In order to derive inferences about likely actions in a situation (and, thus, about the pattern of joint results that may be produced), an analyst must make assumptions about the characteristics of the individuals in a situation. Minimally, these involve assumption about four types of variables:
1. How individuals process Information.
2. The way individuals assign values to actions and outcomes.
3. The way individuals calculate choices.
4. The resources available to individuals.
At the most general level, an analyst puts himself into the position of each of the participants in a situation and tries to reason through the objectives that the individual would pursue, what resources they would bring to the situation, how much knowledge they would have, how they might learn from experience over time, and what type of calculation process they would adopt. Having done this, the analyst infers the likely behavior of different participants and how they would or would not be led to stable results. The assumptions about the individual are the components of an analytical engine that activates (i.e. gives motion or is a moving part) an action arena enabling analysts to predict the results likely to occur.
The most fully developed, explicit model of an individual is the extreme rational choice model used in microeconomic theory and in game theory. Even though this model makes "unrealistic" assumptions about information processing capabilities of individual, and about the evaluation and calculation processes, this model has been useful in generating verifiable predictions in a variety of tightly constrained action arenas. These predictions have considerable empirical support in those "real-world" situation which come closest to approximating the tight constraints of such models. Given the simple structure of some highly repetitive, competitive market situations, the limited capacity of human beings to code (develop appropriate language structures), store, and process information may not be exceeded. In such constrained and simple situations, it is analytically useful to assume that individuals have complete information relative to the tasks they face and the way the action processes are organized.
The extreme rational choice model is, however, an inadequate behavioral model for applications to more complex situations. Institutional analysis is not limited to the use of any single pre-defined model of the individual. Considerable work in cognitive psychology, experimental gaming, and management science uses models of the individual involving different assumptions about the information processing capabilities and valuation procedures than the extreme rational choice model. These efforts draw intellectual inspiration from the work of Simon, Kahneman, Tversky, and others.
In more open and complex situations, we presume that individuals are adaptive, learning and fallible organisms who seek multiple, and at times, conflicting objectives. These assumptions lead us to generate propositions at the theoretical level which differ rather markedly from the conclusions derived from much of modern administrative theory which implicitly relies on a different model of the individual. When one presumes that humans make errors, one is led to a type of theory which stresses the necessity for redundancy. When one presumes that public officials may have interests that differ from those they represent, one is led to a type of theory which stresses the importance of countervailing powers.
⇒ The Prediction of Results
~
Elinor Ostrom, “Formulating the Elements of Institutional Analysis,” in Studies in Institutional Analysis and Development (Conference on Institutional Analysis and Development, Washington, DC, 1985), 24, https://dlc.dlib.indiana.edu/dlc/handle/10535/2145.