The field of psychology offers a great deal of research material on how people form their opinions: what do we support, what are we against and how does this decision-making process occur? In his doctoral research project, Jonas Dalege applied a mathematical model to reduce this plethora of research results to a handful of basic principles in order to predict behaviour.
Dalege's doctoral research integrates multiple effects on opinion-forming, or attitudes, into a single mathematical model. He examines, for instance, the mere-thought effect, which is the effect that people tend to become more extreme in their opinions when they have had more time to consider their own judgement. Something that was 'fine' will become 'great' over time, and what was 'not so great' will become 'no good at all'.
Dalege combines this mere-thought effect with other aspects, such as the effect that people are receptive to substantiated arguments when forming opinions on topics they feel are important – yet when the topic of the opinion is less important to them, people are more easily swayed by the status of the speaker.
By integrating these different effects into a single mathematical model, Dalege reduces them to a few basic principles:
• Opinions tend to become unstable and unclear, and therefore less informative, when we do not focus our attention on them.
• Thinking about a particular topic serves primarily to develop a clear and stable opinion. In this process, the person's opinion is influenced by various aspects (if you find a certain politician likeable, for instance, you will probably decide that he or she is competent as well).
It is then possible to predict behaviour with the help of these basic principles.
'We consciously elected to work with an extremely simple model,' Dalege explains. 'It's more effective to first see what can be explained using a simple model, because it's easier to then progress to a more complex model than it is to work your way back from complex to simple. Another advantage of a simple model is that it is easier to make predictions than when you are using an extremely complex model.'
Dalege adds that a crucial test for any general theory is whether it can explain established phenomena. A central objective was therefore to first conduct research to determine whether the applied model was up to the task.
Because it provides concrete results on the process of exerting influence, the model can offer vital knowledge to those who work with 'influencing factors' in real-world practice. Take, for instance, the National Institute for Public Health and the Environment, which wishes to promote healthy behaviours among citizens, or the government that wants to prompt individuals to form climate-neutral habits.
'Our model demonstrates, for example, that continuously offering substantiated arguments in order to convince the other party may have the opposite effect. That's because the other party is constantly looking for arguments that are in line with who they are as a person, especially when they have had more time to think about a particular topic,' Dalege explains.
The next step is to use the model to assess the number of predictions. The theory is currently focused on the opinions and assumptions that people already have, but in the future, the plan is to devote additional attention to how those opinions are formed in the first place.
The research was conducted using computer simulations in a model that was derived from statistical mechanics. This relied on the use of virtual subjects whose parameters were based on those of real-life subjects from the large-scale survey previously conducted into effects on opinions. The study revealed that the mathematical model can, in fact, successfully integrate many significant effects.
Jonas Dalege, Denny Borsboom, Frenk van Harreveld & Han L. J. van der Maas (2018), 'The Attitudinal Entropy (AE) Framework as a General Theory of Individual Attitudes', Psychological Inquiry, 29:4, 175-193, DOI: 10.1080/1047840X.2018.1537246