Causal Mapping

As is also evident elsewhere in this book, not only are there a number of different ways

that causal mapping can support Information Systems, but there are different forms of

Causal Mapping. The particular version considered in this chapter is based upon the

work of a cognitive psychologist — George Kelly (1955) — whose propositions about

how individuals make sense of their world resulted in a powerful body of theory known

as personal construct theory (PCT). From this body of theory Kelly developed an

instrument: Repertory Grids (Fransella & Bannister, 1977; Bonarius, Holland & Rosenburg,

1981) which has been used in IS research (Hunter & Beck, 2000; Tan & Hunter, 2002). In

addition, and, more pertinently for this chapter, cognitive mapping was developed to

reflect more depth and greater appreciation of individuality than that which was offered

by repertory grids (Brown, 1992; Eden, 1988). However, as many organizational tasks

require the involvement of a number of participants, a range of different forms of group

mapping have been developed (Ackermann & Eden, 2001; Eden & Ackermann, 1998) to

extend the use of causal mapping that is founded in personal construct theory. These

different forms include:

The “Oval Mapping Technique” — a manual, rather than computer-assisted, process.

Individuals are provided with oval shaped cards (about 11x19cms, or large rectangular

post-its) and identical pens (to help increase anonymity). They are asked to write down

any concerns, aspirations, issues or assumptions that come to mind. Each contribution

is written on a separate oval card. These ovals are posted up on a flipchart paper-covered

wall, enabling others to read and “piggyback” off them. During this process of generating

a scatter picture of the important aspects of the situation under consideration, a facilitator

works at clustering the material into themes so as to manage the large amount of material.

It is not untypical to have 200+ contributions surfaced from a group of eight by the end

of a two-hour session. Once the rate of contributions has slowed down or stopped, the

group, with the help of the facilitator, works through each cluster exploring how the

content of the different statements causally influence one another, both within and

across clusters. This process inevitably surfaces further views as different chains of

causality are presented and captured. The result is a structured causal map. The map

represents not only the different themes/clusters but also their degree of interaction with

one another. The process invariably enables the participants to move from an often very

divergent view of the situation to a more convergent one.

The second form of group mapping involves the use of a mapping software package —

Decision Explorer1 — to capture and structure the material. Instead of having participants

write down their contributions, the facilitator captures the contributions as they

are stated in facilitated discussion, thus in real time. The facilitator regularly checks with

participants that views have been captured correctly, and explores their relationships

with other already captured statements. This enables participants to concentrate on the

discussion — using the developing map as an aide mémoire. As a result of working

electronically the group can interactively “play” with the captured material exploring, for

example, the consequences of different options, examining possible alternatives, and

agreeing objectives/goals (end points of the causal chains). In doing so they are always

aware of the causal ramifications of their developing agreements. In addition, rapid

searches can be carried out for statements focusing on a particular topic, as well as a range

of analyses used to enable exploration of the structure of the map. Analysis results, such

as the identification of “central” statements that influence and are influenced by many

others, of “potent options” that can affect many goals, and of significant outcomes or

goals, subsequently can be colour coded and categorised. Finally by having a number

of views available (similar to spreadsheet packages where there can be a number of

sheets) of different aspects, managing the complexity of the large body of material

surfaced becomes easier.

The third and final form has participants provided with laptops connected together

through a local area network, and connected through Decision Explorer to a public

screen — the combined system is run through the software Group Explorer2. This allows

participants to directly enter their contributions (both statements and relationships) into

the map developing on the public screen as soon as they think of them. This allows for

total anonymity and reduces the pressure on the facilitator to capture the material. In

addition, this mode of working provides other useful features such as the ability for each

participant to rate the importance of statements on a scale or to prioritise them. For

example, through using the “preferencing system” participants are able to demonstrate

anonymously which of the options they will support and which they will not, providing

a reality check on the proposed way forward. Alternatively, participants can use the

rating tool, which allows them to explore which of the various options might provide best

support for an objective and whether there is any consensus about this.

Each of these has been applied to a range of different decision making areas (see

Ackermann & Eden, 2004) including problem structuring (Eden & Ackermann, 2001),

strategy development (Eden & Ackermann, 1998) and the modelling of disruptions and

delays occurring on large engineering projects (Ackermann et al., 1997).

