Building a conceptual model
- The text on this page is taken from an equivalent page of the IEHIAS-project.
The purpose of the conceptual model is twofold:
- To help communication, and ensure agreement, between the various stakeholders about what the 412issue actually is – and within that what is important and what is not;
- To provide a clear and concise description of the system of interest (i.e. all the main components and their relationships), as a basis for assessment.
Building conceptual models is not easy, and several, repeated attempts are likely to be necessary. One of the most difficult challenges is to define clear boundaries to the issue, in ways that do not unfairly limit or bias the assessment. (see link to Setting boundaries).
Initially, also, the models often become progressively more complex, as new elements of the issue (and in some cases new stakeholders, with different interests) emerge. Later, however, it is helpful to simplify the models, in order to remove elements that are not important or unclear, and focus on those aspects that are considered most important (see the link to the examples to the left, and under See also, below).
Conceptual modelling can be done largely intuitively – for example, by simply drawing up a diagram through a process of collective debate. More formal methods are available, however, which help to rationalise the process; probably the two most useful and flexible are mind-maps and systems diagrams.
Which is most appropriate is likely to depend on the experience and skills of those involved, as well as the complexity of the issue. No matter which technique is used, it is also helpful to structure the model in terms of the causal chain (or some other, predefined framework) in order to make sure that it is logical, consistent and complete.
In the real-world, almost everything is ultimately connected with everything else. The sorts of complex issues that are often considered for integrated environmental health impact assessment therefore have ambiguous and porous boundaries, in all sorts of ways:
- In terms of original causes – the roots of many health problems can be traced back through an almost infinite sequence of antecedent causes into factors that are very far-removed and remote;
- In terms of impacts – direct health effects often lead to an ever widening range of secondary and further consequences, not only in terms of health but also other social and economic impacts;
- Spatially – many issues are not confined to narrowly defined geographic areas but instead extend their roots or effects, albeit indirectly, across a wide and ill-defined area;
- Temporally – many health issues involve long delays between cause and effect, and lead to secondary effects that can persist for many years (or even generations).
These four dimensions of an environmental health issue are not wholly independent. Often they interact. For example, short-term effects from proximal causes tend to be more localised and the consequences more direct, because opportunity for the risks to spread more widely are limited. In contrast, more remote causes tend to imply longer timescales and larger spatial scales for analysis, because time is needed for the causes to translate into human exposures, and in the process the hazard has the opportunity to spread through the population.
It can therefore be useful to map issues in relation to these dimensions, to help see how far to go along each one, and thus where the boundaries might lie. One way of doing this is to generate a form of polar diagram, as below. Examples of diagrams for different issues are also shown. As can be seen, some issues imply very localised and short-term causes and impacts; these may not always merit full integrated assessments, for the causal pathways are short and direct and the problems are highly constrained. Others - the ones that probably need a more integrated approach - have much broader extent in terms of their causes, consequences or time and space scales. Many issues also clearly comprise a nested series of problems, operating at somewhat different scales and with more or less indirect causes and effects. The polar plots can help to make these explicit, and thus encourage a more informed decision about where to place the boundaries to the assessment.
Guidelines on boundary setting
- In general, it is better to be more, rather than less, inclusive at the issue-framing stage since, if the assessment methodology is correct, factors which are not important in reality will lead to small effects, relative to the uncertainties involved, and thus contribute little to the central estimates of impact;
- Factors should not be omitted simply because of the lack of firm evidence or data, since this biases assessments to things that are already well-known and measured. Instead plausibility of effect should be the primary criterion for inclusion. Factors which are less well-evidenced, however, will introduce greater uncertainty into the assessment (and thus lead to a wide variation in the estimated impacts).
- Extending assessments over a larger area than necessary will increase the effort and cost involved, and may exacerbate problems of finding relevant data. In proportional terms (e.g. where impacts are measured as a percentage change in mortality or cost), it may also lead to some dilution of impact - though where outcome indicators are measured in absolute terms (e.g. as excess mortality or cost) this will not be an issue.
- Assessments should not be constrained by artificial boundaries, which ignore significant transfers (e.g. of pollutants or affected people), since this will result in an under-estimation of the overall impact. Geographically, therefore, assessments should generally be based on ‘catchment areas’ or ‘zones of influence’, which reflect the full extent of the processes of concern.
