Opinion pooling is a method whereby actors who hold different opinions from one another are able to change their opinions according to the opinions of other actors to whom they are related. This process of change is based on the assumption that actors are rational, respect the individual integrity of each other’s opinion, and are prepared to make a genuine compromise between the opinions of others to whom they are linked. A true consensus where all the actors are in agreement can only occur however if the actors are linked to one another, either directly or indirectly. We first describe this process and we then present the standard linear model of opinion pooling, illustrating how the dynamics of opinion change takes place and how an equilibrium occurs. We illustrate this using numerical values given to the opinion of each actor which at the beginning of the process are a set of ‘best guesses’. This sequence converges to an agreed number which is the ‘agreed guess’ at the end of the process. We then illustrate the process using a set of opinions based on the desirability of different locations defined as a ‘map’. We show how a unique equilibrium emerges and then we change the focus and assume the desirability indices are coordinates of places where actors are located. When rationally averaged, these coordinates change and the process leads to actors moving to the same location. We finally illustrate this process using a highway location problem, and then conclude by suggesting how we can predict the form of the network rather than simply focussing on the process of pooling.
Authors: Michael Batty
Download CASA Working Paper 220 (file size 3.5MB, file format PDF)