Full text: ARCH+ : Studienhefte für architekturbezogene Umweltforschung und -planung (1969, Jg. 2, H. 5-8)

Recognizing the merit of these arguments, we emphasize 
again our contention that the process of modelling, not 
the model itself, is often the most valuable aspect of this 
effort. - The planner who is involved in the development 
of a model can gain new Insight into urban phenomena 
and relationships. The above arguments present issues to 
be studied, problems to be met and attacked, not ex- 
cuses for chaos. 
This is not to argue that planners should try to model the 
world. The appropriate role of models in urban planning 
today is admittedly limited and qualified. 
But that role is important: 
- to assist In the development of theory 
- to provide a laboratory for testing policies and pro- 
- to emphasize evaluation of short-term relationships 
- to provide predictions and projections 
Figure 1 is a general schematic representation of the 
urban planning process. We believe that the greatest 
potential for models and modeling today lies in the area 
of theory development and project evaluation, represen- 
ted by that part of the diaaram within the dotted box. 
While much can be learned from observing a finished 
model, or by scanning a computer printout of its applica- 
tion, more can be learned from the discipline of being 
required to state assumptions and theories explicitly, and 
by the opportunity to test one’s ideas of urban relation- 
ships while groping for the model’s parameters and form. 
The recurring theme of this section will be the value of 
the modeling process; we will carry it through our dis- 
cussion of model development, validation, and appraisal, 
and use it as the essential principle in our argument for 
the involvement of certain people in the effort . 
Model Development 
In an earlier paper (5) we described the four elements 
central to the development of a model: 
- model subject - what is the model about? 
- model function - what is the model to do? 
- model theory - on what theory is the model based? 
model method - how does the model use its theory? 
Defining these elements in response to the above questions 
is the first step of a modeling effort. 
Roughly speaking, the subject of the model is that 
entity or activity which is projected, allocated, or 
manipulated by the model. There are four essential classes 
of subject matter in current urban models: land use (6), 
transportation (7), population (8) and economic activity 
(9). The function of an urban planning model is usu- 
ally to project or allocate the subject, or to derive new 
subjects from it; most models perform two or more func- 
tions in varying combinations. The theory underlying 
the model is that set of relationships - stated or implied 
assumed to prevail between the model subject and the 
larger environment. Distinguished on this basis, models 
generally can be sorted into two classes: micro-analytic 
behavior or "choice" models, and macro-analytic 
growth-forces or "index" models. The operational 
method of the model is the mathematical or symbolic 
form used to carry out the projection, allocation, or 
To these thought, and without repeating the content of 
the earlier paper, we would add some general observa- 
tions on the availability of data and on the number of 
factors or variables included in the model. Clearly, the 
purpose for which the model is being developed is basic 
to decisions about data and factors included. If the model 
is to be a generalized or research effort designed for the 
examination of relationships between urban phenomena, 
or to state a general theory in symbolic and manipulative 
form, its essential purpose is to increase knowledge and 
insight about fhe processes studied. For such models 
hypothetical or generalized data can be used, and the 
number of factors or variables can be limited by assump- 
tions. On the other hand, if the model is to be applied 
in an ongoing planning effort, accuracy in depicting the 
real world is desirable. A specific set of data generally 
must be used, and simplifying assumptions cannot be 
easily made. 
Although an applied or practical model is usually devel- 
oped for a particular city and used only in its planning 
activities, the time has probably come when we should 
try to generalize these models by extending them to other 
cities. If these models do not seem to "fit" when tested in 
other, similar cities, their validity - even for the city 
for which they were built - may be challenged. It would 
seem that data from one city should apply to other cities, 
at least within certain class categories. For instance, 
differences in natural setting such as topology, climate, 
spatial distribution, and so forth, should not affect func- 
tional data. Moreover, we should be able to rely on the 
stability of large numbers: the social and economic be- 
havior of large groups are relatively invariant and pre- 
dictable. (This is fortunate indeed, since the physical 
construction that planners undertake based on this behav- 
ior is quite durable.) 
In the development of applied models the data available 
Frequently determines the nature of the model and this is 
sometimes detrimental to the effort. In a practical situa- 
ton, with time and money constraints, most analysts be- 
gin by surveying the available data. They then attempt 
to shape the model to make the best use of these data and 
avoid as much as possible the expense of gathering new 
data. This approach can severely restrict the choice of 
theory and method employed in the model. For example, 
if a model depicting market behavior were planned and a 
look at the data showed that information about market 
suppliers would have to be specially gathered, the analyst 
might well reconsider his approach and select a gravity 
model instead. This is understandable; little is to be 
gained by developing a model meant for practical applica- 
tion, for which there is no data. 
The level of aggregation of the subject matter also will 
depend to some extent on the availability of data: If the 
necessary data are available on municipal divisions only, 
the analyst is well advised to structure the model on this 
basis and not on census tracts - unless, of course, he is 
orepared to generate his own census tract data. 
The seriality of data may also determine the form of model 
used for if the analyst can obtain data for only one point 
in time - and few areas have data for more than two or 
three points - he may be forced to develop an equilibrium 
modeF rather than a time-series or growth-forces model. 
ARCH+ 2 (1969) H.8

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.