A rchive Date
[ 21-11-2005 ]
Category
[ Science ]
sub-Categoy
[ Psychology ]
|
Grounded Theory Methodology. The Pressing Need for a Coherent Logic of Justification
David L. Rennie
SPECIAL SECTION: WINDELBAND ON ‘NOMOTHETIC' AND IDIOGRAPHIC'
Theory & Psychology 1998 Vol. 8 (1): 101-119
Department of Psychology
York University
Toronto, ON, Canada
Abstract. The originators of the grounded theory approach to qualitative research now disagree on certain procedural aspects of the methodology, while agreeing on others, and dispute its epistemological implications. In this article it is argued that the rift can be traced to a conflict over the logic of justification of the approach. Strauss and Corbin endorse Dewey's instrumentalism, including its prizing of the experimental method, and introduce a form of hypothetico-deductivism into the grounded theory method. Alternatively, although subscribing tacitly to the experimental method, Glaser does not tie it in with instrumentalism, and insists that grounded theory properly involves only the inductive phase of inquiry. It is argued that both instrumentalism and induction are inadequate as rationales for the grounded theory method. A new logic of justification, termed methodological hermeneutics and derived from Margolis's reconciliation of realism and relativism, has been developed by the author. When applied to the two positions, it leads to the conclusion that Glaser's procedures are the most consistent with the objectives of the method.
Key words: B.J. Glaser, grounded theory, A. Strauss
The grounded theory method was introduced by Glaser and Strauss (1967) as a discovery-oriented approach to the development of theory in sociology. This form of qualitative research has since been taken up in a number of social science disciplines, including psychology. It is attractive to inquirers who prefer to immerse themselves in data before leaping into theory. More fundamentally, perhaps, in being a form of qualitative research, it appeals to investigators who like to work with natural language rather than with numerics as signs of the phenomenon of interest, and who feel restricted by the natural science canons on how to make inquiries in social science. Among its applications in psychology, the method has been used in inquiries into topics in, for example, behavioral medicine (Pennington, 1988), community psychology (Pilowsky, 1993) and psychotherapy (see Polkinghorne, 1994).[p.101]
It is thus unfortunate that, with its use on the rise, Glaser and Strauss now disagree about many of its procedures and what can be concluded from them. Strauss and a colleague, Julia Corbin, have altered the original grounded theory procedures in several respects and emphasize that some of them must be implemented if the researcher is to claim to be doing grounded theory analysis (e.g. Strauss & Corbin, 1990). Moreover, they maintain that grounded theory analysis is more verificational than Strauss's earlier works with Glaser evidently led people to believe. Meanwhile, Glaser (1992) disputes these changes and the claim about verification, countering that it is Strauss and Corbin who are no longer doing grounded theory analysis.
This is not to say that the rift is total. Glaser and Strauss and Corbin still concur that constant comparative analysis, theoretical sampling and theoretical memoing (see below) are required and contribute to objectivity. They also agree that a grounded theory analysis reflects the perspective of the analyst, and, in this sense, both sides are in tune with the challenge to foundationalism and essentialism made by deconstructionists and postmodernists. But, curiously, both are also about equal in not tying their relativism in with the modernist-postmodernist debate.
Although they do not characterize it as such, their dispute appears to result from a fundamental disagreement about what amounts to the logic of justification of the methodology. Doubtless influenced by his affiliation with symbolic interactionism (Blumer, 1969), Strauss has noted from time to time that the approach expresses Dewey's instrumentalist form of pragmatism (J. Dewey, 1938/1991).2 In their writings, Strauss and Corbin (Corbin & Strauss, 1990; Strauss & Corbin, 1990, 1994) bring the method in line with instrumentalism, becoming more rationalistic in the bargain. Coming from a background in discovery-oriented, quantitative sociology, Glaser complains that their method undermines the discovery-orientation of the approach and ironically throws it back to the very kind of theorizing that the method is meant to challenge.
Also, in keeping with Dewey's valuing of the experimental method, Strauss and Corbin advocate a combination of deduction and induction as a mode of verification during grounded theory analysis, whereas Glaser insists that the approach is purely inductive and leads only to theory, not its verification; at the same time he maintains that the categories developed within an analysis are validated in virtue of the method itself. On balance, some of the current differences between Glaser and Strauss are clear, while others are subtle.
Under the assumption that method as procedure needs to be in keeping with method as logic of justification (J.K. Smith & Heshusius, 1986), when applied to the dispute in question, this requirement raises a number of questions. Given that Strauss has never specified the main tenets of instrumentalism, to what extent can his new procedures be seen as expressions of it? Alternatively, how well does induction as Glaser characterizes it [ p. 102] explain his procedures? Beyond these particulars, granted that there are procedures about which Glaser and Strauss still agree, how compatible are these with either instrumentalism or induction? As an extension of the last puzzle, when the procedures about which they both agree are looked at squarely, what sense is to be made of the conflicting claims about verification? Finally, there is an overarching question about where all of the similarities and differences between Glaser and Strauss and Corbin fit with the current modernist-postmodernist debate about ontology and epistemology (e.g. Gergen, 1994; M.B. Smith, 1994).
The following essay represents the results of a search for the answers to these questions. The approach used is both deconstructive and constructive. In the first phase, the attempt is made to identify the rationale Glaser, on the one hand, and Strauss and Corbin, on the other, appear to use in support of their research practices, and to appraise those practices and claims made for them in the light of the rationales. In the second phase, a new logic of justification (Rennie, 1996) is applied as a way of evaluating the two approaches. The argument begins with a summary of the features of the method about which both parties still agree.