- Extending assessments over too long a period (e.g. too far into the future) will likewise increase difficulties in carrying out the assessment, and may increase substantially the uncertainties involved. If the time-period for the assessment is unduly truncated, however, substantial under-estimation of the impacts may occur, by ignoring delayed and secondary effects.
- On the whole, assessments should aim to cover the full life-cycle of any development (e.g. policy, technology), and at least the full life of those exposed to any risks (or likely to receive benefits) within that period. Assessments should be extended wherever there is evidence for direct inter-generational effects (e.g. via reproductive health).
Mind-maps are a simple means both of brainstorming ideas about an issue, and representing the results in the form of a clear, structured diagram.
The original procedures and rules for making mind-maps were very rigid, and their purpose was primarily to help people learn. In recent years, however, they have found a wide range of applications in many areas of science and decision-making, and methods of mind-mapping have accordingly become more flexible and diverse. Non-expert users thus find them relatively easy to apply.
Mind-maps can be considered to comprise three main components:
- One or more ‘end points’ – in this context, the things that are being assessed
- A series of ‘items’, each linked directly or indirectly to one of these end-points
- The links between each item or end-point (limited to one forward and one backward link for each)
The basic steps in mind-mapping are as follows:
- Starting with a blank page, enter first the object(s) which are to be assessed
- Working outwards, identify the key items that influence these objects first directly, and then indirectly
- Link each item to the entity (another item or end-point) that it affects with an arrow or line
- Review the resulting diagram, and make any changes considered necessary
A growing volume of guidance on mind mapping is now available, and this is supported by a wide range of software tools, many of which are freely downloadable.
System diagrams (also known as causal loop diagrams or flow diagrams) provide a structured means of representing complex systems. They differ from mind-maps in several ways:
- they are based on a more formal and complex set of principles and rules, which recognises different types of entity (or component) and relationships or links between them;
- they are explicitly structured to represent flows (of information, matter or energy) or causal relationships between the components of the system;
- they recognise and allow for reciprocal links and feedback.
Because of these more sophisticated concepts, system diagrams come much closer than mind-maps to providing a basis for actual modelling of the system. Links, for example, can be represented by quantitative rules or mathematical formulae which describe the relationships involved; each entity is hence a product of its inputs.
Types of system diagram
In practice, any issue can be seen in different ways. Different types of system diagram can thus be developed, depending on the perspective adopted, and the interests and needs of the users. A wide array of different (and often overlapping or contradictory) typologies and terms have emerged to distinguish between different diagrams. In the context of issue-framing for impact assesment, however, three main approaches can usefully be recognised:
- Structural (or functional) diagrams, which break down the system into its component parts, and show the functional (or in some cases spatial) links between them; these are used primarily to identify what factors need to be included in the assessment.
- Transactional (or pathway) diagrams, which describe the flows of pollutants or other agents (e.g. disease carriers) through the system; these are used mainly to identify the exposure pathways that need to be followed in the assessment.
- Transformational (or process-response) diagrams, which represent the changes that occur in the system as a result of these interactions and flows; these are used to highlight the impacts.
Selecting and designing system diagrams
All three of these approaches are clearly relevant in issue-framing, for no one of them captures everything that might need to be considered in the assessment. It may therefore seem sensible to try to combine them into a single diagram. This, however, tends to result in a very complex diagram that is difficult to understand. In most cases, therefore, it is better to settle on a single approach and supplement this with textual notes to describe the aspects that have been ignored; or a suite of diagrams can be developed, representing the different perspectives on the issue. If appropriate software is used, these can be hyperlinked, so that users can readily move between them.
Because of their inherent complexity, system diagrams need to be built according to a clear set of rules. It is important, for example, to distinguish clearly between the elements that make up a system: the compartments (or entities) from which it is built; the agents that pass through the system and connect (and are stored within) these compartments; the properties (or attributes) which characterise each compartment and agent; and the factors that influence these properties. It is also important to use clear and consistent symbols to represent each of these elements. Many software packages are available which provide a predefined set of symbols, and apply helpful constraints in diagram-making.
In some cases this complexity limits the use of system diagrams as a means of collaborative issue-framing. On the other hand, even when other techniques (such as mind-maps) have been used to conceptualise the issue, it is often helpful then to create a system diagram, in order to give a more rigorous framework for the assessment process.