Comparing managers' mental models of competition: why self-report measures of belief similarity won't do

Abstract

In their study designed to investigate the relative impact of 'task and institutional influences on managers' mental models of competition,' Daniels et al. (2002) elicited cognitive maps using two complementary ideographic mapping procedures: a card-sort technique and a variant of the repertory grid. Given that the resulting individual maps were each based on differing organizations and attributes, Daniels and his colleagues assessed belief similarity -- their key dependent variable -- by asking their participants to rate the overall similarity of the various maps so elicited to their own mental models which prevailed at the time the comparative rating exercise was subsequently performed, some three to six months later. Drawing on research into the cognitive processes underpinning similarity judgements, I argue that this approach to the comparison of cognitive maps suffers from a number of severe limitations which are likely to bias the results in favour of the research hypotheses under test, thus leading to increased type I errors. Alternative procedures for eliciting and comparing individuals' mental representations of competition are briefly considered.

Descriptors: institutional theory, task environment, institutional environment, mental models, cognition, cognitive mapping

Introduction

In recent years, there has been a rapid growth in theory and research that has sought to develop understanding of the nature and significance of actors' mental representations of business competition (for reviews, see Hodgkinson 1997a, 2001a, 2001b; Lant and Phelps 1999). Two major trends are discernible within the literature. One group of scholars (e.g. Lant and Baum 1995; Porac et al. 1989,1995; Porac and Thomas 1990, 1994; Peteraf and Shanley 1997; Reger and Huff 1993) have focused their attention primarily on the inter-organizational level of analysis, arguing that through related processes of social construction, inter-organizational learning, social identification and institutionalization, actors' mental representations of competition become highly unified over time, leading to the emergence of stable groupings of firms (see also, Greve 1998). A second group of scholars (e.g. Calori et al. 1992, 1994; Daniels et al. 1994a; Hodgkinson and Johnson 1994; Johnson et al. 1998) have focused their attention pr imarily on the search for significant intra- and inter-organizational individual differences in mental representations of competitive structures, arguing that such differences might play a central role in a variety of non-trivial processes associated with strategy formulation and implementation (cf. Lant and Phelps 1999).

As one of the first attempts in the related fields of organization studies and strategic management to explore the relative contributions of 'task and institutional influences on managers' mental models of competition', the study reported by Daniels et al (2002) has broken new ground. Simultaneously drawing on the two main strands of theory and research briefly summarized above, Daniels and his colleagues have sought to identify systematically the extent to which task and institutional factors each contribute to the convergence and divergence of managers' mental models. The study is well-grounded theoretically, the industry in which they have chosen to conduct their investigation has been carefully selected, and the nature of the participating organizations and individual participants are entirely appropriate for the hypotheses under test. The way in which Daniels and his colleagues elicited and compared their participants' mental representations of competition, however, is an entirely different matter, which is worthy of extended debate.

According to Daniels et al. (2002), nomothetic cognitive mapping procedures might lead artefactually to greater levels of convergence than free response methods, by virtue of the standardized stimuli and attributes employed in the elicitation process. In an attempt to overcome this potential problem, Daniels and his colleagues used an ideographic, card-sort procedure in conjunction with a variant of the repertory grid technique. Previous studies (Daniels et al. 1994a, 1995) have demonstrated that these procedures yield maps with a number of convergent properties, thus apparently strengthening the researchers' claims concerning the reliability and validity of their approach to cognitive mapping. Moreover, many other studies into managerial and organizational cognition in general, and mental models of competition in particular, have similarly employed ideographic mapping techniques. The primary advantages of these techniques are that they neither constrain participants to focus on 'competitors or attributes of competition predetermined by researchers or research on other managers', and that, unlike nomothetic procedures, they do not 'force participants to consider organizations and bases of competition that they may not ordinarily consider relevant' (Daniels et al. 2002: 40). Whilst these features are undoubtedly attractive, as Daniels et al. (2002: 56, note 4) readily acknowledge: 'ideographic methods can artefactually increase the divergence amongst cognitive maps through demand characteristics, which emphasize surface-level triviality in the maps during interviews'. They also correctly observe that, in view of the fact that several of their own hypotheses imply differences amongst mental models, 'this is problematic' (ibid.).