Aspects of the Grounded Theory Method About Which Glaser and Strauss Both Agree
The grounded theory method is made up of a set of procedures designed to promote the systematic study and representation of the meaning of data. The goal of the method is to develop a theory that is accountable to the data under study. The gathering of the data and their analysis proceed concurrently. Reduction in the complexity of the data is achieved by the conceptualization of what they appear to have in common. This objective is achieved through the procedure of constant comparison. Each datum is compared to other data and commonalities among data are represented by codesand categories (to reduce confusion about these two terms, I prefer to use only the latter; see Rennie, Phillips, & Quartaro, 1988). A given datum is assigned to as many categories as seems fitting. As categories are conceptualized, interest is taken in the apparent relationships among them. Categories in turn may be grouped according to a meaning that appears to unite them. Grouped categories are named as a higher-order category. The categories it subsumes are considered its properties.
This categorizing of categories is carried through higher orders of abstraction, with the objective of conceptualizing a supreme category that colligates all others. This category is referred to as the core category, the key concept that organizes the theory. Meanwhile, attention is paid to the extent to which the taxonomy that develops is adequate to account for the meanings in new data as the sampling of data continues. When it is judged that the new data add no [p.103] further meaning and hence require no further categorization, the taxonomy is judged to be saturated. The researcher may then wish to address the generalizability of the taxonomy by sampling from alternative sources of data (theoretical sampling). In the grounded theory method, analysts attempt to keep their biases at bay by recording them in a research diary (theoretical memoing)at all stages in the inquiry; by attempting to set biases aside (i.e. to put them in 'brackets', as phenomenologists describe the tactic); and by being as explicit as possible about them when writing up the returns from their studies. Nevertheless, the objectivity of the analysis rests with the objectivity of its groundedness. It is not the objectivity of the 'God's-eye-view' (Putnam,1990) of the phenomenon under study.3
I now turn to Dewey's instrumentalism, the apparent rationale for the approach at the hands of Strauss and Corbin.
Dewey's Instrumentalism
Most commentators agree that Dewey's instrumentalism as a logic of inquiry needs to be understood in terms of his life-long identification with Hegel's thought (e.g. Jane M. Dewey, 1951; Murphy, 1951; Santayana, 1951; Thayer, 1968). Like his fellow pragmatists (Peirce, James, F.C.S. Schiller, Mead), Dewey became impatient with the seemingly insoluble difficulties created by the dualisms of Enlightenment philosophy (e.g. appearance-reality, subject-object, mind-body) prompted by two contrasting metaphysics—idealism and materialism. Nevertheless, he was comforted by Hegel's absolute idealism as an antidote to what Dewey felt to be the nihilism of monistic materialism, despite a concerted attempt to stay within a monistic naturalism.
In Dewey's naturalism, which drew upon Darwin, there is experience— life adapting to the problems of living.4 Experience is of a situation and is continuous with what has gone before and is to come. Experience is action that becomes efficient as habits. Habits do not always suffice for adaptation, however, and, in people, when habits fail action occurs as deliberation, planning. In this action, various operations are hypothesized and their consequences are considered, culminating in conduct of the preferred plan of action. The conduct has consequences which, if satisfactory, pave the way to a new set of habits. Hence, experience is in nature and is an ongoing process of experimentally solving problems of living in the service of adaptation. This is not only experimentation as used in the laboratory, however; it is the experimentation involved in the solving of problems in all walks of life. Perhaps most important, it is experimentation in the service of growth, purpose.
Overall, in instrumentalism, there is no 'mind'; there is covert and over action. There is no Truth; there is warranted assertibility. There is no [p.104] final test of warranted assertibility; there is eventual agreement in a community of inquirers (J. Dewey, 1922, 1929, 1929/1988,1938/1991; Piatt, 1951; Santayana, 1951; Thayer, 1968; White, 1964).
In keeping with pragmatism as a school of thought, instrumentalism is innovative and bold. It has also given rise to criticism. It has been challenged on the grounds that the optimism of growth and progress and the idealization of a democratic community of inquirers appear to go beyond the boundaries of monistic naturalism (Murphy, 1951; Parodi, 1951; Thayer, 1968). It has also been observed that instrumentalism is deeply psychological but in a way that reflects little interest in the particular psychological mechanisms and structures playing a part in the development of knowledge (Allport, 1951). Accordingly, this lack of specification makes it difficult to know where to start should an attempt be made to study the psychology of knowing. In the same vein, Allport points out, Dewey had little interest in the Geisteswissenschaften (human studies) as promulgated by Wundt because they relied on introspection, in Dewey's view.
The Method of Strauss and Corbin
Strauss and Corbin's modification of the basic method is fourfold: the investigator's recalled experiences pertaining to the phenomenon under study are accepted as legitimate empirical data; hypothesis testing is made integral to constant comparison; consideration of the conditions influencing the phenomenon should not be limited to those indicated by the data themselves; and the application of an axiomatic schema that converts all social phenomena into a process is made mandatory (Corbin & Strauss, 1990; Strauss, 1987; Strauss & Corbin, 1990, 1994). Each of these changes is addressed in turn.
In the initial version of the method, what are considered data are signs (commonly, written text) external to the analyst, and the analyst's conceptualization of the relations among the signs is to be supported by the signs themselves. In contrast, in their revision, Strauss and Corbin broaden acceptable signs to include the analyst's own recalled experiences. In short, although not exclusively so by any means, Strauss now makes introspection a contributing source of data bearing on the phenomenon under inquiry.
With respect to hypothesis testing, Strauss and Corbin hypothesize relations that are not necessarily supported by the data under consideration at that point in the analysis. Instead, the relations are conceptualized on rational grounds whereupon succeeding data are looked for (including, in keeping with this amendment, the data of the analyst's own experience) as confirmations or disconfirmations of the hypothesized relation(s).
