This paper focuses on partial models for solving urban problems to contrast our achievements as social scientists with our aspirations as prescribers of public policy. The objectives of this paper are (1) to review some of the reasons that an ideal set of solutions for urban problems has not been produced by social scientists and (2) to describe two efforts currently underway at the Urban Institute to arrive at useful, if not ideal, answers. An ideal methodology for seeking answers to urban problems would incorporate the principle that nearly everything is related to everything else. Thus the model would be interactive in the sense that a given variable could be both an independent variable and a dependent variable to allow for "feedback effects". The urban metropolitan area and its population would be regarded as a complex organism in which a diverse set of outcomes would follow from the decisions of various subsets of decision makers, each subset complying with known behavioral relationships within specified constraints. Another paradigm for this kind of modelling effort is a micro-analytic simulation model. The methodological base for this type of modelling is the assignment of behavioral characteristics to individual micro-units according to known probabilities of occurrence of those behavioral characteristics within identifiable subgroups of the population. (Author/JM)