Decision Explorer, Version 3.0


Decision Explorer is cognitive mapping software for use in qualitative data analysis produced by Banxia Software. It is a software tool to help the user illustrate and elucidate relationships between ideas and perspectives. Decision Explorer can be used to analyze interview data, to aid in gathering data, to build qualitative models to aid in dynamics modeling, and to show how values are structured within a particular context. In this review, I will focus primarily on the use of Decision Explorer to examine interview data to structure a decision process. According to the Decision Explorer information, this is the "classic use" of the software.

Intended audience

Decision Explorer seems to be aimed a fairly broad audience. The software can be used in a business context to structure problems for analysis or in project management. The program is also useful for academic researchers as a tool for eliciting information in interviews or for structuring interview data that has been gathered. In terms of prior knowledge, Decision Explorer is best suited for someone who is familiar with the process of cognitive mapping. In addition, Decision Explorer seems to have well-organized users network. They publish newsletters a few times a year and provide a useful bibliography on their web-site.

Range of Functions

Decision Explorer offers a number of tools to draw cognitive maps and to analyze their content. Basic map drawing consists of entering concepts to the model and linking those concepts. Concepts are linked using arrows that can represent a causal, temporal, negative, or user defined relationship. For simple problems, simply drawing the map can provide a fair amount of insight to the problem. The concepts in the map can be coded through the use of different colors and font styles in order to represent different components of the decision process. For example, the user can distinguish policy constraints from economic constraints through the use of styles in map drawing. Figure 1 shows a model that I constructed for the decision of whether or not to centralize Californiaís vehicle inspection and maintenance program Smog Check. This model is based on data gathered from an exploratory interviews.

Figure 1: Decision Explorer model of the decision of whether or not to centralize Californiaís Smog Check Program

To go beyond drawing and linking the concepts, Decision Explorer provides a large number of analysis tools. All analysis tools can be invoked through menus on the top of the screen or through a command line. Some of the analysis tools that are available are best suited to large problems (defined by having a large number of concepts). The Decision Explorer Userís Guide states that analysis tools are best suited for models that are composed of greater than 40 concepts. Some of the simplest analyses in Decision Explorer that are good for practically any size model are the analysis of heads, tails, and consequences. These analyses allow a useful way for the user to review what the model is actually saying (especially useful for a novice user). The commands list heads (LH) and list tails (LT) display all concepts that have no arrows leading out of them (heads) and those that have no concepts leading into them (tails). The command "Explanations" provides route leading to a given concept in the model. For example, Figure 2 shows the output of the command "Consequences" for the result "Reduction of air pollution from mobile sources" for the model I constructed on the decision of whether or not to centralize the Smog Check program (shown in Figure 1).

+27 centralize smog check testing

may lead to

+39 insure long-lasting and effective repairs

which can lead to

+40 reduction of air pollution from mobile sources

+27 centralize smog check testing

may lead to

+29 hire centralized contractor

which can lead to

+30 control over test monitoring

which can lead to

+39 insure long-lasting and effective repairs

which can lead to

+40 reduction of air pollution from mobile sources

2 routes

Figure 2: Output of the command "Consequences" in for the model shown in Figure 1. This shows the sequence leading to the result "reduction in air pollution from mobile sources" for the decision to centralize the Smog Check program. The numbers in the output refer to concept numbers in the model.

When models become quite large, Decision Explorer has several useful tools available to explore portions of the models. For instance, in the case of very large models (greater than 150 concepts) a cluster analysis can be used to create more manageable sets of concepts. Cluster analysis attempts to define mutually exclusive sets within the model based upon the userís inputs (the results must be treated with caution in order to insure appropriate boundaries have been drawn). Another tool that is useful for any size model is the creation of sets. Sets are user-defined groups of concepts within the model that share a common characteristic. For example, for the model in Figure 1, I defined a set of concepts called "consumer" that contained all concepts that relate to consumers in the Smog Check program. Similar to using styles in map drawing, the creation of sets allows the user to define a common feature that unites a group of concepts. The benefit of creating sets is that once you have created and named sets, you can use more Decision Explorer analysis tools to explore relationships between sets.

In addition to allowing the user to simplify and streamline data, Decision Explorer has several tools to aid in showing relationships between concepts or sets in the model. The user can invoke the command "potent" in order to determine the most influential concepts in the set. Similarly, the command "domain" determines the "density" of concepts around a certain specified concept. Each of these commands allows the modeler to see which concepts are most influential in the model that has been created. Decision Explorer enables the user to create hierarchical sets based upon one set, to reduce the model to a specific set, or to explore concepts at the boundaries of sets. These types of analyses allow the user to focus on specific portions of the model and to relate sets of concepts to one another.