In terms of conditions bearing on the phenomenon under study, these methodologists advocate that the understanding of a given category should [p.105] be broadened by taking into consideration all of the conditions that can be thought to influence the experience represented by the category. The conditions taken into account need not be indicated by the data themselves. Instead, conditions should be introduced rationally under the assumption that such conditions must be at play. They urge the use of Strauss's conditionalmatrix, which schematizes conditions bearing on social phenomena, beginning with action pertaining to the phenomenon's immediate environment and extending to international influences. 5
Lastly, Strauss and Corbin maintain that the meaning of a given category must be enriched by applying to it a coding paradigm. This is a schema consisting of 'conditions, context, action/interactional sequences and consequences' (Strauss & Corbin, 1990, p. 96). The net effect of this modification is to make all social phenomena processual, a change from the earlier version of the methodology in which analysts are encouraged to develop categorical systems that are either structural, processual or both, depending on the interpreted meaning of the data.
Glaser (1992) criticizes all of these changes. He is concerned that in supplementing external signs with personal experience, Strauss promotes movement away from groundedness and toward subjectivity and dogmatism. Glaser proposes that, to the extent that this drift occurs, it restores the rational approach to theorizing that the grounded theory method was designed to offset. Secondly, he disputes the notion that the loose form of within-analysis hypothesis testing is a form of verification. Instead, he asserts that verification requires the application of the testing of hypotheses in the normal way, by which he appears to mean the way of natural science, subsequent to the development of a grounded theory. At the same time, while disclaiming that grounded theory research praxis is verificational, he makes a claim for validation, remarking, 'While creativity is necessary for generating categories and their properties, the researcher must always validate their fit and relevance by saturation, interchangeability of indices, relationship to the core categories and integration into the emerging theory' (Glaser, 1992, p. 18). I shall return to this point presently.
It is with Strauss and Corbin's stress on the coding paradigm and the conditional matrix, however, that Glaser takes greatest issue. Glaser argues that social phenomena are not necessarily processual and that to insist that they are may force the data. He is not against the use of theoretical schemata to facilitate understanding of the relationships among categories and the development of higher-order categories, but only if they fit the data. He points out that in Theoretical Sensitivity (Glaser, 1978) he lists Strauss and Corbin's coding paradigm as one of 18 possible theoretical schemata and advises the reader that this list in turn is merely suggestive. Glaser appears mystified as to why Strauss and Corbin have seized upon the one theoretical schema and require its application. In terms of the conditional matrix, Glaser cautions that conditions it specifies should be entertained only if they are [p.106] called for by the data; to do otherwise once again forces a priori conceptualizations upon the phenomenon of interest. As with the coding paradigm, Glaser is vexed and seemingly bewildered by Strauss and Corbin's fixation upon conditions.
In most respects, Glaser (1992) advocates the method as specified in the first two works on it (Glaser, 1978; Glaser & Strauss, 1967). His main change is to render more visible the distinction between validation and verification, as he sees it, in support of his argument that the purpose of the method is to develop theory that is accountable to data, not to verify theory, and certainly not to import rationally theory that is supported by forcing the data to fit it (Glaser, 1992).
I shall now turn to the compatibility between the implicit assumptions supporting the latest thinking of Glaser and Strauss and their actual research praxis, turning first to Strauss and Corbin.
Grounded Theory Methodology and Instrumentalism
Compatibility
Instrumentalism as a logic of inquiry recognizes that perspectivism is inevitable in the development of knowledge. Problems have many facets. Moreover, individuals will approach the same problem from different vantage-points. The result is that solutions to problems must be expected to be pluralistic, but that need not imply solipsism; commonality in judgment about the warrant of an assertion is made possible by subjecting it to the scrutiny of the community of inquirers. In the same way, in Strauss and Corbin's grounded theory method, it is recognized that different analysts of the same data are to be expected to formulate somewhat different categories/theories; this perspectivism is accepted with the proviso that all formulations should be grounded. Moreover, apart from any other procedures that might be conducted to test the credibility of a formulation it is assumed that the procedure of grounding the formulation will result in it resonating with members of the community, in it being judged credible in its own right. Hence, this characteristic of grounded theory, which Strauss has left untouched, is consistent with instrumentalism.
What about the modifications? In several respects, there is compatibility between the philosophy and the method. Under their tutelage, all social phenomena are processual; they are necessarily adaptational, involving context, operations and consequences; and they are influenced by far-ranging conditions—all a straightforward expression of Dewey's process of problem-solving. Moreover, Strauss and Corbin's emphasis on hypothesis testing is in keeping with Dewey's experimentalism. 6. [p.107]
Incompatibility
There are two aspects, however, of the grounded theory method that do not fit instrumentalism as a logic of justification. The first applies to all versions of the method, while the second is an outgrowth of the first that has particular bearing on Strauss and Corbin's revision.
The matter of what counts as legitimate data. The first concern has to do with the nature of data. The grounded theory method accommodates any type of data as long as they can be taken to mean something. Accordingly, grounded theory methodologists have no difficulty with verbal reports; instead, they generally prefer them. Dewey's monistic naturalism, however, is not compatible with the use of verbal reports as data. In instrumentalism, reports on inner experience are ruled out because of the dualism created by the notion of an inner condition, or state, that is reported on through the use of language as a mode of representation. Moreover, even reports on past, overt experience are not satisfactory because they invariably are influenced by the context of the present situation. Any inner activity such as deliberating or planning is relevant only as it eventuates in future, overt action, with overt consequences.
