Introduction

К оглавлению
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 
51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 
68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 
85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 
102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 
119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 
136 137 138 139 140 141 142 143 144 145 146 

The purpose of this chapter is to provide an overview of some of the key problems that

researchers and policy makers using causal mapping techniques have wrestled with over

the ensuing years, both in order to illustrate the range of choices confronting the wouldbe

user of these techniques and to highlight the strengths and limitations of particular

approaches. Despite the fact that causal and other forms of cognitive mapping techniques

are generally more labor-intensive and time-consuming than other research

methods, in recent years the emerging field of managerial and organizational cognition

has developed dramatically (e.g., Eden & Spender, 1998; Hodgkinson & Thomas, 1997;

Meindl, Stubbart & Porac, 1994; Narayanan & Kemmerer, 2001; Porac & Thomas, 1989),

to the extent that its reach is now extending across virtually all of the major sub-fields

of management and organization studies, including information technology-related

applications (Nelson, Nadkarni, Narayanan & Ghods, 2000a; Swan, 1997). Researchers

have employed a rich variety of methods in an attempt to gain insights into actors’ belief

systems, ranging from the relatively simple process of having participants list basic

concepts (de Chernatony, Daniels & Johnson, 1993; Gripsrud & Gronhaug, 1985) to more

sophisticated procedures such as the development and multivariate analysis of questionnaire

items (Fombrun & Zajac, 1987) and repertory grid and related multidimensional

scaling and related clustering techniques (Daniels, de Chernatony & Johnson, 1995;

Daniels, Johnson & de Chernatony, 2002; Fournier, 1996; Ginsberg, 1989; Hodgkinson,

1997a; Hodgkinson, Padmore & Tomes, 1991; Hodgkinson, Tomes & Padmore, 1996;

Reger & Huff, 1993). Fortunately, a number of comprehensive reviews of the many diverse

methods for accessing thinking in organizational settings have been published elsewhere

(e.g., Fiol & Huff, 1992; Hodgkinson, 2001; Hodgkinson & Sparrow, 2002; Huff,

1990; Jenkins, 1998; Lant & Shapira, 2001; Mohammed, Klimoski & Rentsch, 2000; J.

Sparrow, 1998; Walsh, 1995).

In this chapter we shall confine our attention to a consideration of one particular class

of cognitive mapping techniques — causal mapping — that has risen in popularity in

research domains as diverse as strategic management (e.g., Fahey & Narayanan, 1989;

Hodgkinson, Bown, Maule, Glaister & Pearman, 1999; Hodgkinson & Maule, 2002;

Maule, Hodgkinson & Bown, 2003; Narayanan & Fahey, 1990), human resource management

(Budhwar, 2000; Budhwar & Sparrow, 2002), and technological innovation (Swan,

1995; Swan & Newell, 1998). In the words of Huff (1990, p.16):

“Causal maps allow the map maker to focus on action — for example, how the

respondent explains the current situation in terms of previous events, and what

changes he or she expects in the future.”

It is the direct links to action implicit within this approach that make it such a powerful

method, applicable across a wide range of contexts. However, as noted in this volume

by Narayanan (2005), causal mapping techniques have been much under-utilized within

the inter-related domains of information systems (IS) and information technology (IT).

This is highly surprising, given the obvious parallels with general systems theory and

the potential of these techniques to shed light on systems-designers’ and users’

understanding of a range of hardware and software capabilities and limitations (cf.,

Nelson et al., 2000a), thereby extending the repertoire of cognitive engineering tools and

techniques available for use in these domains (Schraagen, Chipman & Shalin, 2000;

Seamster, Redding & Kaempf, 1997). However, if this potential is to be realized, it is vital

that important methodological insights already gained in the context of other domains,

where causal mapping techniques have enjoyed widespread prominence, are brought to

bear in the context of IS and IT applications. Since Axelrod (1976) produced his classic

book that introduced causal mapping to the field of policy analysis, a number of

significant methodological issues have risen to the fore across a range of fields, which

in turn has stimulated much thinking and further advances.

In this chapter we map out some of the key methodological choices confronting the

would-be user of causal mapping techniques, drawing upon the wider body of research

that has been conducted using these techniques in other domains, over almost a 30-year

period, both in order to illuminate the nature of those choices and to accelerate progress

in these new, inter-related focal areas of application, by distilling the very valuable

lessons that have emerged from extensive prior usage in these other domains. In so doing,

our purpose is to accomplish three principal aims: (1) to illustrate the range of methodological

choices associated with causal mapping techniques; (2) to highlight the

strengths and limitations of the particular approaches identified; and (3) to offer some

general guidelines to aid the would-be user of these techniques. Our recommendations

are not intended to be prescriptive, but to assist potential users of causal mapping

techniques in making methodological choices that are appropriate in particular contexts

of application.

In Figure 1 we present a schematic overview of the principal stages involved in the causal

mapping process, as discussed in this chapter. Undoubtedly, this representation

oversimplifies the complex realities involved. (In practice, for example, the mapping

process is often an iterative one, with feedback sought from participants during or soon

after the construction and analysis stages.) Nevertheless, it serves as a useful framework

to guide those new to the process of causal mapping and provides a clear overview of

the organizing logic we have employed in structuring our chapter.

The chapter is organized in seven principal sections. Following this introduction, we alert

the reader to ongoing philosophical debates concerning the ontological status of causal

maps (also known as cause maps), outlining our own position in respect of these. In the

third section, we identify a number of issues concerning knowledge elicitation that

researchers need to address if they are to make well-informed mapping choices and we

highlight a number of strengths and limitations associated with particular approaches.

Next, we turn our attention to basic metrics for the analysis of individual cause maps. In

the fifth section we discuss issues associated with the aggregation and comparative

analysis of causal maps, while in section six we consider some measurement issues which

are fundamental to the entire mapping process. Finally, we draw together our key

recommendations and overall conclusions. In an accompanying appendix we provide a

brief overview of some of the available computer software systems for supporting users

throughout the various stages of the mapping process, from data collection/elicitation

to analysis and comparison.