Research techniques and findings
Relation to neighboring disciplines
Regional science, a new interdisciplinary field within the social sciences, focuses on the locational dimension of human activities in the context of their institutional structure and function and on the significance of this dimension’ in the understanding of social behavior and forms. The locational dimension identifies the spatial relations of people to their activities and also to the natural and man-transformed physical environment. Regional science relies heavily on mathematical models to frame its theories and draws on the theories and findings of other social sciences, particularly location theory (Alonso 1964; Isard 1956a, chapter 2). [SeeSpatial economics, article onthe general equilibrium approach.]
The word “regional” implies the systematic approach to space in the sense of the human habitat. “Science” expresses the intention to apply rigorous techniques of investigation and to develop theoretical structures and concepts of general applicability. And “regional science” is meant to connote a field that transcends the bounds of any one social science discipline: it is related to regional economics, ecology, theoretical geography, regionalism in the sense of the political scientist, and parts of other social sciences, but differs from these fields in that it takes a more general approach to the role of space in social phenomena. Regional science also has a close affinity to a number of applied fields keyed to problems of adaptation to, or manipulation of, phenomena in space. Among these fields are city and regional planning, transportation, public administration, agronomy, and industrial engineering.
This article first discusses the formal structure of regional science, including the types of analytic propositions and conceptual primitives that characterize its studies. We then turn to the research areas that have been most successfully explored and report on some of the major findings; we thereby illustrate the substantive content and the bounds of the field. Third, we comment on the relation of regional science to neighboring disciplines; and finally we touch on the educational programs in regional science and the state of the profession. We restrict this review to work in the United States; parallel activity is under way in Europe, in Japan, and, to a lesser degree, in India, Pakistan, Latin America, and elsewhere.
Like other social sciences, regional science has three facets, which, in any one study, may emerge. Phenomena in space and choices related to such phenomena are presented and analyzed from each of three viewpoints: the normative, the descriptive, and the deductive. Thus, the distribution of urban places in a region, such as a nation or a part of a nation, can be approached normatively, as in the evaluation of conditions in many developing nations with high concentrations of population in metropolitan capitals (e.g., Harris 1959 and references cited therein); descriptively (or behavior-istically), as in the analysis of existing distributions of hierarchies or rankings of urban centers (e.g., Zipf 1949, chapters 9 and 10); and deductively, as in the development of a theory of regional urban structure from propositions regarding personal demand functions, distances between consumers, economies of scale, and transport rates (e.g., Isard 1956a, chapter 11).
The regional scientist studies three major classes of decision makers: individuals (or households), entrepreneurs (businessmen or firms), and public bodies (such as city governments and regional planning organizations). These classes are of interest for two distinct reasons. First, each makes locational choices. Second, each responds to its environment, among other ways, by analyzing and anticipating the locational behavior and choices of the various elements of its environment. Thus, the regional scientist is concerned with developing appropriate techniques for studying locational choices as well as the consequences of such choices (for example, the resulting patterns of urban settlements ).
The substance of the decisions represents yet another facet of the regional scientist’s interest. He seeks to include in his analysis diverse economic, social, political, and other variables and constraints as these are formulated in a spatial context (Isard & Isard 1965; Isard & Tung 1964). The decisions that are at the center of regional science analysis are those which establish the locations of various classes of activities in the environment. These decisions determine the scale of each class of activity at a given point in space and the nature, magnitude, and direction of flow of people, goods, funds, messages, and so forth between various loci of activity. The basic task, then, of regional science analysis is the identification of (a) location, (b) activity magnitude at the location, and (c) flow between locations. Regional science sees these as part of an interrelated system. Different locations involve the substitution of, for example, transportation charges for other factor costs. Different locations may also involve changes in the scale of operation as, for instance, market areas change. Furthermore, any change in location or scale at a given point has a concomitant effect on the flows emanating from and coming to the point. In short, regional science location analyses focus on questions of substitution of various categories of costs and revenues. The fundamental research problems are the determination of actual changes in accounts as spatial shifts occur, the identification of optimum location patterns under particular sets of welfare assumptions, and the identification of equilibrium conditions under postulated socioeconomic environments (Isard 1956a, chapters 4-10).
