A Methodology for Constructing Collective Causal Maps*

Decision Sciences, May 2006 by Scavarda, Annibal José, Bouzdine-Chameeva, Tatiana, Goldstein, Susan Meyer, Hays, Julie M, Hill, Arthur V

ABSTRACT

This article develops a new approach for constructing causal maps called the Collective Causal Mapping Methodology (CCMM). This methodology collects information asynchronously from a group of dispersed and diverse subject-matter experts via Web technologies. Through three rounds of data collection, analysis, mapping, and interpretation, CCMM constructs a parsimonious collective causal map. The article illustrates the CCMM by constructing a causal map as a teaching tool for the field of operations management. Causal maps are an essential tool for managers who seek to improve complex systems in the areas of quality, strategy, and information systems. These causal maps are known by many names, including Ishikawa (fishbone) diagrams, cause-and-effect diagrams, impact wheels, issue trees, strategy maps, and risk-assessment mapping tools. Causal maps can be used by managers to focus attention on the root causes of a problem, find critical control points, guide risk management and risk mitigation efforts, formulate and communicate strategy, and teach the fundamental causal relationships in a complex system. Only two basic methods for creating causal maps are available to managers today-brainstorming and interviews. However, these methods are limited, particularly when the subject-matter experts cannot easily meet in the same place at the same time. Managers working with complex systems across large, geographically dispersed organizations can employ the CCMM presented here to efficiently and effectively construct causal maps to facilitate improving their systems.

Subject Areas: Causal Maps, Knowledge Networks, Process Improvement, Quality Management, and Systems.

INTRODUCTION

Causal maps are an essential tool for managers who seek to improve complex systems in the areas of quality (Pande & Holpp, 2001; Evans, 2005), information systems (Nelson, Nadkarni, Narayanan, & Ghods, 2000), and strategy (Kaplan & Norton, 2004). These causal maps are known by many names, including Ishikawa (fishbone) diagrams (Enarsson, 1998), cause-and-effect diagrams (Evans, 2005), impact wheels (Sorach, 2006), issue trees (Universität St. Gallen, 2005), strategy maps (Kaplan & Norton, 2004), current reality trees (Goldratt, 1994), and risk-assessment mapping tools (Hodgkinson, Tomes, & Padmore, 1996).

Managers can apply causal maps in at least five ways. First, causal maps can be used as a diagnostic tool to focus attention on the root causes of a problem (Evans, 2005). Second, they can be used to identify the critical control points (the levers) for a system (Kaplan & Norton, 2004). Third, they can be used to guide risk management and risk mitigation efforts (Card, 1998). Fourth, they can assist managers in both formulating and communicating strategy (Kaplan & Norton, 2004). Fifth, they can be a powerful tool to help managers teach the fundamental causal relationships in a complex system.

In the social sciences, a causal map is generally considered to be a particular type of cognitive map, which is an individual's mental model of the relationships (causal or otherwise) among the elements of a system. Typically, causal maps are drawn with nodes representing concepts, ideas, or areas. The nodes are linked with unidirectional arcs that represent beliefs about the causal relationships among these nodes. The arcs may indicate only the presence or absence of a relationship or may indicate the strength of the relationship between two nodes. Synthesizing causal maps from a number of experts results in a "collective causal map."

The only two methodologies currently in practice for capturing cognitive data for a causal map are brainstorming and interviews. While these methods are effective in many contexts, they make it difficult to collect information from a large group of dispersed and diverse experts, especially when multiple rounds of data collection are needed. These methods usually rely on a small group of homogenous experts and can be subject to groupthink, which is defined as biases introduced by group dynamics (Nadkarni & Nah, 2003), and superstitious learning, which is defined as incorrect attribution of causality (Levitt & March, 1988).

Surowiecki's (2005) book, The Wisdom of Crowds, argues that groups are remarkably intelligent and are often smarter than the smartest people in them. He further argues that wise crowds need to have (i) diversity of opinion, (ii) independence of members from one another, (iii) decentralization, and (iv) a good method for aggregating opinions. He contends that, if these four conditions are met, the collective intelligence of the group will produce better outcomes than a small group of experts.

This article develops a new methodology for creating causal maps called the Collective Causal Mapping Methodology (CCMM) that is consistent with Surowiechi's theory. This methodology is a formal, inductive, three-round, Delphilike, Web-based asynchronous data-based approach for collecting and analyzing cognitive data from geographically dispersed and diverse experts. CCMM is particularly valuable when the problem domain includes many subdisciplines holding different views and when the study team does not have the resources to interview many experts. CCMM follows the Six Sigma philosophy of "data-based decision-making" (Pande & Holpp, 2001) being better than either brainstorming or interviews and, therefore, less likely to be impacted by "superstitious learning" (Levitt & March, 1988).