ABSTRACT
Yet if models are to be used, it is important to clearly define and separate the stages of procedure and to realize the limitations of each stage. In this way confusion and mistrust may to some extent be avoided. For example, a basic distinction is that between model and theory as procedural stages. ‘ Models . . . constitute a bridge between the observational and theoretical levels’ (Haggett and Chorley, 1967, p. 24). They are simply aids to the con struction of new theories concerning observed information, and more than one model may appear to fit the same data equally well. The predicted theories or hypotheses, on the other hand, must be testable by reference to known or knowable facts. If, due to the limitations of the data, it cannot ever be tested, then a hypothesis cannot be considered useful. Examination of a hypothesis may result in its retention, modification or rejection (Clarke, 1968, p. 643). A
further result of the application of models is that a certain amount of ‘ feed back5 may occur. For example, it may be found desirable to gather new types of information or to stress rather different problems. Another sort of feedback is that new information might be learned about the models themselves and possible modifications suggested.
