of drug dealers only after she and her husband had moved to California to attend a graduate school course in sociology and had come into contact with dealers through a neighbour, My con-irast, some ethnographers have somewhat more precise notions of I heir focus of study at the outset. Mloor (1978) conduclcd an observational study (including data from conversations) of FNT (car, nose and throat) clinics, because he 'was concerned to establish whether or not geographical differences in Ihc incidence of adeno-tonscillcctomy among children could be attributed to . . . differences between FNT specialists in different geographical areas in their routine assessments' (Bloor, 1978, p. 545). While not quite as specific as Bloor in this last Statement, Ball (1981) describes his study of comprehensive schooling as aiming to examine 'the dynamics of selection, socialization and change ... as well as the playing out of social structural and cultural forces in the school' (p. xv).
Irrespective ol whether the research problem is closely defined, qualitative researchers tend to the view that the predominantly open approach which they adopt in the examination of social phenomena allows them access to unexpectedly importmi! topics which may not have been visible m them had they foreclosed the domain of study by a structured, and hence potentially rigid, strategy. It is even possible for the researcher to discover lhat a particular focus is irrelevant. In his study of Whalsay, Cohen (1978) had intended to look at the ways in which changes in ihc technological and economic infrastructure of fishing may have affected skippers' authority. He was also concerned with the ways skippers
0
68    QUANTITY AND QUALITY IN SOCIAL RESEARCH
had responded to changes in the bases of their authority. However, after the start of the field-work Cohen found that the problem he had formulated was 'empirically Irrelevant* and considers that the factors which conspired to render his research problem unproblcm-atic were of greater importance. In particular, he argues that when he viewed the problem within the contexl of the community's pattern of social relations and culture the issue of skippers' authority could not be sensibly extracted from its wider milieu. What is of particular interest here is the qualitative researcher's ability lo recognize the irrelevance of his research question from within the framework of the community (the contextualist emphasis) and the ability lo change direction in the formulation of his problem.
theory and Concepts'
In line with their preference for a research strategy which docs not impose a potentially alien framework on their subjects, qualitative researchers frequently reject the formulation of theories and concepts in advance of beginning their field-work. In particular, they view the imposition of a pre-ordained theoretical framework as deleterious because it may excessively constrain the researcher and also may exhibit a poor fit with participants* perspectives. By and large, qualitative researchers favour an approach in which the formulation and testing of theories and concepts proceeds in tandem with data collection. This issue will be the focus of a more detailed discussion in Chapter 4. In the meantime, a fairly general treatment of the connection between concepts and research is supplied.
Blumer's (1954) writing on concepts is widely cited and accepted in broad terms by many qualitative researchers. He argued strenuously against treating concepts in terms of fixed empirical referents which are then applied to the real world. This is the basic procedure of much quantitative research, which sets up precise operational definitions against which reality may be gauged. Against such an approach he proposed treating social scientific concepts as sensitizing concepts which provide 'a general sense of reference and guidance in approaching empirical instances' (p. 7). This approach to the connection of concepts and data means that a concept provides a set of general signposts for the researcher in his or her contact with a field of study. While the concept may become increasingly refined, it does not become reified such that it loses contact with the real world. One concomitant of this approach is that the qualitative researcher is attuned to the variety of forms that the concept may subsume. As such, a sensitizing concept retains close contact with the complexity of social reality, rather than trying to bolt it on to fixed, prcformulated images. In a study
NA TURE OF QUALITA TIVE RFSFA RCII    69
of power in a medical school in the USA, Bücher (1970, p. 26) preferred 'lo postpone sharp definitions of my terms and concentrate on the empirical situation in the expectation that definitions appropriate to my setting will emerge from analysis of the data'. This procedure allowed the author to draw out the different forms and bases of power within the school and to look at its operation in both formally designated offices and beyond. The general understanding of concepts, then, seems to imply that they are both inputs and outputs in relation to the research enterprise; that is, they provide a general frame of reference at the outset and are also refined by the researcher during the field-work period.