In each of these processes the technique of causal mapping provides the means of

developing a graphical representation of an individual’s or group’s perception of issues

by building up chains of argumentation. Therefore, statements/nodes (facts, assertions,

options, issues, goals, etc.) are captured along with their relationships — where the arrow

(relationship) implies causality. In Figure 1, the individual is discussing the use (or not)

of a financial information system. Node 13 and 7 are both assertions resulting in the

perceived consequences of 9 and 6. This is a small causal map — frequently individual

maps comprise 80+ statements and group maps in excess of 400 statements.

Furthermore, the analysis of the structure of the causal map can reveal patterns. For

example, busy points typically suggest key issues, superordinate statements (those at

the top of the chains of arguments) imply values or goals, and feedback loops imply

potential dynamic behaviour. In Figure 1, node 17 has a considerable amount of material

supporting it — some of which is shown in detail, e.g., nodes 15, 18, 20 and some of which

is currently hidden (nodes 27, 25, 23), suggesting that it is likely to be a key issue. Node

8 is displayed as a goal — this was something that the individual felt was good in its own

right. Finally on the left of the figure is a feedback loop (comprising nodes 17, 19 & 20)

— in this case operating as a vicious cycle (a self-sustaining negative situation). In this

way the user is able to examine both the whole (in terms of emergent properties of its

structure) and the detail to help develop a fuller understanding of the interaction of

different considerations, and so make a more informed decision.

The map enables a better understanding to be developed as statements are explored

alongside their context. Adhering to the formal coding guidelines (see Appendix A) is

necessary for the analysis of emergent properties of the map to be reliable, but it also

provides a powerful aid to group thinking. For example, the guideline asking that each

Figure 1. An example map

statement be worded in an action-orientated form encourages those developing the maps

to be clearer about what might be done and why that would be expressed by an assertion.

Attending to the direction of causality between two statements prompts consideration

of what is the means and what is the desired end or outcome of two action-oriented

statements, often revealing underlying values.

Although mapping can help individuals better understand their world, when considering

Information Systems design it is mapping’s ability to support group negotiation that is

of most interest here. Capturing the views of those participating, along with the full

context of their views as chains of argument means individuals are better able to

determine what they are concerned about and what they want regarding the system

within the context of the opinions of others. Through this extended picture, a greater

understanding of meaning is elicited (through the context) and thus an appreciation of

the rationale for particular views. Moreover capturing all of the contributions provides

participants with a sense that the process is “just” (Kim & Mauborgne, 1991). From this

often quite extensive map3, a shared understanding begins to emerge. One objective is

for those involved to begin to understand, in use, what is meant by very general

descriptions of the proposed system (Checkland & Holwell, 1998).

Using Mapping to Develop an

Information System

As intimated above, developing a clear and shared understanding of the purpose of the

system by all who have some stake in its development and use is important for ensuring

that the system is used and used within appropriate bounds. Information is always given

meaning by its context, and often IS’s are designed with a presumption of context. A good

understanding can only be derived from a clear knowledge of purpose and the limits to

its application.

Understanding who the stakeholders might be, who will give the information meaning

and in what way might they be involved (or effect the usage of the system) is critical.

Stakeholder analysis and management techniques such as those described in Boddy,

Boonstra and Kennedy (2002) or Ackermann and Eden (2003) provide a good starting

point.

Ackermann & Eden employ particular forms of mapping to determine not only who the

stakeholders are, but also the details regarding their disposition, relationships with

others, and the nature of the power and interest they may use to influence the success

or otherwise of the information system. By building a grid whose axes are power and

interest, participants are able to position stakeholders according to their relative power

to influence success (as determined by the purpose of the IS) and interest in usage. The

grid usefully shows those who have both the interest and power to ensure success or

failure, and so attention to their views regarding the development of the IS will be crucial.

A better understanding of these stakeholders can be elicited by exploring in more depth

the bases of power and the nature of interest. Those who have power and no interest in

the outcome can easily determine failure (intentionally and unintentionally) and so must

be carefully managed. Mapping the influence network among these powerful stakeholdUsing

ers provides important clues as to which of them can be used as opinion formers —

increasing the chances of success. Analysis of these maps follows the same conventions

as those used to analyse causal maps.

By facilitating not only the means of contributing effectively to these deliberations, but

also structuring the contributions to enable effective management of the unfolding

complexity (rather than reducing it), sufficient and productive time can be spent in the

exploration stage. The purpose of the exploration is to consider in depth the emergent

properties and develop agreed action packages. This ability through capturing the

richness and diversity of views along with managing their structure represents the stages

of Intelligence and Design, in what Simon (1959) describes in his four stages of decision

making.