The unique quality of regional science is not only its concern with the location decision and the individual decision maker within an environment but also its concern with the location itself and the locational framework. The locational framework, as a space, and locations, as points within the space, remain concepts that are analytic primitives. Points define locations in a bounded space; this space is the set of possible locations; and specification of points in the space establishes distances between points. But distances can be expressed in many ways. The distances of interest may be expressed in terms of physical units (e.g., miles), in terms of travel time, in terms of travel cost, or in terms of any of a variety of sociological attenuations. These distances can, to some extent, be translated into each other. Development of a satisfactory aggregative weighted distance measure awaits further progress of the field, yet it is clearly a requisite for a valid regional science (Dacey 1963; Deutsch & Isard 1961; Olsson 1965; Webber 1964).
Total space is not undifferentiated or homogeneous; neither are patterns of activity spatially chaotic. Regionalization is the task of dividing the total set of points into relevant locationally defined subsets. In order to divide an area into regions, the analyst is faced with four tasks. He must identify the population, or what might be labeled the “constituent set,” which is the set of points, lines, or areas (or the phenomena these represent) that are to be disaggregated. Second, the analyst must position these in the set of locations in space. Third, having a problem or a hypothesis in mind, he must establish one or more scales that “measure” a member of the population in terms of relevant attributes. Thus, the population is specified in terms of location and in terms of characteristics. Finally, cutoff points must be chosen along the attribute scales; this procedure places the member of a constituent set in one category or another, each category corresponding to a partitioning of the location set (Teitz 1962; 1964). The result is a pattern of regionalization; but because of the choices made at each of the four stages, an infinity of regionalization systems can be generated for any one constituent set. Little appears to be gained from the quest for an ideal set of regions; rather, each task or problem has associated with it a given set of regions in a system of regions (Isard 1956b).
A system of regions can be described in terms of several properties. First, is the system exhaustive? Second, is it disjoint? These two properties establish whether a given location is defined in terms of the region system and, if so, whether it is uniquely defined. Third, are the locations assigned to regions contiguous? Fourth, what is the principle of internal organization within a region? Thus, a disjoint regionalization is found in the designation of the Standard Statistical Metropolitan Areas (which are not exhaustive, however, with respect to the total territory of the United States). Coverage by radio stations is exhaustive (though clearly not disjoint). The counties within a state are both disjoint and exhaustive. Each of the states is made up of contiguous subregions, while the system of states, or the United States, is not, because of Alaska and Hawaii (Bunge 1962; Teitz 1964; Tobler 1963; Whittlesey 1954).
There are at least two basic systems of internal regional organization. One system of defining a region is by partitioning the population into spatially designated subsets that are internally homogeneous and externally heterogeneous. This gives rise to the identification of uniform regions. Examples of this approach are the basic soil and climate regions of the United States and the landuse mappings of urban areas. Each delineated area has common internal characteristics that distinguish it from other areas. It should be added that the presumption is that socioeconomic variables of some consequence reflect this regional designation and partition. Further, such a region is an area where policies may be uniformly applied-a matter of considerable administrative interest. An alternative system of regionalization stresses internal cohesion or connection. This approach gives rise to the nodal-tributary region. It has been suggested that any one of numerous linkages or flows between populations specified by their location may be the basis for such a region.
The second regionalization principle is, perhaps, most clearly exemplified by the notion of a market area with characteristic flows to and from a central place. The set of consumers is spatially partitioned by the location of their suppliers. A set of cores and associated tributary areas results. The region is held together by the ties of interchange between purchaser and seller, and the regional boundary is the break point between the trading areas. At the same time, the set of nodal-tributary regions can form the framework for an administrative hierarchy. Here, too, there enters an element of homogeneity in the set of flow links that portray trade relationships (Berry & Garrison 1958; Berry & Pred 1961; Friedmann 1956; Garnsey 1956; Nystuen & Dacey 1961). [SeeCentral place.]
