Causal Mapping
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
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.