Experience as structure. In his penetrating analysis of the development of Dewey's social psychology, Allport (1951) observes that Dewey's Logic: The Conduct of Inquiry marks the fullest expression of his functionalism. By this stage in the development of his thought, Dewey had no interest in states of experience, such as attitudes; what counted was the use of intelligence (which he called inquiry) in the promotion of adaptation. Strauss and Corbin have expressed this instrumentalism methodically through insistence that the coding paradigm is required in grounded theory analysis. In the application of this axiom, all experience is converted to process. Correspondingly, experience as stable structure is rendered inadmissible as an ontological category.7
The question has to be raised as to whether this position is sound. Allport was concerned about its epistemological implications. On the one hand, he mused that Dewey may have been ahead of his time in stressing that experience is an ever-changing fluid process. On the other hand, Allport felt that to take this position makes any attempt to classify human experience futile and hence puts the science of psychology (and, it might be added, all social science) on a foundation of shifting sand.
Dewey's preoccupation with the conduct of inquiry in service of adaptation was driven by his great and enduring concern about the condition of people in the world combined with his passion to break free from entanglements of dualism. It was on these rational grounds, and not on the grounds of empiricism, that he derived his ontology and epistemology. In a similar vein, Strauss's insistence on his conditional matrix and coding paradigm for [p.108] grounded theory method appears to have come about more through rational than empirical and/or semiological considerations. Certainly this is Glaser's complaint.
Glaser's Appeal to Induction
In his fervent claim that the grounded theory method produces only theory, Glaser is staying on the high ground of the experimental method of natural science: His grounded analysis is inductive. This is its strength. Induction keeps the investigator close to the data, and justifies the assertion that the claims to understanding are grounded. Verification is another matter entirely. It is deductive. It is not part of a grounded analysis; it comes later. Yet, as we have seen, Glaser also believes that the method itself inherently is validational, through various indicators of internal consistency.
Glaser thus holds on to both verification and validation. Verification is valued but is not a part of a grounded theory analysis, while validation is both valued and a part of the method. In protecting the grounded theory method from the backslide into rationalistic hypothetico-deductivism at the hands of Strauss and Corbin, Glaser need not talk in the language of verificationism. This concept of the logical positivists (who came to be disenchanted with it; C. Green, personal communication, April 1996) has been allied by psychologists with behaviorism, positivism and the correspondence theory of truth. The grounded theory methodology looked at squarely involves none of these things, however. Its innovation compared to the received approach to method in social science is to work with the subjectivity and reflexivity of both the subject being addressed and the researcher doing the addressing. The theories it produces are about the meaning of experience—they are 'emic' as much as 'etic'—and it is very difficult to preserve the subjective by appealing to the objectivism of natural science methodology.
On the other hand, as I show below, Glaser's claim that the method is validational is correct. He argues his case weakly, however, because of the theory of inference to which he subscribes. This theory is that inference entails either induction or deduction, and that support for the tenability of the inductions arises from deducing hypotheses from the inductions, and testing the hypotheses. In the received approach to method in psychology, the emphasis on the deductive mode of inference is so strong that it comes to be seen as the key component of inference, as evidenced by the term 'hypothedco-deductive' method, Apart from this bias, the difficulty with the two-component theory of inference is that it provides no way of supporting validity claims for the original inductions, other than the deductive step. Yet this is precisely what Glaser attempts to do because he tries to explain, in the language of inference, the checks for internal [p.109] consistency that constitute the grounded theory method. But he can appeal only to the checks themselves; he cannot tie them in with the theory of inference to which he subscribes.
The foregoing analysis leads to the conclusion that the monistic naturalism of instrumentalism cannot accommodate the assumption made in the grounded theory method that there is private experience representable by verbal reports, and that the emphasis in instrumentalism on process to the exclusion of structure is suspect. On the other side of the coin, the analysis indicates that Glaser's intuition about the grounded theory method being inherently validational is sound, but needs to be supported by a better theory of inference than the one to which he subscribes. Accordingly, it is apparent that neither instrumentalism nor the received logic of method serves well as the logic of justification of the grounded theory method, necessitating the need for a new direction.
A New Logic of Justification for the Grounded Theory Method
The two markers creating a sight-line for an appropriate logic of justification are the realistic and the relativistic aspects of the method. An appropriate logic of justification needs to take into account the relativity of the grounded theorist's formulation to his or her horizon of understanding but in a way that is not as constraining as instrumentalism and the experimental method of natural science.
Recently, I developed a logic of justification, called methodical hermeneutics (Rennie, 1996), for grounded theory.8 Summarizing its development and features briefly, I turned to the works of Margolis (1986, 1987, 1989, 1995) to establish a metaphysical and epistemological base from which to derive coherently implications for method. Margolis argues compellingly for the reconciliation of realism and relativism, maintaining that Descartes's mind-body dualism, leading to subject-object dualism and the appearance reality aporia, was a mistake to begin with. He declares that thinkers like Merleau-Ponty (1962) are correct in asserting that we are embodied and that our qualities of human beingness are incarnate expressions of that embodiment. He also subscribes to the insight of contemporary continental philosophers such as Wittgenstein (1958) and Heidegger (1927/1962) that, rather than being dualistically separated from the world, we are symbiotically related to forms of life in the world. In turn, then, the notions of foundationalism, essentialism and truth as correspondence to reality are untenable. Instead, all knowing is relative to the perceptual framework of the knower. This relativism is not subjective idealism, however; in virtue of the symbiosis of ourselves and the world, it is a relativism about the real world. Hence, Margolis reconciles a relativistic epistemology with a non-reductive materialist metaphysics. [p.110]
Drawing upon this metaphysics and epistemology, I argue that the human sciences, especially, are hermeneutical in virtue of their involving an interpreting subject addressing a self-interpreting referent. Correspondingly, then, they are rhetorical. This is not an empty rhetoric, however, because, in virtue of the reconciliation of realism and relativism, there are saliencies that apply to arguments, making some interpretations better than others. The grounded theory method instantiates this hermeneutics more completely than do the more conventional, received approaches to method in the social sciences. Moreover, the method tends to reduce the danger of vicious circularity and the limitation to particularism characteristic of traditional hermeneutics. This gain is achieved by integrating hermeneutics with Peirce's theory of inference (Peirce, 1965) and the phenomenological procedure of bracketing.
