A model is a conceptual tool which serves for abstracting the important elements in a social phenomenon and to show how these elements fit together. A good example of a model is an underground map. These maps are topological – only the routes are represented and these are reproduced in consistent intervals and angles. A tube map puts in (or ‘abstracts out’) what people need to know (connections and lines) and leaves out what they do not need (eg. roads, railway lines, precise distances, names of shops and pubs). It is exceptionally easy to follow a tube map and difficult to plan your journey without one. In contrast general maps, even ones that can be tailored on line to your requirements, are more hit and miss. They are often difficult to follow as the detail can be either overwhelming or at times you can be left frustrated by the absence of detail. There is an enduring tension between fidelity (a model needs to represent reality) and usability (a model needs to abstract out the important details).

Models are often dynamic or at least expected to show how elements interact and influence each other. Dynamic models can be produced using computer simulations, for example economic events can be modelled in order to explore what if scenarios. Static models are, well, static but can be supported with double headed arrows and loops to show how elements interact (see for example the oft repeated diagram of reasoned action in Fishbein, & Ajzen, 2010).

There are many types of model. For example Martindale (2010) differentiates between analytical models and sensitising ones, the former seeks often to explain why something happens and to offer a degree of prediction, as against a sensitising model which is inviting someone to view the situation in a particular way. One problem here is that once represented in diagrammatic form models seems to take on a life of their own and become predictive even if they were designed only to be sensitising. (Activity theory is a good example of a model which was intended to be sensitising but ended up in some research as a causal – see discussion in the case studies section.)

Statistical models often appear impressively supported by correlational measures. However they should not be over interpreted. They show what is happening at a particular point in time, based on a set of assumptions. Those assumptions may change and people might behave in different and unexpected ways.

Mathematical modelling has often been critiqued for paying too much attention to association and less to explanation. Yet without just such explanations the relationship between elements may be misunderstood. The direction of impact will not be known and indeed both elements are better explained via a third factor. This tendency towards association without explanation has been sharpened by the push for Big Data studies and data analytics (e.g. Kitchin, 2013). There are, in contrast, many examples of modelling which do engage deeply with evidence and do search for explanation as well as association.

Finally models may be used in a more educative ways, i.e. to promote action based on what would happen if things stayed the same. Sterman et al (2015) offer an example in the form of a role play simulation on climate negotiations – this a kind of game theory modelling.

References

Fishbein, M., & Ajzen, I. (2010). Predicting and changing behavior: The Reasoned Action Approach. New York: Taylor & Francis.

Kennedy, T., Smyth, R., Valadkhani, A., & Chen, G. (2017). Does income inequality hinder economic growth? New evidence using Australian taxation statistics. Economic Modelling.

[This is a case study of economic modelling looking at the effect of income equality on economic growth. The study appears very technical if, like me, you are not an economist. However the essential ideas are easy to follow and it appears likely from this case, and from other cases cited, that income equality may negatively impact on economic growth. This is important to know and to consider but many of us would hold back from offering this as a generalisable law – what comes over from a raft of examples might be fixed for a certain period of time or it might not, we don't know with certainty].

Kitchin, R. (2013). Big data and human geography: Opportunities, challenges and risks. Dialogues in human geography, 3(3), 262-267.

Martindale, D. (2013). The Nature and Types of Sociological Theory, London: Routledge.

Sterman, J., et al. (2015) World Climate: A role-play simulation of climate negotiations. Simulation & Gaming, 46(3-4), pp.348-382.