By describing a case study that employed mapping as a means of developing an IS,

aspects of the process and some of the benefits of mapping will be illustrated and

explored. As with most case studies in IS development, the material is sensitive, and so

the material presented below is less expansive than preferred.

Developing an Information System for

Student Tracking4

The organization discussed below is part of a large University. Rather unusually, it is a

self-funded unit receiving no public monies. Therefore, it operates in many ways as a

business with one of the main objectives being to not make a loss, and make enough

surplus to reinvest in its educational products. At least, sufficient revenue is required

so that the organization can pay staff, keep the estate in order, and provide its

commitment to a high quality total service to students. The major contribution to finances

is unsurprisingly student fees, although executive development programmes, amongst

other activities, contribute. Currently the unit delivers its key products in many different

locations (product offerings not only at home but in four countries in South East Asia,

four in the Middle East, and two European countries).

One of the consequences of having such a widely dispersed programme (at any one time

the school has around 2,000 students on its books) is that it is paramount that an effective

and efficient information system is available to track enquiries, manage admissions, and

monitor the progress of students. However as is often the case, an incremental growth

in locations and student numbers along with a growing recognition of the power of

information technology had resulted in an ad hoc approach to the development of

information systems. Consequently, there were six different systems being used, involving

a range of different programmes — databases, spreadsheets, and word-processing

packages. Getting data from one to another was problematic, frequent errors occurred and

thus accurate and timely information was difficult to attain. A new system was required.

In order to ensure that the new system was as effective as possible, it was necessary to

take into account views from staff with responsibilities for marketing, operations,

academic delivery, and unit management along with those from the computer support

group. Moreover in terms of operations, it was important to consider the views not only

of those working in the unit but also those operating in the different sites across the

globe. Their views needed to be included. Management reviewed the budget and it was

agreed that $350,000 would be available for software development with the total budget

for staff time and training aiming at a maximum of $750,000 total costs. They were keen

to ensure that the new system would be within budget and on time.

As a means of involving the different perspectives and also ensuring a sustainable and

achievable product, the unit decided to run a workshop using the causal mapping process

and associated software5. In this manner, it would be possible to capture all the different

views along with their explanations and consequences, thus building up what was hoped

to be a full picture of requirements, aspirations, issues, etc. Thus, the workshop was

expected to ensure that those attending felt an involvement in the system development

and that all users and the computer support staff understood the rationale for the

intended system.

The management team spent some time considering specifically who to involve, as they

wanted to ensure that two members of staff represented each constituency. They also

wanted to employ the services of a neutral party to facilitate, allowing those attending

to concentrate purely on contributing. Once the composition of the group was settled,

and a facilitator chosen, a date for the workshop meeting was set.

The day commenced with an introduction from the facilitator, who explained the process

to be adopted. Each participant had in front of them a networked laptop. Through this

medium they would be able to contribute their concerns, aspirations and requirements

simultaneously and anonymously. It was believed that this would be helpful, as firstly

there was a lot to cover in the day, and secondly there were some concerns that some

participants might feel constrained due to the presence of management. The intention

was to start with possible reasons for wanting a new system, before eliciting their

aspirations for the new system.

The group began by surfacing their concerns about the current situation — the reasons

why a new system was needed. They did this by typing in the statement and seeing it

appear both on their laptop workstation and on the public screen (which enabled them

to see the views of others and so build on them). It soon became clear that many found

this a cathartic process — they were obviously very fed up with the present system! After

about 10 minutes participants felt that they had covered this topic.

While they had been doing this, the facilitator had been working to position the

statements so that they were roughly clustered and hierarchically positioned. This

allowed her to then, with help from the group, begin the process of linking them6 together

to develop the causal network where the links represented causal relationships. For

example, the group believed that 5 “rigid structure of the systems rather than flexible

structure” led to 6 “unable to get decent statistical information” which in turn meant

that 7 “they experienced difficulties with managing exam board decisions” (See Figure

2). It was explained that the numbers associated with each statement were for reference

only and simply made manipulation easier, and that the three dots (an ellipsis) was a short

hand for “rather than” — enabling the capture of contrasting situations.

During the discussion of the triggers for a new system it became clear to the group that

they were already surfacing implicitly some of their aspirations for the system (or at least

the contrast of the goal). Two concerns, namely “drop out rates” and “inefficient

management of student progress” were, if managed, important (negatively expressed)

goals. To reflect this insight, the facilitator edited the original wording so that these

concerns now read in a more aspirational manner (for example, instead of stating “drop

out rates too high,” the statement now stated: “reduce drop out rates rather than drop

out rates too high”). In addition, to reflect the change in status of the statement, a new

style (representing through the use of a different font and colour a new category) was

created called goals and these two statements given this attribute.

