A comment on Supernak's critique of transport modelling

Abstract

In a recent article in this journal Supernak (1983a) advanced a forceful critique of contemporary urban transport modelling. This note re-considers some of the issues discussed by Supernak, but from a different philosophical point of view. The resultant analysis of contemporary modelling differs importantly form Supernak's and leads to a different set of prescriptions for the reformation of the field.

Urban transportation modeling: The discussion continues

“…We agree with Polak that '...the search for "significant" regularities as indicators of "laws of behavior" may be misleading and ultimately unsuccessful' (Polak, 1987), especially if the analysis ignores contextual and causal relationships. We must seek reliable functional relationships in our data, but cautions are in order.…”

“…This apparent misunderstanding influenced -not surprisingly -the entire line of Polak's reasoning and was the main reason for stating that his 'resultant analysis of contemporary modeling differs importantly from Supernak's and leads to a different set of prescriptions for the reformation of the field' (Polak, 1987). This paper is our response to issues discussed in Polak's article.…”

“…His analogies from natural sciences were intended only to illustrate some basic methodological points, but were interpreted by Polak as indicating that a) 'transport phenomena are essentially law-like and that they manifest certain empirical regularities,' and b) 'transport phenomena are, in essence, no different in nature from the subject matter of the natural sciences and are therefore amenable to study in the same manner as the physical world' (Polak, 1987).…”

“…Supernak's 1983 call for clear structural definition and proper stratification of analysis units, in relation to situational specifics implies a program for achieving what Polak too apparently considers essential, viz., understanding in terms of attention paid to situation specifics, but we reject such vitalistic notions as 'the nature of the modern car' (Polak, 1987) as explanatory constructs. Logical distinctions -especially such vitalistic ones -may obscure functional relationships!…”

“…As in the early development of scientific modeling in other fields, transportation modeling has often suffered from too great a reliance on logic, divorced from empirically-based understanding of the real world phenomena being modeled. Polak (1987) asks: 'How good is a model that proves consistently inaccurate in its forecasts?' Although there are deficiencies in existing models that adversely influence accuracy of forecasts, poor forecasts don't always result from worthless models.…”