The following quote from Miles and Huberman (1994) well illustrates the sit-uation with regard to qualitative techniques of analysis.
The most serious and central difficulty in the use of qualitative data is that methods of analysis are not well formulated. For quantitative data, there are clear conventions the researcher can use. But the analyst, faced with a bank of qualitative data has very few guidelines for pro-tection against self-delusion, let alone the presentation of unreliable or invalid conclusions to scientific or policy-making audiences. How can we be sure that an ‘earthy’, ‘undeniable’, ‘serendipitous’ finding is not, in fact, wrong. (p. 2.)
In addition to highlighting the relative infancy, and some may say, crudity, of techniques of qualitative data analysis compared to quantitative data analysis, Miles highlights the central issue which this comparative infancy of qualitative data analysis techniques gives rise to, namely the issue of trying to ensure that the results through the interpretation of qualitative data are reliable and valid.
We have seen in Chapter 3 that the whole process of developing knowledge and understanding through research centres on the development of theories which are both reliable and valid. Both deductive and inductive research meth-ods centre on reliability and validity of data collection and analysis. As Miles points out, without these twin pillars of theory development and scientific research then, the researcher can never be truly sure, and certainly will find it difficult to convince others that his or her research findings are correct.
Much qualitative data analysis, it has to be admitted, does stem from the often subjective and individual interpretations of the researcher. Put another way, with much qualitative data two researchers could put entirely different interpretations on a set of data depending on their perspectives, experiences, agendas, predispositions and so on. Because of this, in the past, researchers wishing to develop knowledge and understanding, and particularly when try-ing to develop theories, have sought to avoid collecttry-ing and ustry-ing qualitative data, or alternatively have sought to develop techniques for transforming essen-tially qualitative data into a quantitative form through, for example, scaling techniques, and so on. However, as already mentioned, in the study of man-agement and organizations, it has increasingly been recognized that much of the data collected in the process of a research or management consultancy exer-cise will be, by its very nature, qualitative rather than quantitative. In addition, and related to this, it is increasingly appreciated that such qualitative data, far from being ‘inferior’ to its quantitative counterpart is in fact often the most powerful and useful data. If anything these days, the use of qualitative tech-niques of data collection and the need to analyse qualitative data, probably
predominate in management research and consultancy exercises. As a result, there has been a growing interest in it accompanied by the development of bet-ter and more powerful techniques for analysing qualitative data. As we shall see, most of these developments in qualitative data analysis centre on tech-niques for improving the reliability and validity of qualitative data analysis techniques. In other words, researchers have attempted to move towards tech-niques of analysis which fit the scientific model of theory development and analysis with an emphasis on codifying and classifying data, removing subjec-tive interpretations, and developing data which is amenable to statistical analy-sis and verification. We have to be careful here, in as much as in trying to impose a more rigid scientific approach to the analysis of qualitative data, we may in fact, by attempting to impose a system of order and consistency, which does not fit the natural data collected, detract from the essential nature of qualitative data and its potential value in providing ‘earthy’, ‘undeniable’, and
‘serendipitous’ findings which, although they cannot be proved ‘right’ (in a sci-entific sense) offer major insights into organizational and management issues.
In other words, we should not let a desire for more objective and scientific tech-niques of analysis panic us into imposing an artificial framework of rigour onto qualitative analysis, thereby detracting from the essential nature and inherent advantages of qualitative data analysis. Churchill (1999) captures this caveat well in referring to the analysis of focus group data.
One has to remember that the results (of focus group analysis) are not representative of what would be found in the general population, and us are not projectable … The unstructured nature of the responses, makes coding, tabulation, and analysis difficult. Focus groups should not be used therefore to develop head counts of the proportion of peo-ple who feel a particular way, focus groups are better for generating ideas and insights than for systematically examining them. (p. 113.) To reiterate, the very strength of qualitative data stems from its open-ended and often subjective nature, particularly when it comes to analysis. However, as already mentioned, the growth in qualitative techniques of data collection allied to a desire to develop more systematic and scientific techniques of analy-sis for such data has led to attempts to develop less subjective approaches to analysing quantitative data. It is now accepted that it is possible and poten-tially advantageous to try and develop more systematic techniques of analysis along the lines of the processes conventionally used in analysing quantitative data. Gradually then, analysers of qualitative data are moving away from sim-ply presenting their interpretations of thousands of words of field notes col-lected using qualitative techniques in ways in which the researcher simply informs us of the patterns drawn from this enormous amount of data without telling us how, or allowing other researchers to replicate the analysis of the data.
Similarly, researchers are now recognizing that analysis of qualitative data which enables the validity of findings to be assessed by others is indeed possible and often desirable. As already discussed, the main theoretical approaches to
qualitative data analysis which embody this relatively new acceptance of the possibility and desirability of more systematic/scientific analysis of qualita-tive data are those of content analysis and grounded theory. More recently however, the desire to develop more analytical methods of qualitative analy-sis have led to a welter of individual techniques for analysing qualitative data embodying more systematic and analytical frameworks. These include for example: cognitive mapping, pattern coding, cause and effect diagrams, case ordered effects matrices, scatter plots and perhaps one of the most interesting techniques of qualitative analysis to be developed, namely the technique of semiotics. Because of its relevance and potential usefulness to management research in organizations, we shall now consider semiotics as a technique in more detail.
11.4 Semiotics
There are many ways of analysing qualitative data that have already been dis-cussed, but in this section we are going to focus on one of these in more detail namely – semiotics or semiology, the names are interchangeable. This is an essentially interpretive methodology based upon linguistic theory. Language is a means of communication between people sharing a common culture, or at least with sufficient commonality to enable communication to take place. This cultural commonality can be in terms of a societal culture or a business culture.
In communication language provides not merely a representation of objects and events that the communicator of information has in mind but also a represen-tation of the desires, intentions and goals of the communicator. These are either consciously embedded into the communication according to the communica-tor’s intentions or unconsciously embedded despite the communicacommunica-tor’s inten-tions. As such these intentions are subject to analysis and interpretation.
In order for communication to take place, it is necessary for the language used to have some formal structure to ensure common understanding, and this is the function of syntax and grammar, as well as of the meaning ascribed to individual words. In speech, as opposed to written language, this formal structure of language is simpler and the rules are often broken, with meaning being partly given by contextual information as well as speech content. Thus it is possible for a conversation to be understood by all parties to it that would be meaningless if written down and shown to a third party because the vital contextual and implied content of the communication is missing. Such short-hand can, however, be considered to be nothing more than a special case of the elaborated code used by technical experts in the language. Halliday (1978) argues that the format of this code both determines social structure and is determined by such social structure, and that the use of language socializes the child into the adult.
I am grateful to David Crowther, London Metropolitan University, for providing the material on Semiotics.
Linguistic studies have shown that language is used to identify social class (Klein 1965; Labov 1966; Hewitt 1989) but that language also defines identity much more narrowly in terms of the social group to which one belongs. This view has been identified by Le Page (1968) who states:
Each individual creates the systems for his verbal behaviour so that they shall resemble those of the group or groups with which from time to time he may wish to be identified. (p. l94.)
Thus this view suggests that language acts like a membership card and assumes that language usage and behaviour is adopted to gain membership. However, a contrary argument from feminist discourse suggests that language is used as a source of power and dominance, and is used in this manner by the dominant group in order to exclude others. Thus, Lakoff (1975) considers language in the context of power and dominance, stating:
The language of the favoured group, the group that holds the power, along with its non-linguistic behaviour, is generally adopted by the other group, not vice versa. (p. 136)