The general approach of qualitative researchers to concepts and theories is to be mistrustful of their specification prior to the start of the research enterprise. This is not lo say (hat a method like participant observation is incapable of testing theories and allowing concepts to be operationally defined at the outset of a qualitative study. Becker (1958) proposed an approach to participant observation in which the testing of theories was a prominent ingredient, in order to infuse it with (as he put it) a more 'scientific' flavour. Thus participant observation may not be entirely incompatible with the kind of approach detailed in Figure 2.1 in the context of quantitative research, but qualitative researchers have tended to perceive it as ideal for the extraction of actors' rather than social scientists' prior conceptual schemes. Further, the disillusionment with the underlying principles of quantitative research, which has already been alluded to, almost certainly has militated against the enlistment of participant observation for such purposes by researchers working within the qualitative tradition.
Conclusion
Whereas quantitative research was described in the previous chapter as drawing the bulk of its intellectual inspiration from a natural science approach, and from certain tenets of positivism in particular, qualitative research derives from, and has been stimulated by, traditions which are distinctively different from such an orientation. The connection between the intellectual traditions delineated above and the chief characteristics of qualitative research is most evident in relation to the commitment to seeing through the eyes of the people being studied. All five of the intellectual currents discussed reveal this general concern. Indeed, some writers (e.g. Deutscher, 1973) see all perspectives which take (he actor's point of view as the empirical point of departure as 'phcnomcnologicaľ. The emphasis on description is in large part a product of the tendency in these perspectives to adopt a naturalistic
70    QUANTITY AND ('í   11 11 Y IN SOCIAL RESEARCH
approach which retains fidelity to the real world. The stress on eofi textual and holistic undcrsinudiug can be attributed to ;i preference for grounding accounts of social reality in subjects' perceptions of their environment: the symbolic intcractionist emphasis on the definition of the situation, lot example, illustrates this tendency. Symbolic imeraclionism is also responsible for the accent on pro CCSS m that its exponent* tend to view social life as a scries Ol Interlinked events, mediated by people's interpretive devices. The tendency for conceptual and theoretical reasoning to be seen as something which either occurs en passant or towards the end of the research enterprise can be attributed to the qualitative researcher's distrust of stances which may fail to do justice to the subject's orientation to the social world.
Interest in qualitative research has gained increasing momentum since the late 1960s. Yet qualitative research as such predates this period. While the exlcnt to which the work of the Chicugo School ol sociology in the I92Ü1 and 1930s Can be interpreted us qualitative research (in the sense in which the term is currently used) has been questioned, it has none the less been viewed as an early example of such research (Plait, 1983). Since then, the work of writers like Whyte and Gans has provided notable examples of qualitative research before the burgeoning ol" interest in the 1960s. Further, the notion of verstehen and the perspective of symbolic interactionism have long been familiar to social scientists. Why then does one find a fairly sudden increase in interest in qualitative research if some of the intellectual traditions on which it is supposed to rest and the methods with which it is associated predate the 1960s? 1 wo \.v lois are particularly noteworthy, ľirsí, ii is possible to deteel con siderable disillusionment with the fruits of quantitative ivscuich among the early writers on qualitative research. Th«M wrlteri invariably explored the major features of qualitative research in contradistinction to those o( quantitative research. The second factor is the growing awareness of phenomenology, and more particularly Schutz's version of it, which occurred in the 1960s. This work seemed simultaneously to offer the epistemological basis lor a critique of quantitative research and a novel approach in its own right. One the one hand, the growth of phenomenology acted as a spur to congruent perspectives like symbolic interactionism and verstehen; on the other hand, it spawned interest in methods like participant observation and unstructured interviewing, which seemed to allow the phcnomcnological approach to be set in motion.
■^■^■^■^H
1
NATURE OF QUA! 11   MIM    RESEARCH     71
Notes
1   Pariieipunt ohscrvniion should not be confused with 'pine' observation. I his technique involves (he researcher obsei víny others, hut with no participation. It has been used by I.. I olland (1973) in her .studies of urban life, although she buttressed her observations with interviews. Many participant observers use pure observation some «*I the lime, but the relative absence of involvement with the subjects of the research has meant lhal pure Observation is rarely used alone by qualiiative researchers because It Is unlikely to allow access to the world-views of those being studied.
2  However, there arc signs of increasing interest in the use of the life history method («c Plummer, 1983).
3  A useful summitry of some major phcnomcnological ideas can be found in Husserl (1927), which also provides a flavour Of his style.
4  On conversational analysis and its relationship with the ethnographic phase nl elhnoincihodolow, see Atkinson und Drew (197«)
78