Regionalization is not the only approach to the problem of classification of phenomena over space. One alternative has been the development of descriptive models that are analogues of the Newtonian physics of masses. Thus, it is possible to describe a distribution in terms of potential of the particular mass (e.g., in terms of people or of income earned). Such a potential measure represents the total influence at each point in space of all mass components, as the influence is attenuated by distance. The values of the potentials at each location indicate each location’s proximity to other locations. This information is usually presented in map form. The usefulness of potential measures is that a number of socioeconomic variables have been found to correlate with the level of potential. An analytically related index of spatial structure is the population energy: the expected degree of spatial interaction between the masses at a pair of points is inversely proportional to the distance and directly proportional to the product of the population or other relevant masses at those locations. Again, certain flows have been found to behave according to this relationship. Socioeconomic analogues to force and gradient have also been developed. This social physics formulation is highly suggestive; however, it remains mainly a descriptive and predictive tool. The measures and relationships it emphasizes are too aggregative to reflect the individual elements (Isard et al. 1960, chapter 11).
In addition to efforts directed to construction of macrogeographic measures of spatial structure, there have recently been successes in development of a formal spatial statistics. This work has concentrated on formulating measures of centrality, dispersion, and correlation over space analogous to those in traditional statistics. Corresponding to the mean, median, mode, and geometric mean there are spatial central measures. Formulas for standard distance deviations have also been derived. These measures have been calculated for a number of points in time to show spatial shift of the population (Dacey & Tung 1962; Neft 1965; Warntz & Neft 1960). Order statistics represents a further development in the analysis of spatial patterns. Both quadrant and nearest neighbor techniques have been applied. The use of these techniques has made possible tests for nonrandomness that provide the analyst with a test for similarity between patterns (Dacey 1963). [SeeGeography, article onstatistical geography.]
The process of regionalization and the analysis of patterns of social phenomena have a purpose beyond construction of deductive models or precise description. A central purpose of regional science is to identify and analyze the problems of regions and to suggest solutions. Many of the regional science techniques have been adapted and developed for the purpose of making and implementing normative decisions by, for example, regional and metropolitan planning bodies.
The regional scientist characteristically attacks a region’s problems with a comprehensive or systems approach, which leads him to an interdisciplinary view of the individual, entrepreneur, and public body. Traditional economic reasoning and economic costing can neither adequately explain nor optimally guide location decisions. Considerable thought has gone into modeling decisions within a framework in which entrepreneurial profit or individual utility maximization plays a less dominant role than in classical economics. Specifically, social profit, public welfare, and noneconomic considerations tend to play a major role in model formulation.
The systems approach also emphasizes social and economic interdependences in which the behavior of one actor is seen to affect all other actors in the region and in which each actor has a characteristic and known relationship with others inside and outside the region. The concept of interdependence stresses not only the economic interrelations of men, firms, and organizations over space, but also their social and political interactions. Inquiry into these interdependences over space has proceeded at the highly abstract level of general equilibrium models as well as at the more operational levels of interregional input-output analyses and systems of interregional social accounts. Such operational models not only describe the structure of a system of regions but are particularly well suited to test and assess impacts of alternative public decisions and plans (Isard etal. 1960).
Research techniques and findings
In discussing the major research in regional science, we shall emphasize operational techniques, particularly those which allocate scarce resources, given available technology, to the production of wealth for individuals and organizations motivated primarily by the wish to maximize utility or profit. These techniques can also be applied by public bodies concerned with social welfare, present and future. The range of techniques may be suggested by reference once more to the three classes of actors or decision makers: entrepreneurs, individuals, and governments.
In accord with traditional economics, the entrepreneur is often seen as a decision maker who wants to maximize revenue less costs, where elements of both revenue and cost vary over space [see Firm, theory of the]. These calculations are, it is assumed, made for a known and fixed time period over which conditions do not change. The entrepreneur possesses complete information, and, in the formal model of the location decisions, he knows how to use it. The location decision essentially involves a calculated substitution among transportation and other factor costs (Hoover 1948, chapter 3; Isard 1956a, pp. 222-235).
For example, one may wish to predict the decision of an entrepreneur selecting a location for a new manufacturing plant. For this prediction, comparative cost analysis may be employed, on the assumption that this analysis is a reasonable representation of the entrepreneur’s decision process. The comparative cost approach consists of listing the locationally variable costs in serving a given market from a number of possible production sites and then picking the minimum cost point. A relatively simple extension of this model allows for variation in revenues as alternative production sites under consideration tap different markets. A further extension allows for the effects of finished product transport costs on the effective demand of consumers. The iron and steel, aluminum, shoe and leather, and synthetic fiber industries are among those whose locational structure has been studied intensively in this manner. A review of the comparative cost technique will be found in Isard et al. (1960), which includes an extensive bibliography of pre-1960 studies.