Another use of Decision Explorer analysis is in the interview and data gathering processes. Each of these analyses could aid in focusing further interviews and can provide useful checks on the validity of the model that is constructed. The literature on Decision Explorer encourages users to share their model output with informants. For instance, if a model is constructed based on an interview the resulting model can be constructed and shown to the informant for verification. This review of the model can provide another channel for communication between the interviewer and the informant and can provide a tool for gathering further information.

Utility to an ERG Researcher

I see two primary uses for Decision Explorer by researchers in ERG. The first is as an organizational tool and the second as an aid in interviewing. The latter is more applicable for researchers who are primarily focusing on the decision making process in their research and not as much on the context that goes into shaping the decision criteria. In the first use, Decision Explorer is a good tool for organizing information and displaying connections between data. The use of cognitive maps forces the researcher to think systematically about the problem he or she is studying and to analyze data in a way that emphasizes connections and relationships. I think that this aspect of Decision Explorer could be useful for any researcher who is looking for a way to organize data collection and analysis.

The more advanced analysis tools in Decision Explorer are better suited to researchers who are primarily interested in exploring the intricate details of a decision process. Yet, even for these researchers, a lot of data is required for the analysis tools to be useful. In this case, Decision Explorer could be useful for both eliciting information from informants and in devising an overall research plan. As was discussed previously, models that are constructed in Decision Explorer can be useful tools in the interview process for a researcher who is focusing on decision making. The model can be used to verify information and relationships and to work with an informant to identify and fill data gaps. In addition, a Decision Explorer model can be used by the researcher, alone, to identify holes in the data, where different informants disagree, and what information is needed to complete the model. In this capacity, Decision Explorer can be used as an organizational and data gathering tool.

Overall Evaluation

As far as software goes, Decision Explorer is fairly easy to use. Drawing the concepts and links in the cognitive map is fairly straightforward. I think that the primary limitation of Decision Explorer is that the user needs to be fairly familiar with the nuances associated with drawing cognitive maps in order to get the most out of the program. Nonetheless, the on-line help has a very useful overview of cognitive mapping and the Decision Explorer web-site provides pages of references to papers that use cognitive mapping and/or Decision Explorer. Overall, Banxia does an excellent job of providing the background material for the software.

Despite all of the background provided with the software, I feel that the information may come a little late if data has already been collected. I think that Decision Explorer would be most useful if it is selected as a data analysis tool before data collection begins. In this case, data could be collected so that it is most useful for cognitive mapping. Otherwise, it is quite time-consuming to go through long transcripts of interviews and extract the information that is needed to construct the model. Nonetheless, this strategy is most appropriate for researchers who are primarily interested in the intricacies of the decision making process and not the broader context of the decision making (e.g. cultural and social factors).

In terms of usability, I had a few gripes with the navigability of the program. The documentation provides an overview of the analysis functions, but is not entirely clear on where they are useful. It was not completely clear what advantage one analysis had over the other, and in some cases how they were even different. Another problem that frustrated me to no end was that once I had drawn my map, I found it very difficult to restore it to its original state after analyses had been done. Each of these difficulties made it difficult for the novice cognitive mapper/Decision Explorer user (a.k.a. me) to explore the program and interpret the results. I found the simpler analyses much more transparent than some of the more complex tools. Aside from these difficulties, in most cases where I had problems or confusion, I found the help file in the program and the Userís Guide that is available with the demonstration version to be quite useful.

Overall, I think that Decision Explorer is a good program for a researcher who is interested in studying a the details of a decision making process through the use of cognitive mapping. Yet, even in these cases, I think that it is important to consider factors that are not part of the model such as cultural or social constraints and other forms of information that are not easily incorporated into the cognitive map. For researchers who are interested in this broader context of decision making or who are looking for an organizational tool, I think that the effort that would go into learning to use Decision Explorer could probably be used more productively elsewhere.

All this said, I think that there are two major limitations in my evaluation of Decision Explorer. The first is that I did not have a large data set to work with. I think that the analytic abilities of the program are most useful in cases where there are a large number of concepts and links to work with. The second limitation is that I am not intimately familiar with cognitive mapping. For a researcher who knows the theory behind cognitive mapping well, I think that the benefits of Decision Explorer would be more transparent.



Banxia Software Limited. 1998. Decision Explorer Newsletter. Issue 1, Spring/Summer 1998.

Banxia Software Limited. 1998. Decision Explorer Newsletter. Issue 2, Winter 1998/1999.

Banxia Software Limited. Decision Explorer Userís Guide Version 3.

Brightman, Jennifer R.. 1998. An Introduction to Decision Explorer: Decision Explorer Workbook 1. Banxia Software, Ltd..