Distinctions Among Different Types of Generalizing in Information Systems Research

It is incorrect and even harmful that many information systems researchers typically criticize their own intensive (qualitative, interpretive, critical, and case) research as lacking “generalizability.” We untangle and distinguish the numerous concepts now confounded in the single term “generalizability,” which are generality, generalization, generalize, general, and generalizing. These clarified terms allow us to identify four distinct forms of generalizing (everyday inductive generalizing, everyday deductive generalizing, scientific inductive generalizing, and scientific deductive generalizing), each of which we illustrate with an information-systems-related example. With these clarified terms, we show how information systems researchers who perform intensive research can properly lay claim to generality for their research. Distinctions among Different Types of Generalizing in Information Systems Research For researchers in information systems (IS), the concept of “generalizability” has been developing and maturing with the growing acceptance of intensive research approaches (e.g., qualitative, interpretive, critical, and case research). Often misconstrued to be a property of statistically based research alone, generalizability has been gaining recognition as an achievable ideal in intensive research as well (Lee, 1989). Still, it remains common for intensive researchers to flagellate themselves in their published articles’ “discussion” sections for the lack of generalizability of their findings. There they typically blame this supposed failure on their having examined “only” a single case, or “only” three technologies, or “only” two organizations, or “only” one point in time, and so forth. We believe that such self-flagellation is not necessary (i.e., the completed intensive research can indeed claim “generalizability” if it is properly performed and presented). Such unwarranted self-criticism can even be harmful (i.e., it incorrectly elevates large-sample statistical research above all other forms of scholarly inquiry). The purpose of this study is to clarify the different processes of generalizing so that scientific researchers in IS can better achieve and securely claim “generalizability” (or, as we will rename it, “generality”) in their research. We will illustrate these processes with some examples of how authors of specific published articles in IS research can better describe the “generalizability” of their studies. The first section of the paper after this introduction will identify and distinguish different terms now confounded in the single term, “generalizability.” The terms that we will unconfound and extract from “generalizability” are generality, generalization, generalize, general, generalizing, and even (after we define it) generalizability itself. In the section after that, we will define four types of generalizing and classify them according to the dimensions of “reasoning process” (inductive vs. deductive generalizing) and “context” (inquiry in everyday life vs. inquiry in scientific research). The same section will illustrate the four types with 1 We take this characterization of “intensive research” from M. Lynne Markus and Allen S. Lee’s call for papers for a special issue of MIS Quarterly on intensive research. They, in turn, took the term “intensive research” from Karl Weick and used it to refer to the diversity of forms of empirical information-systems research falling outside of the quantitative and positivist genre, including qualitative positivist (and non-positivist) research, interpretive research, critical social theory research, and case study research.