The costs of operation of any given activity may be closely tied to the scale and costs of operation of other spatially and technically related activities. Recognition of the significance of external economies due to spatial juxtaposition has led to the development of the industrial complex technique. Plausible complexes of industrial units are examined, using input-output analysis and the substitution principles of comparative cost analysis [see Input-output analysis]. A study of proposed oil-refining, petrochemical-fertilizer, synthetic-fiber complexes for the industrial development of Puerto Rico represents one example. Results of such research include identification of activity levels and locations optimal for the set of complexes studied. Industrial complex analysis can be used both to predict behavior of firms and to suggest policies to encourage industrial development. (For a discussion of this approach, together with a case study, see Isard, Schooler, & Vie-torisz 1959.)
Linear programming is one step beyond comparative cost and industrial complex analyses, which are restricted in the sense that they require preselection of output and shipment alternatives, which may well precondition results obtained. Linear programming, on the other hand, treats both the outputs at each location and shipments between locations as variables to be determined [see Programming]. The classical transportation model of linear programming has had numerous applications. Further developments in linear programming allow determination of optimal production and shipment patterns for multilocation and multistage firms. Finally, more general models of whole industries emphasizing the use of location-ally specified resources, including land itself, have been made operational. One interesting output of such studies is the generation of a set of location rents consistent with the optimal pattern of production shipments. These location rents provide a link between the linear programming models and classical location theory. Applications of the linear programming approach include Heady and Egbert (1962), Henderson (1957; 1958), Manne (1958), Marschak (1958), Miller (1963), Stevens (1961), and Vietorisz (1956).
The optimizing models that have just been described do not correctly depict the decision processes of entrepreneurs. The entrepreneur operates under constraints of imperfect knowledge; he also reacts to a host of noneconomic pressures and incentives. His calculus clearly extends to include elements of cost and revenue that the above formal models ignore; indeed, certain personal and institutional factors would appear to dominate in many location decisions. The divergence between optimal and actual behavior may also in part be explained by a secular trend toward homogeneity of locations in the United States and diminution of measurable differentials among locations. An increasing number of industries can be labeled “footloose,” and decisions about their location are made in response to institutional and other factors rather than strictly economic considerations. Economically nonoptimal behavior is not as severely penalized. Studies by McLaughlin and Robock (1949), Hoover and Vernon (1959), Perloff and his colleagues (1960), and Fuchs (1962) analyze the noneconomic factors influencing locational decisions.
Studies of household location decisions have focused on aggregates of individuals rather than on the location decisions of individuals. In contrast with the framework for studying entrepreneurial spatial decisions, where the unit of analysis is the individual firm, household behavior studies have been largely unable to identify any measurable criterion for optimal behavior or, for that matter, to explain manifested behavior at the level of the single household. One possible justification for dealing with individuals en masse is that individual behavior is assumed to be unpredictable in much the same way as the behavior of individual molecules. This assumption has led to development of the above-mentioned gravity and potential models of social physics and also to research studies of commuters, migrants, home purchasers, and shoppers. Such studies have typically analyzed behavior as a function of three factors: (1) distance, (2) socioeconomic characteristics of nodes, and (3) socioeconomic characteristics of the landscape between nodes. Thus, there has been a wide variety of studies verifying the hypothesis that interaction between a node and its tributary region (or between two nodes) decays with increasing distance. Further, the effective or perceived distance has been found to differ from physical distance, money-cost distance, and time distance because of the existence of intervening places, information flow patterns, and institutional effects. (See, for example, Carrothers 1956; Marble & Nystuen 1963; Stouffer I960; Warntz 1959; and Wingo 1961.)
The deductive models of mass behavior have either implicitly or explicitly been based on the assumption that interaction decreases with distance and increases with the number of opportunities. This, however, has led to two classes of similar models: first, those in which the physical analogy has been stressed, and, second, those in which the empirical utility of the models has been emphasized.
