Combining IS research methods: towards a pluralist methodology.

This paper puts forward arguments in favor of a pluralist approach to IS research. Rather than advocating a single paradigm, be it interpretive or positivist, or even a plurality of paradigms within the discipline as a whole, it suggests that research results will be richer and more reliable if different research methods, preferably from different (existing) paradigms, are routinely combined together. The paper is organized into three sections after the Introduction. In [section] 2, the main arguments for the desirability of multimethod research are put forward, while [section] 3 discusses its feasibility in theory and practice. [section] 4 outlines two frameworks that are helpful in designing mixed-method research studies. These are illustrated with a critical evaluation of three examples of empirical research.

(Critical Pluralism; IS Research Methods; Methodology; Multimethodology; Paradigm; Qualitative Research; Critical Realism)

1. Introduction

The question of which research methods are most appropriate for information systems research has been a focus of concern for some time. Orlikowski and Baroudi (1991) considered three broad research paradigms--positivist, interpretivist, and critical--and found that between 1983 and 1988 97% of IS research articles used a positivist framework. Since then there has been growing interest in, and commitment to, a range of nonpositivist (sometimes called post-empiricist) approaches, particularly based on interpretivism. This is evidenced by a shift in the editorial policy of some of the main IS journals, notably MISQ (Lee 1999, Walsham 1995), and a recent literature survey (Nandhakumar and Jones 1997) that found that 16% of papers used a broadly interpretive methodology. For general coverage of a variety of methods and paradigms see Nissen et al. (1991), Galliers (1992), and Goles and Hirschheim (1999). For recent papers in favor of interpretivism see Avison and Myers (1995), Harvey and Myers (1995), and Myers (1994); for a survey of the critical perspective in IS see Ngwenyama and Lee (1997).

In part this diversity of approach is because information systems, broadly defined, draws on and provides a nexus for many diverse research fields and disciplines. IS is much more than simply the development of computer-based business systems--electronic and information technology is now so fundamental within society that IS as a discipline must concern itself with the general evolution of human communication (Mingers and Stowell 1997). Thus, it has to draw upon a very wide range of disciplines--technology, psychology, economics, sociology, mathematics, linguistics, semiotics--that encompass very different research traditions. This puts IS in a position similar to other management areas such as organizational studies, which are also characterized by a plurality of research paradigms, each with particular research methods. Consideration then turns to the possible relations between these paradigms. The traditional view, particularly since Burrell and Morgan's (1979) seminal analysis of paradigms in organization theory, has been one of isolationism in which the paradigms are seen as essentially based on mutually exclusive and contradictory assumptions, and individual researchers do, or should, follow a single paradigm. This principle is sometimes justified in terms of the superiority of the (usually positivist) paradigm but, more recently, Benbasat and Weber (1996) have argued in terms of the need for uniformity within the IS discipline as a whole:

 
   Our own view is that we need both a paradigm (one or more) and diversity in 
   the IS discipline. A paradigm will serve to provide coherence to the IS 
   discipline and to characterize the phenomena that make it different from 
   other disciplines. In short, it is needed to articulate the core of the 
   discipline (p. 397). 

Without this, they fear, the discipline will shatter or be taken over by a more established discipline.