Continuing, the group then spent time considering what might be other goals of the

system. Again using their ability to directly type in their contributions along with any

links (by now they were familiar with the process) a draft goal system emerged (see Figure

3). Not surprisingly, money featured prominently! However, as the group began to

explore the links between the goals, it became apparent that a number of participants felt

that they needed to consider the issues more fully — as they were unsure whether they

had surfaced all of the goals, or whether they fully understood them. It was time to open

up the discussion further.

Using a new view (similar to sheets in Excel) the group then began to consider the new

system more deeply. Contributions came rapidly. Within 20 minutes, another 50 or so

statements were surfaced. The group began the process of examining this new material

(see Figure 4) — suggesting possible links, amending statements to make their meaning

more clear, and adding new material. To ensure that there was equal representation, the

facilitator created a number of new categories representing the different stakeholder

Figure 2. The first stage: expressing concerns or triggers for a new system (note those

statements that are boxed are those that were converted to goals)

constituencies and with the help of the group applied these to the contributions. Having

done this, she was then able to note that both marketing and operations had contributed

the most (11 and 20 respectively) with computer services raising 10, academics only

raising four statements and management five. Not a surprising result given that both

marketing and operations would use the system most, with computer services having to

maintain and upgrade it.

This insight prompted further reflection and analysis. By examining the tails (those

statements that had no statement linking in/supporting it), it was possible to ensure that

all systems implications were addressed by ensuring that each tail was a statement of

system requirement, or of system characteristic. Where necessary, this meant adding a

statement from computer services staff regarding the particular function or action

required. Moreover the group were able to check that each system implication led to a

goal, thus ensuring that all system characteristics or requirements did contribute to the

stated overall goals. Both these analysis prompted the group to identify further goals

— increasing the system of goals from nine goals to 14 goals.

In addition, through using the software’s analytical features7 (discussed above) it was

possible to do some logic checks. Firstly, an examination of the “busy” points of the map

was undertaken (essentially comprising a count of the number of statements linking in

and out). The group were pleased and not surprised that six “unable to get decent

statistical information” emerged as the most busy (having eight statements linked to

it) with 7 “experience difficulties with managing exam board decisions,” 17 “efficient

Figure 3. The developing emergent goals system

Figure 4. Developing the map further

management of student progress rather than inefficient,” and 38 “academics provide

more appropriate advice to students” each having links to seven statements. The group

reflected that this seemed correct.

When examining the most potent statements (those statements that supported the

greatest number of goals and were therefore important action points) they were also

reassured to note that three of the four most potent statements were IT oriented: 28 “use

from dial-in facility,” 27 “make it machine operating system independent,” which was

supported by the third 39 “ensure all are able to use the system rather than those with

the relevant version of Windows.” All of these supported the final potent statement,

which was 21 “have real-time access to student progress.”

The workshop ended with a clear agreement of what was wanted from the system, which

addressed the needs and concerns of those who would be using it. As a result, it meant

that computer services staff were able to put together the document that would be used

for the tendering process. In addition the map acted as an organization memory, enabling

renegotiation as the development process unfolded and a number of the agreed actions

had to be reviewed.

Postscript

The system was seen as very successful. Eighty percent of the users were pleased by

the result, the total cost was within budget and the system was delivered on time. The

software is the sole record-keeping system in use, substantially reducing errors, and

preventing the need for data having to be re-entered. The system, to date, has never failed

and application software updates, data migration issues and the occasional bug causing

run-time errors on individual desktops are the only causes of system downtime.

However as with all systems, now that the system exists there are additional features

desired by users. As with most information systems, these are learned from using the

system — with users not aware of them until after the system has been completed —

unfortunately mapping can’t pick up these (echoing the findings reported by Orlikowski,

Walsham, Jones & DeGross, 1995).

This case provides a brief illustration of how the mapping process supports multiple

stakeholders in determining an IS development strategy for which all felt a high degree

of ownership. It reflected the views of all there: both users and developers — quiet

members of staff as well as those who were socially confident. The resultant model helped

members to develop a common language and based upon this negotiate an agreed way

forward. Most importantly it enabled participants to understand the reasons for requirements

and how they causally linked to supporting goals.