Abstract
This chapter provides a brief introduction to the grounded theory (GT) approach to research, discussing how it has been used in information systems (IS) research, and how GT studies may be conducted to provide a significant theoretical contribution to the management information systems (MIS) field. The subject is of particular interest at a time when GT attracts frequent criticism for a lack of rigor. This chapter deals with what makes for a rigorous contribution to “grounded” theory in MIS. It addresses developments and controversies in the generation of grounded theories, examining the use of GT as a coding method vs. the use of GT as a method for generating theory. The discussion focuses mainly on the constructivist/interpretive perspective adopted in most qualitative data studies, as this is the way in which GT has been used most often in MIS. The chapter concludes with a roadmap for the use of GT in MIS research and a discussion of the contribution made by GT studies in MIS.
TopIntroduction
The Grounded Theory (GT) research method has grown more popular in recent years. This is partly in response to an increasing awareness of the limitations of applying a priori, deductive theories to human transactions embedded in a social context, and partly in response to the immaturity of Management Information Systems (MIS) as a discipline. The GT approach is used to generate a substantive theory – a theory that is grounded in specific mechanisms, contexts or environments. This fits well with the need to produce in-depth empirical studies that develop a dynamic body of theory that evolves with the MIS field itself. The construction of a grounded theory relies on a systematic analysis of qualitative data, to theorize about “what is it that is happening here?” The result is a theory that is grounded in empirical evidence, rather than developed from existing conceptual frameworks. The GT approach may be used to analyze qualitative data to produce quantitative data that are analyzed statistically, or it may employ a qualitative, interpretive data analysis throughout. The latter approach is the most frequently encountered in MIS and so this chapter will focus mainly on these studies.
Grounded theories are situated, not only in “the data,” but also in the context in which data was collected. They may be considered idiographic theories, that are “concerned with the individual [case], pertaining to or descriptive of single or unique facts and processes” (Dey, 1999, pg. 217). Quality criteria for idiographic theories of action emphasize transferability or adaptation to different contexts, rather than the generalizability concerns that are applied to nomothetic or formal theories (S. . Gasson, 2003; Lincoln & Guba, 2000). The GT approach may be used over time to generate formal (nomothetic) theories, that are more generalizable as they are derived from multiple studies and contexts. This requires a substantial amount of time and relies on researchers who are capable of reflexive theoretical abstraction. The majority of GT studies contribute idiographic theories that provide deep insights into the research problem for a limited number of situations or contexts.
The aim of this chapter is to provide an overview of the GT method, to address controversies and issues surrounding its use, and to provide some guidance on how it may be used to contribute meaningfully to MIS research. The chapter is organized as follows. First, some background is provided on the GT approach to research and the GT “method.” Secondly, I discuss controversies and developments in both the approach to, and the methods for Grounded Theory generation. Thirdly, the contribution of the GT approach to MIS is demonstrated by means of illustrative studies. The fourth section presents a roadmap for GT research in MIS, examining the unique challenges that our field presents to GT researchers, discussing the constraints presented by seed categories and a priori theoretical models, ethics, boundaries, and scope in MIS research, and issues of generalizability. The chapter concludes with a brief discussion of the contribution of GT to the MIS field.
Key Terms in this Chapter
Theoretical Saturation: The point at which analysis of additional data through constant comparison across data samples, cases, or situations provides no new insights into the substantive theory of action generated from the data.
Substantive Theory: A theoretical model that provides a “working theory” of action for a specific context. A substantive theory is considered transferable, rather than generalizable, in the sense that elements of the context can be transferred to contexts of action with similar characteristics to the context under study (for example, studies of small-group IS design in US management consultancy companies). This contrasts with Formal Theory, which is based upon validated, generalizable conclusions across multiple studies that represent the research population as a whole, or upon deductive logic that uses validated empirical theories as its basic axioms.
Post-Positivist: Based on the belief that most knowledge is conjectural, this research paradigm emphasizes deductive logic, or warrants, in supporting theory generation. .Post-positivism admits reported experience (for example, surveys), sociological or psychological experiments (where the data must be inferred from other phenomena) and observed human behavior as data. Because of the wider criteria for data acceptability than is the case for positivism, post-positivism is often used to describe an approach to research where large amounts of qualitative data are categorized to produce quantitative data to be analyzed using statistical methods.
Social Constructivist: Where the individual views the external world as represented by a set of names, concepts, and labels that are used to structure reality. We understand the world by socially-situated processes of framing and filtering that employ consensus concepts, names, and labels. Very few people ever adopt these extreme positions in totality. For example, researchers who describe their position as social constructivist would admit to some parts of physical reality as having an existence independent of their own. Similarly, researchers who describe their position as realist (the epistemological term “positivist” is more often employed), admit to some elements of their understanding of the external (to them) world as resulting from subjective perceptions.
Realist: Where the individual views the external world as having an existence independent of their own.
Ontology: Related to how we view the nature of the external world. The two extreme positions are:
Idiographic: An approach to research knowledge that is concerned with the study of individual or specific cases, pertaining to contingent and often subjectively-perceived phenomena that relate to an identifiable context. This may be contrasted with the nomothetic approach most usually employed in positivist research studies which are seen as representing a population of individuals and focus on the variables and behaviors that characterize a generalizable set of contexts.
Epistemology: Our beliefs about the nature of knowledge and our relationship to the “real world” (i.e. how we know reality). The most common positions are:
Positivist: Observable, measurable experiences can contribute to knowledge. Theories result from the application of scientific methods of analysis to data, resulting in a research approach that emphasizes quantitative data collection and statistical analysis methods to ensure the validity of findings.
Systemic Inquiry: This is a philosophy of research and action that relates elements of the situation together, viewing a problem-situation as an interrelated set of cause-effect relationships or phenomena. This philosophy was proposed by Churchman (1979) and operationalized in the Soft Systems approach advocated by Checkland (Checkland and Holwell, 1998).
Interpretivist: Reported experience, phenomena, and observations are seen as social constructions: filtered through the interpretations that result from the individual’s prior experience. This approach attempts to account for interpretations by research subjects and the researcher, in suggesting findings from a research analysis. The resulting theories are seen as contestable and context-specific, providing explanatory power, rich descriptions, and in-depth understanding of how and why to act, rather than providing prescriptive rules for action.
Grounded theory: A theory that is generated from patterns in, and relationships between, elements of the data collected, rather than based upon extant theories of action. The Grounded Theory method is based upon Glaser and Strauss (1967) and requires two key process elements, 1. Data are categorized (or coded) according to an emergent set of categories that define key elements of the situation. These are related together by means of code integration to derive families of codes, or code splitting to derive sub-categories. A substantive theory is generated when the researcher can define core categories in the data and important patterns of relationships between categories, that apply across data samples. These patterns are made explicit through the generation of theoretical memos as the analysis proceeds,2. A constant comparison method is employed to analyze which elements of the emerging theory apply across multiple data samples, at which point the emerging theory may change. Constant comparison ends when theoretical saturation is reached (see below).