A Method for the Nonstatistical Assessment of Behaviorally-Based Policy Models

Abstract

Advanced computer-technology, sophisticated econometric computer-software, and the increasing availability of disaggregated data have all facilitated the development and use, for planning and policy analysis, of statistically-estimated mathematical models of human behavior. In this paper the author presents a method to judge subjectively the quality of these models and illustrates how the results of the method can be used, with statistical measures of the quality of a model, as indicators of the predictive abilities of the model. Ideas are drawn from the literature on artificial intelligence, and the concern is how to represent and manipulate subjective judgments. Five models which are related to household energy-use behavior are subjectively assessed, statistically estimated, and used to predict future behavior. In four of the five cases, the subjective assessments, in combination with the statistical measures, correctly indicate the quality of the predictive accuracy of a model.

Get full access to this article

View all access options for this article.

References

Ajzen

Fishbein

, 1980 Understanding Attitudes and Predicting Social Behavior (Prentice-Hall, Englewood Cliffs, NJ)

Bosher

Schweppe

Gruhl

, 1978, “Validity and uncertainty of predictive energy models”, in Proceedings of the Sixth Power Systems Computation Conference, Darmstadt (IPC Science and Technology Press, Guildford, Surrey) pp 93–99

Buchanan

Shortliffe

, 1984 Rule-based Expert Systems (Addison-Wesley, Reading, MA)

Longley

P A

, 1984, “Comparing discrete choice models: some housing market examples” in London Papers in Regional Science 14. Discrete Choice Models in Regional Science Ed. Pitfield

D E

(Pion, London) pp 163–180

Macgill

S M

, 1986, “Research policy and review 12. Evaluating a heritage of modelling styles” Environment and Planning A 18 1423–1446

Prade

, 1985, “A computational approach to approximate and plausible reasoning with applications to expert systems” Transactions of the Institute of Electronic and Electrical Engineers, PAMI (Pattern Analysis and Machine Intelligence) 7 260–283

Shafer

, 1976 A Mathematical Theory of Evidence (Princeton University Press, Princeton, NJ)

Tonn

, 1986, “A framework for the nonstatistical assessment of behaviorally-based policy models” Environment and Planning A 18 799–802

Tonn

Berry

, 1985, “An illustration of household energy-conservation modeling using the search—list decision-heuristic” Environment and Planning A 17 1243–1261

10.

Tonn

Uppuluri

, 1986, “Uncertainty methodologies used in expert systems: a synthesis of approaches”, in Proceedings of the Colloquia, International Symposium on Methodologies for Intelligent Systems ORNL-6362, Oak Ridge National Laboratory, PO Box X, Oak Ridge, TN 37831, USA; pp 59–68

11.

Tonn

White

, 1986, “Residential wood-use in the Pacific Northwest: 1979–1985” ORNL/CON-216, Oak Ridge National Laboratory, PO Box X, Oak Ridge, TN 37831, USA

12.

Wrigley

, 1982, “Quantitative methods: developments in discrete choice modeling” Progress in Human Geography 6 547–562