According to Peirce, there are three modes of inference—abduction, induction and deduction. Abduction is hypothesizing, induction is a testing of abductions, and deduction is the demonstration of apodictic truth by deriving a conclusion that is entailed tautologically in its premises. Thus, for Peirce, new knowledge is only abductive. Abduction does not require a method of inquiry, however, although it is not exclusive of method. Instead, it is any form of new idea, including intuitions and hunches. Within this framework, an interesting feature of qualitative research is that it involves the symbiosis of abduction and induction. This symbiosis is achieved by bracketing preconceptions of the phenomenon of interest and delaying the development of conceptualizations until they are derived from immersion in the data pertaining to the phenomenon. Moreover, the adherence to a conceptualization thus conceived is contingent on its being supported by the data as a whole. Thus, abductions are derived through the activity of induction. Symbiotically, they are validated in virtue of that same induction.
Thus, although the grounded theory method is hermeneutical, it improves on traditional hermeneutics. In the method, the conceptualization of concepts from particularities liberates hermeneutics from its emphasis on particularities, characteristic of traditional hermeneutics. Moreover, there are positive implications for the hermeneutic circle, which is the cyclical understanding of particularities by initially projecting onto them a 'whole', or frame of reference, which in turn is influenced by the (interpreted) meaning of the particularities. In traditional hermeneutics, the whole that is projected tends to be rationally derived, which potentiates the danger that the hermeneutic circle may become vicious. In grounded theory, on the other hand, vicious circularity is reduced because the holistic understandings applied to particularities are derived from the particularities themselves, in virtue of the combination of bracketing and the symbiosis of abduction and induction. [p.111]
Another feature of this logic of justification, derived as it is from the reconciliation of realism and relativism, is that it takes into account the postmodern critique of essentialism and foundationalism, while retaining the modern concepts of self and reference. As such, it addresses constructively the current debate among discourse analysts, in particular (see Madill & Doherty, 1994), and postmodernistic qualitative researchers of all types about the question of authorial intent (cf. Denzin & Lincoln, 1994). Authorial intent is retained without sacrificing social constructionism (cf. Gergen, 1985). Looking at the grounded theory method along these lines provides a coherent framework for making appraisals of Glaser and Strauss's versions of it.
Strauss and Corbin's and Glaser's Methods in the Light of the Logic of Justification
I shall begin this critique by attending to Strauss and Corbin's amendments to procedure involving the heightened use of rationally driven hypothesizing, the required use of the conditional matrix and the coding paradigm, the analysts' use of self-experience as data, and their claims for verification on the strength of their use of hypothesis testing during the course of a grounded analysis.
Regarding hypothesizing on the grounds of reason, although it is useful to be actively involved in the analysis of the data, the whole point of the grounded theory method is to listen to the data and to avoid undue theoretical sensitization to them. There is a danger that systematic and deliberate hypothesizing from the earliest stages of the analysis may predispose the analyst prematurely to import theoretical frameworks. In terms of the logic of justification, the rationalism involved increases the likelihood that the hermeneutic circle will be vicious. Apart from this consideration, there is a practical consequence. The transcripts of Strauss's training sessions with his students (Strauss, 1987) suggest that this procedure slows the analysis to a crawl—an aspect that has vast implications for the scope of the phenomenon that analysts might wish to address when contemplating adherence to this procedure. Hence, both the logic of justification and practicality call for less rationalistic hypothesizing than is advocated by Strauss and Corbin.
In terms of the application of the conditional matrix and the coding paradigm, in principle they are coherent with the logic of justification in the sense that, in being an interpreter, the grounded theorist has to start from somewhere in entering into the hermeneutic circle. There is a danger posed by the insistence on the conditional matrix and the coding paradigm, however: namely they increase the chances of making the hermeneutic circle vicious; after all, 'automatic' appeal to these structures is a form of [p.112] rationalism. In this sense, Glaser is correct in his concern that these procedures force the data.
Regarding Strauss and Corbin's move toward the deliberate use of self experience as a source of empirical data, the difficulty raised by this procedure is that it is not in keeping with the symbiosis of abduction and induction; thus it, too, is another way of developing conceptualizations that are insufficiently grounded. More broadly, too much of a shift in this direction results in a method closer to that of philosophical phenomenology (see Spiegelberger, 1960) and, correspondingly, in a drift away from the empirical and semiotic interests of the social sciences.
Lastly, with respect to Strauss and Corbin's claims about verification, a problem is created by applying this concept to the grounded theory method in that the concept, apart from its outmodedness, orients analysts toward positivism and the correspondence theory of truth. When the realism and relativism of the grounded theory method are taken into account, it is coherent with the logic of justification to think in terms of the plausibility of the category, structure or theory rather than in terms of their verification (see also J.K. Smith, 1989; J.K. Smith & Heshusius, 1986). More fundamentally, when, in giving expression to Dewey's fondness (in theory) for the experiment as the finest form of inquiry, Strauss and Corbin turn to hypothesis testing within a grounded theory analysis, they fail to appreciate sufficiently the validational impact of the symbiosis of abduction and induction during constant comparative analysis.
