On emergence and forcing in information systems grounded theory studies: the case of Strauss and Corbin

Abstract

Grounded theory method (GTM), which has been increasingly used in the information systems (IS) field, is a contested method. GTM has even been viewed as a family of methods by Antony Bryant and Kathy Charmaz in the respected ‘Handbook of Grounded Theory’. One debate that is of particular relevance is about the metaphor of ‘emergence’ and the most basic rule of GTM – that researchers should not force preconceived conceptualizations on data. This debate has its origins in a dispute between the two co-founders of grounded theory, Barney Glaser and Anselm Strauss. Glaser criticized Strauss for the introduction of the stage of axial coding and the use of one single coding paradigm. In his view, the paradigm is too rigid, forces data, hinders emergence, and leads to conceptual description instead of grounded theory. It is perhaps surprising then, that this debate has so far been conducted without any empirical evidence in support of the proposition that the Strauss and Corbin version of grounded theory might result in forcing. In this article, we analyse IS studies in top journals where Straussian grounded theory procedures – which have found most adherents in the IS discipline – are utilized. We provide detailed insights into the use and the impact of axial coding and the coding paradigm. We find that the researchers’ use of Straussian coding procedures in the IS field is on a level that is conscious and deliberative. We relate our findings to the broader debate about the contested nature of GTM, and find that axial coding and the coding paradigm are an exemplar that shows that GTM is an evolving method that is subject to idiosyncratic interpretations and flexible deployment. Our findings, we argue, are in line with the more recent development of constructivist grounded theory that holds that grounded theories are not discovered, they are constructed, based on conscious decisions and interpretive acts. We also put forward three propositions, and five guidelines, intended to assist IS researchers in constructing grounded theories using Straussian coding procedures, while adhering to the primacy of avoiding preconceptions.

Appendices

Appendix A

Details of literature search strategy

Table A1 provides an overview of the searched IS journals, as well as the database coverage (for the relevance of transparently documenting the literature search process compare, for instance, vom Brocke et al., 2009).

Table a1 Journals reviewed in the literature review

Our search strategy was as follows:

• 1 We started with a full text search using the search terms ‘Strauss’ and ‘Corbin’ in order to identify those articles citing the work of Anselm Strauss and Juliet Corbin. We justified this by arguing that any study that makes use of S & C coding procedures is most likely to reference their work. Table A3 provides an overview of the (methodological) publications by Anselm Strauss and Juliet Corbin that IS researchers cite in the investigated articles. We soon realized that some articles refer to S & C coding procedures without referencing Strauss and Corbin. Lim et al. (2005), for example, cite Orlikowski (1993) when it comes to the stage of axial coding. We then also included ‘axial coding’ as a search term. Our initial search was a logical expression of the following form: <‘Strauss’ or ‘Corbin’ or ‘axial’>. We then eliminated those articles where it was obvious that they do not relate to the use of S & C coding procedures, that is those articles

  • ° where other authors with the names Strauss or Corbin are cited, but not the aforementioned advocates of grounded theory, Barney Glaser and Anselm Strauss (e.g., a couple of articles cite Levi-Strauss), and that do not refer to axial coding (e.g., a couple of articles use notions such as ‘coaxial’ that are not related to the use of the paradigm); and

  • ° that only refer to work of Strauss in conjunction with Glaser (most notably Glaser and Strauss, 1967) and that do not refer to axial coding.This exercise produced a total of 115 articles (41 in EJIS, 23 in ISJ, 7 in ISR, 9 in JAIS, 12 in JMIS, and 21 in MISQ) that built the basis for our further analysis.

• 2 We then made a distinction between empirical and non-empirical studies, identifying 100 empirical studies and 15 non-empirical studies. While empirical studies were expected to help us answering the research questions by investigating the use of S & C coding procedures as well as the results of research, non-empirical studies were hoped to provide further insights into how S & C grounded theory is perceived and understood in the IS discipline. In some cases, we identified articles that are primarily conceptual, but use empirical data for illustration purposes. Of the 15 non-empirical articles, 9 are methodological, 3 are literature reviews, 1 a theoretical (conceptual) article, and 2 we classified as opinion pieces. For the classification scheme, we used the Publication Manual of the American Psychological Association (2010).

• 3 We then went through the 100 retrieved empirical articles in order to establish if the work of Anselm Strauss and/or Juliet Corbin is referenced in the context of the applied research method or if, at least, axial coding is mentioned. We also excluded those articles where the work of Strauss and Corbin is not cited in the context of the underlying methodology, but with regard to their actual (empirical) research; this is the case in three articles (Horton and Wood-Harper, 2006; Azad and King 2008; Bjørn and Ngwenyama, 2009). We further identified one article where it is stated that grounded theory is not used (Khoo and Robey, 2007). This left us with 96 empirical articles that have been published in the top six journals of the IS discipline that reference Strauss and Corbin in the context of the applied research method.

• 4 In a last step, we performed a more detailed analysis of the 96 retrieved empirical studies in order to identify in what manner S & C coding procedures are used. We read the articles, with a focus on the applied data analyses, and searched for mentioning of open, axial, and selective coding as well as paradigm items such as conditions, actions/interactions, strategies, and consequences, or the explicit notion of the coding paradigm.

Table A2 provides an overview of our search strategy, which focused on the identification of those studies that applied axial coding and the paradigm. While we took great care in applying this procedure, we cannot exclude the possibility that a few articles may have been missed.

Table a2 Overview of search strategy

Table A3

Table a3 Publications by Strauss and Corbin cited in the investigated studies in the context of the research method applied (references to Straussian grounded theory)

Appendix B

Table B1

Table b1 Articles that are explicit about the use of axial coding

Appendix C

Examples where it is possible that the S & C paradigm was applied, but there is not sufficient evidence

Work (2002), in a study of patterns of software quality management, writes about the use of axial coding: ‘The second technique is axial coding which develops the categories further. The specific features of each category, such as the conditions which cause the category to occur, are validated against the data’ (p. 64). Other paradigm items are not explicitly mentioned.

Maldonado (2010) writes: ‘Axial coding defines causal relationships and the intervening conditions that mitigate those relationships’ (p. 762). Again, no further paradigm items are mentioned.

Orlikowski (1993) says: ‘This iterative examination [axial coding] yielded a set of broad categories and associated concepts that describe the salient conditions, events, experiences, and consequences associated with the adoption and use of CASE tools in SCC’ (p. 314). With regard to this article, Urquhart et al. (2010) note: ‘It is not clear whether the Strauss and Corbin (1990) coding paradigm […] was used to assist the coding’ (p. 374).

Other examples where one may speculate about the use of the paradigm include Goulielmos (2004) as well as Baskerville and Pries-Heje (2004).

Appendix D

Analytical dimensions

This appendix provides an overview of the analytical dimensions that were used in order to analyse those studies that applied the S & C coding paradigm.

Table D1

Table d1 Analytical dimensions