Certain normative approaches to mass behavior have been developed. Despite our previous statement that models of human behavior over space have been almost entirely oriented to mass probabilistic behavior, some attempts can be noted that ascribe optimizing behavior to the individual. A classic example is the so-called “traveling salesman” problem, in which the individual wishes to pick the shortest route that passes through each of a set of specific nodes (Dacey 1960). An example of the attempt to simulate household location decisions within an urban area has been developed in connection with a large-scale transportation study: each household is perceived to attempt to maximize its location advantage while competing with all other households for the available residential space (Herbert & Stevens 1960). This model is related to an abstract model in which households attempt to maximize their utility in the competition for urban space while landowners attempt to maximize returns from the land (Alonso 1960).
One of the characteristic problems faced by a government unit (say a city government, a planning organization for a metropolitan region, or a national development agency) is the anticipation, assessment, and adaptation to change brought on by major exogenous or outside impacts, for example, a reduction in military expenditures or an increase in population. Regional science techniques, particularly those based on recent developments in applied economics, are well suited to these tasks; application of the techniques by planning bodies is discussed at the theoretical level by Mitchell (1961), Davidoff and Reiner (1962), and Isard and Reiner (1962).
Such problems are best analyzed within a framework, or model, that relates the system’s (e.g., nation’s) demographic and economic variables to the regional demographic and economic variables of interest. The framework emphasizes the interdependence of the regions, and in practice it is constructed by synthesizing studies of various aspects of the system. Isard and his colleagues (1960, chapter 12) present in some detail various “channels of synthesis” and assess their operational significance.
The central feature of a typical framework is an interregional input-output matrix, which generates employment, population, migration, output, etc., by region when fed information specifying, by region, the final demands for each output. The basic findings of comparative cost and industrial complex studies can be incorporated into the framework by removing from the input-output matrix those industries or parts of industries whose location patterns are most meaningfully analyzed by use of these techniques. The results of these studies indicate specifically how certain industries and parts of industries are to be treated in specifying the final demands.
But comparative cost and industrial complex analyses require a set of basic assumptions for the system regarding, for example, birth rates, death rates, net immigration, technology, and tastes. From these assumptions, population, labor force, and average productivity of the system are projected for the key future year or years, and the osocial accounts (gross system product and gross system income) are then estimated. The social accounts are expressed in terms of levels of major aggregates, such as government expenditures and capital formation. System outputs by industrial sector are derived via input-output or other techniques. Initial regional markets are then established on the basis of current data and with the aid of relative growth charts, trend analysis, consumption expenditure studies, and ratios such as the coefficient of localization and location quotient (ibid., chapter 7). The final demands by region are then fed into the fused comparative cost-industrial complex-interregional input-output framework to obtain the values of the regional variables. However, because of discrepancies, especially between estimates of initial regional markets and the set of regional markets consistent with the results of the computation, it may be necessary to re-estimate initial regional markets and rerun the framework. The rerun process is continued until results are obtained that are in harmony with the regional market assumptions underlying the results and with resource scarcities and community attitudes.
In practice, studies involving the synthesis of several regional science studies are often restricted. Some pertain to a single region rather than to a system of regions; some involve fewer techniques and yield less disaggregated results. On the other hand, recent developments in general social, political, and economic equilibrium analysis for a set of regions, and in cooperative procedures for a set of regions, point to the possibility of still more extensive and comprehensive channels. Such systems are premised on a market containing social and political commodities as well as economic goods. Individuals function in a multiplicity of both active and passive roles; organizations engage in the production of economic and noneconomic goods, being motivated to maximize effective profits. Government units engage in production and distribution of programs of goods and services designed to maximize constituency welfare. Social norms and values are introduced and help govern actions chosen by the several behaving units. This broad social theory is designed not only to operate under the conventional price and market systems of economics but also with cooperative procedures developed from a fusion of game theory and normative concepts. The application of these cooperative procedures has been extended particularly to multi-region systems wherein elements of both harmony and conflict underlie development and investment programs (see Isard & Isard 1965; Isard & Smith 1966a; 1966b; 1967). Because of their ability to make consistent projections for a wide variety of variables of direct concern to public bodies, these channels of synthesis represent a direction of basic interest to central and regional planning agencies.