Turning to Glaser, the logic of justification strengthens his claim that a grounded theory analysis is inherently validational. It does so in virtue of the symbiosis of abduction and induction during constant comparative analysis. Beyond that, however, the logic raises a caveat about Glaser's insistence that, in effect, the procedures of normal science are necessary to verify a grounded theory. Apart from Glaser's untimely subscription to the outmoded logical positivistic concept of verification, interpreting the method within the framework of the logic of justification means that Glaser need not turn to normal science as the ultimate court of appeal. In virtue of its integration of hermeneutics, phenomenological bracketing and the symbiosis of abduction and induction, the method is sufficient unto itself. This is not to say that a grounded theory might not be strengthened through the procedure of deriving hypotheses from a grounded theory and subjecting them to test. So it might, but only as another approach to induction, not deduction because, as seen, deduction is merely tautological in Peirce's theory of inference. Moreover, the logic of justification points out that all truth claims in qualitative research are rhetorical. This being the case, it is coherently possible for a grounded theory to be persuasive in its own right; it may not require the successful testing of hypotheses derived from it in order to convince the researcher's audience of its truth value. [p.113]
Conclusion
When Glaser and Strauss developed their grounded theory method in the 1960s, they came up with something rather wonderful compared with what was customary research praxis in sociology. It turned method upside down: instead of using data to test theory, they were used to develop it. Glaser and Strauss made their achievement despite coming from quite different backgrounds. (Interestingly, on this score, Glaser recently has wondered if he and Strauss ever really understood each other; see Glaser, 1992.) Glaser was trained in inductive, quantitative sociology, whereas Strauss was a symbolic interactionist of the Chicago School. The development of their method appears to have been very pragmatic. In their early works, little attention was paid to other qualitative research approaches. The technique of bracketing was incorporated with no mention of its origins in phenomenology. Ethnomethodology (Garfinkel, 1967) grew up with grounded theory, unrecognized by Glaser and Strauss. Despite their anachronistic claim that it is acceptable for different sets of grounded theorists to produce different theories, deconstructionism and postmodernism have never been appealed to in support of their position.
This insularity is perhaps more understandable in the case of Glaser than Strauss. In his 1992 book Glaser reveals that he has been, first and last, influenced by the principles of quantitative, exploratory sociology. In this sense, he has always been of the 'old school', albeit a reformed member. Strauss is different. His kinship with symbolic interactionism (Blumer, 1969) meant that he rubbed shoulders with the relativism of Dewey's epistemology, shared by Mead in his social behaviorism. Yet, in his seeming attempt to reconcile Dewey's instrumentalism with the method, Strauss has not made much effort to question the merits of this line of thinking for the objectives of the grounded theory approach. Nor does he show any sign of capitalizing on Dewey's relativism by exploring how that theme has been taken up by current thinkers claiming to have been influenced profoundly by Dewey and who declare themselves to be postmodern; I am thinking especially of Rorty (1979).
The upshot is that, in their original version of the method, Glaser and Strauss created something that is more profound than either seems to have realized. In one way or another, both tacitly remain attached to the methods of natural science. Yet, in some ways, each is also tacitly in tune with the more contemporary philosophy of inquiry. Strauss has a zesty desire to justify the method as self-sufficient. Irrespective of his unfortunate alliance with Dewey in attempting to establish that point, the attempt alone makes him kin with qualitative researchers of all persuasions who advocate that their research praxis is an approach to social science that is adequate in its own right and not, as its natural science critics might hold, merely ancillary [p.114] to 'real' science. Moreover, like researchers influenced by poststructuralism and postmodernism, Strauss recognizes explicitly the perspectivism involved in inquiry. Glaser, on the other hand, has a surer sense of where the real strength of grounded theory lies, which is precisely in its open-ended, discovery orientation. In insisting on this, he keeps rationalism at bay; at the same time he appreciates perspectivism.
As indicated, I see the method as involving reconciled realism and relativism, a non-foundational theory of truth, hermeneutics, bracketing and Peirce's logic of inference. As such, the application of the logic of justification lends support to Glaser's method and to the criticism of the approach advocated by Strauss and Corbin. Once Glaser's method is explained in terms of the logic of Justification suggested, it is coherent with its stated objectives. On the other hand, Strauss and Corbin's version does indeed tend to throw the baby out with the bathwater, given that it remains ostensibly about groundedtheory.
This conclusion is perhaps surprising in terms of the comparative popularity of the two versions. The works of Strauss and Corbin have a resounding edge over Glaser's monographs, according to the citations in the literature. The appeal of the Strauss and Corbin version probably comes from its promise of simplicity, procedural structure and verifiability. The first two ingredients provide guidance to an approach to inquiry that is rather bewildering in the early going, and the last gives reassurance in a positivistic climate, still strong in psychology, in particular, but in sister disciplines as well. Glaser's defense of what amounts to the original approach is less well marketed, and it is looser and thereby perhaps more discomfiting at first glance. Even so, it is sound, sufficient in its own right despite Glaser's disclaimers to the contrary, and a much better road to conceptualizing something 'really new' about reality. In order to be appreciated for what it is, however, Glaser's approach needs to be viewed within a larger framework than the one he provides—a framework like methodical hermeneutics.
Notes
1. Foundationalism is 'the belief that we possess a privileged basis for cognitive certainty' (Margolis, 1986, p. 3). It is thus an aspect of the Cartesian approach to epistemology, and ties in closely with essentialism, or 'realism construed in an essentialist manner (the belief that ... the structure of the actual world is cognitively transparent to, or representable by, us' (Margolis, 1986, p. 3). Both Glaser and Strauss are unclear on where they stand on this kind of epistemology. As indicated, on the one hand, they disagree with it when maintaining that the grounded theory method is in effect a template that accommodates different [p.115] perspectives between disciplines, and between investigators in the same discipline. On the other hand, they tend to ignore the impact of such relativism when addressing the relationship between the given analyst and his or her data, and slip into essentialist language such as 'discovering' and 'finding' categories or theories. More will be said about this incoherence below.
2. Blumer (1969) proposed symbolic interactionism as an approach to sociology that drew upon Mead's (1934/1967) social behaviorism, in turn related in some reports to Dewey's instrumentalism. In the same work, Blumer outlined an ethnographic method of inquiry that bears a striking resemblance to the principles of grounded theory advanced in the early works of Glaser and Strauss, which is not surprising in that Strauss and Blumer shared the same department. It is pertinent to the present analysis that Blumer has been attacked as not being a Meadian, despite his protests to the contrary, on the grounds that his interpretative method of inquiry violated the experimental method prescribed by Mead (cf. Blumer, 1977, 1980; Lewis, 1976; McPhail & Rexroat, 1979; Mead, 1917/1964).
3. It would deviate from the purpose of this essay to attempt to describe more fully these aspects of the method. More complete overviews of the method in line with its original formulation are available (e.g. Rennie, Phillips, & Quartaro, 1988) and, of course, the original works give the manualization in greater detail (Glaser, 1978; Glaser & Strauss, 1967). Recent summaries of Strauss's latest position are also available (Corbin & Strauss, 1990; Strauss & Corbin, 1994), as are fuller manuals (Strauss, 1987; Strauss & Corbin, 1990). Glaser's current position is also manualized (Glaser, 1992).
4. His difficulty in defining 'experience' is notorious. Thayer (1968) observes that at one point in his career Dewey reflected that the term 'culture' would perhaps have been more appropriate.
5. The conditional matrix is presented as levels of influence on the phenomena: international; national; community; organizational-institutional; suborganizational and sub-institutional; collective, group and individual; interactional; and action (Strauss & Corbin, 1990, pp. 162-164).
6. Glaser stresses the importance of hypothesis testing as well but, as indicated, comes at it from a different direction; even so, this is an aspect of Glaser's position that is coherent with instrumentalism, albeit in a narrow sense.
7. Although his use of the concept 'habit' would seem to imply that Dewey recognized the importance of structure, this would be a false conclusion; over the course of his thought he increasingly characterized habit as a series of potentialities for conduct, the exact expression of which is influenced by conditions and context. Dewey thus gradually brought the concept of habit in line with his basic conception of adaptation as an ongoing process of change (Allport, l95l).
8. In my paper on methodical hermeneutics (Rennie, 1996), I argue that this logic of justification applies not only to the grounded theory method but to empirical phenomenology, narrative analysis and discourse analysis as well. Critics of this paper have judged that I am too quick to generalize the logic to these other genres of qualitative research. I have yet to do the additional investigation required to address the soundness of this criticism. In the interim, I have decided to limit the applicability of the logic to the grounded theory method. [p.116]
References
Allport, G.W. (1951). Dewey's individual and social psychology. In P.A. Schilpp Ed.), The philosophy of John Dewey (2nd ed., pp. 263-290). New York: Tudor.
Blumer, H. (1969). Symbolic interactionism: Perspective and method. Englewood Cliffs, NJ: Prentice Hall.
Blumer, H. (1977). Comment on Lewis' 'The classic American pragmatists as forerunners to symbolic interactionism'. The Sociological Quarterly, 18, 285-289.
Blumer, H. (1980). Mead and Blumer: The convergent methodological perspective of social behaviorism and symbolic interactionism. American Sociological Review, 45, 405-419.
Corbin, J., & Strauss, A. (1990). Grounded theory research: Procedures, canons, and evaluative criteria. Qualitative Sociology, 13, 4-21.
Denzin, N.K., & Lincoln, Y.S. (Eds.) (1994). Handbook of qualitative research. Thousand Oaks, CA: Sage.
Dewey, Jane M. (1951). Biography of John Dewey. In P.A. Schilpp (Ed.), The philosophy of John Dewey (2nd ed., pp. 1 45). New York: Tudor.
Dewey, J. (1922). Human nature and conduct. New York: Carlton House.
Dewey, J. (1929). Experience and nature. New York: Norton.
Dewey, J. (1988). The quest for certainty. Carborldale: University of Southern Illinois Press. (Original work published 1929.)
Dewey, J. (1991). Logic: The theory of inquiry. Carbondale: University of Illinois Press. (Original work published 1938.)
Garfinkel, H. (1967). Studies in ethnomethodology. Englewood Cliffs, NJ: Prentice Hall.
Gergen, K. (1985). The social constructionist movement in modern psychology. American Psychologist, 40, 266 275.
Glaser, B.J. (1992). Emergence vs forcing: The basics of grounded theory analysis. Mill Valley, CA: Sociology Press.
Glaser, B.J., & Strauss, A. (1967). The discovery of grounded theory: Strategies for qualitative research. Chicago, IL: Aldine.
Heidegger, M. (1962). Being and time. New York: Harper & Row. (Original work published 1927.)
Lewis, J.D. (1976). The classic American pragmatists as forerunners to symbolic interactionism. Sociological Quarterly, 17, 347-359.
Madill, A., & Doherty, C. (1994). 'So you did what you wanted then': Discourse analysis, personal agency, and psychotherapy. Journal of Community and Applied Social Psychology, 4, 261-273.
Margolis, J. (1986). The persistence of reality 1: Pragmatism without foundations: Reconciling realism and relativism. Oxford: Blackwell.
Margolis, J. (1987). The persistence of reality 2: Science without unity: Reconciling the human and natural sciences. Oxford/New York: Blackwell. [p.117]
Margolis, J. (1989). The persistence of reality 3: Texts without referents: Reconciling science and narrative. Oxford/New York: Blackwell.
Margolis, J. (1995). Interpretation radical but not unruly. Berkeley: University of California Press.
McPhail, C., & Rexroat, C. (1979). Mead vs. Blumer: The divergent methodological perspectives of social behaviorism and symbolic interactionism. American Sociological Review, 44, 449-467.
Mead, G.H. (1964). Scientific method and the individual thinker. In J. Reck (Ed.), Selected writings of George Herbert Mead (pp. 171-211). Chicago, IL: University of Chicago Press. (Original work published 1917.)
Mead, G.H. (1967). Works of George Herbert Mead: Vol. 1. Mind, self, and society from the standpoint of a social behaviorist (Ed. and with an introduction by C.W. Morris). Chicago, IL: University of Chicago press. (Original work published 1934- Merleau-Ponty, M. (1962). Phenomenology of perception (C. Smith, Trans.). London: Routledge, & Kegan Paul.
Murphy, A.E. (1951). Dewey's epistemology and metaphysics. In P.A. Schilpp (Ed.), The philosophy of John Dewey (2nd ed., pp. 193-235). New York: Tudor.
Parodi, D. (1951). Knowledge and action in Dewey's philosophy. In P.A. Schilpp (Ed.), The philosophy of John Dewey (2nd ed. pp. 227-242). New York: Tudor.
Peirce, C.S. (1965). Collected papers of Charles Sanders Peirce: Vol. V. Pragmatism and pragmaticism; Vol. VI: Scientific metaphysics (C. Hartshorne & P. Weiss, Eds.). Cambridge, MA: The Belknap Press of Harvard University Press.
Pennington, S. (1988). Healing yourself: Understanding how your mind can heal your body. Toronto: McGraw-Hill Ryerson.
Piatt, D.A. (1951). Dewey's logical theory. In P.A. Schilpp (Ed.), The philosophy of John Dewey (2nd ed., pp. 103-134). New York: Tudor.
Pilowsky, J. (1993). The courage to leave: An exploration of Spanish-speaking women victims of spousal abuse. Canadian Journal of Community Mental Health, 12, 15-29.
Polkinghorne, D.E. (1994). Reaction to Special Section on qualitative research in counseling process and outcome. Journal of Counseling Psychology, 41, 510-512.
Putnam, H. (1990). Realism with a human face (Ed. with an introduction by J. Conant). Cambridge, MA: Harvard University Press.
Rennie, D.L. (1994). Human science and counselling psychology: Closing the gap between research and practice. Counselling Psychology Quarterly, 7, 235-250.
Rennie, D.L. (1996). Reconciling realism and relativism: The root of qualitative research as methodical hermeneutics. Unpublished Manuscript, Department of Psychology, York University, Canada.
Rennie, D.L., Phillips, J.R., & Quartaro, G.K. (1988). Grounded theory: A promising approach to conceptualization in psychology? Canadian Psychology, 29, 39-150.
Rorty, R. (1979). Philosophy and the mirror of nature. Princeton: Princeton University Press.
Santayana, G. (1951). Dewey's naturalistic metaphysics. In P.A. Schilpp (Ed.), The philosophy of John Dewey (2nd ed., pp. 234-261). New York: Tudor.
Smith, J K. (1989, March). Alternative research paradigms and the problem of [p.118] criteria. Paper presented at the International Conference on Alternative Paradigms for Inquiry, San Francisco.
Smith, J.K., & Heshusius, L. (1986). Closing down the conversation: The end of the qualitative-quantitative debate among educational inquirers. Educational Researcher, 15, 4-12.
Smith, M.B. (1994). Selfhood at risk: Postmodern perils and the perils of postmodernism. American Psychologist, 49, 405-411.
Spiegelberg, H. (1960). The phenomenological movement: A historical introduction (Vols. 1 & 2). The Hague: Nijhoff.
Strauss, A. (1987). Qualitative analysis for social scientists. Cambridge, UK: Cambridge University Press.
Strauss, A., & Corbin, J. (1990). Basics of qualitative research: Grounded theory procedures and techniques. Newbury Park, CA: Sage.
Strauss, A., & Corbin, J. (1994). Grounded theory methodology: An overview. In N.K. Denzin & Y.S. Lincoln (Eds.), Handbook of qualitative research (pp. 273285). Thousand Oaks, CA: Sage.
Thayer, H.S. (1968). Meaning and action: A critical history of pragmatism. New York: Bobbs-Merrill.
White, M.G. (1964). The origins of Dewey's instrumentalism. New York: Octagon.
Wittgenstein, L. (1958). Philosophical investigations (G.E.M. Anscombe, Trans.). Oxford: Blackwell.
ACKNOWLEDGEMENTS. This article is derived from Objectivity and Relativity in Grounded Theory Analysis: Pragmatism Considered, presented at the annual meeting of the International Society for Psychotherapy Research, Vancouver, British Columbia, June 1995; and Psychological Research as Grounded Pragmatism, presented at the annual meeting of the American Psychological Association, New York City, August 1995. I wish to thank Kurt Danziger, Christopher Green, Gary Johnston, John McLeod, Sandra McNally, Shula Mor, Donald Polkinghorne, Milan Pomichalek and two anonymous reviewers for their comments.
DAVID L. RENNIE is a Professor in the Department of Psychology at York University. After a period of quasi-experimental research on counselor training, in recent years he has applied the grounded theory method to the study of the client's reported experience of psychotherapy. This transition has stimulated an interest in the philosophy of human science. Among his recent publications is S. Toukmanian and D. Rennie (Eds.), Psychotherapy Process Research: Paradigmatic and Narrative Approaches (Sage, 1992). ADDRESS: Department of Psychology, York University, North York, Ontario, Canada M3J 1P3. [email: drennie@yorku.ca] [p.119]
Theory & Psychology 1998 Vol. 8 (1): 101-119 ]
|