Relation to neighboring disciplines
The relation of regional science to economics is close, since many of the economic variables (income, employment, prices), techniques (input-output, cost analysis, social accounts), and assumptions (maximizing behavior, efficiency of production) have been absorbed. The regional scientist spells out the regional structure of an economy and at the same time seeks to add to traditional economic analysis the locational and spatial dimensions of reality. Regional science also criticizes economics for failing to recognize the role of governmental authority and administration in the shaping of regional economies. Finally, the regional scientist views the national economy itself as the sum of a set of regional economies. (An extensive review of regional economics, including a survey of regional science and its relation to economics, may be found in Meyer 1963; see also Fisher 1957.)
Regional science has in common with political science a concern with the proper areal distribution of power, the number of levels and the structure of power in a hierarchy of governmental levels in which metropolitan government is of special interest, and the varying spatial pattern of the cost of administration. It also shares an interest in the new frontiers in political science in the search for quantitative models that can be empirically established (Isard 1957).
Historically, the relation between regional science and sociology has been close, since many of those in the field have had strong roots in the type of regionalism developed by the sociologist Howard W. Odum and in regional sociology, human ecology, and urban sociology. The study of social structure is common to both disciplines, but the regional scientist is more insistent that such study be intimately linked to the analysis of economic structure and the pattern of cities and central places over space, and in a way that can be systematically formulated (Whitney 1957).
Concern with transportation systems, water resource use, agricultural development, and other problems has led to joint interests and studies with transportation engineers, hydrologists, agricultural economists, and other specialists.
It is frequently stated that regional science overlaps most with geography. Both are concerned with patterns of distribution, distances, and flows over space. However, while recognizing the need for comprehensive descriptive systems, for cartographic techniques and studies, and for the need to study physical phenomena together with the social phenomena which occur in space, regional scientists have also stressed the development of abstract model systems. Recent developments in theoretical and quantitative geography have made significant contributions related to the work of regional scientists (see, for example, Bunge 1962).
Finally, regional science has a close relation with city and regional planning. Many of the problems studied by regional scientists represent issues central to the planner, but there remain differences. The planner is much closer to the policy maker in his quest for immediate recommendations. He is more concerned with procedural and design steps and has less concern with development of abstract models.
Organization and profession
The major professional organization in the field is the Regional Science Association, which had a world-wide membership of about 2,600 in 1967, of whom 1,500 lived in the United States. The association, founded in 1954, has several regional sections in the United States. There are a number of national and language groups in Europe, Asia, Africa, and Latin America either in existence or in the process of formation.
The major work of the association is the organization of professional conferences, held annually in the United States and in Europe, and periodically in the Far East and in Latin America. Proceedings of the conferences are regularly published as volumes in the series Regional Science Association, Papers and Proceedings. In addition, the association cooperates in the distribution of a technical journal, Journal of Regional Science, and in the publication of a series of monographs.
Regional science is a subject of graduate study in a number of universities in the United States and abroad. A full-time degree program is offered at the University of Pennsylvania, and programs are currently being developed elsewhere.
Walter Isard and Thomas A. Reiner
[Directly related are the entriesArea; Central place; Planning, social, article onregional and urban planning; Region; Spatial economics.]
Alonso, William 1960 A Theory of the Urban Land Market. Regional Science Association, Papers and Proceedings 6:149-157.
Alonso, William 1964 Location Theory. Pages 78-106 in John Friedmann and William Alonso (editors), Regional Development and Planning. Cambridge, Mass.: M.I.T. Press.
Berry, Brian J. L.; and Garrison, W. L. 1958 Recent Developments of Central Place Theory. Regional Science Association, Papers and Proceedings 4.107-120.
Berry, Brian J. L.; and Pred, Allan 1961 Central Place Studies: A Bibliography of Theory and Applications. Bibliography Series, No. 1. Philadelphia: Regional Science Research Institute. → Reprinted in 1965, with additions through 1964.
Bunge, William W. 1962 Theoretical Geography. Lund Studies in Geography, Series C: General and Mathematical Geography, No. 1. Lund (Sweden): Gleerup.
Carrothers, Gerald A. P. 1956 An Historical Review of the Gravity and Potential Concepts of Human Interaction. Journal of the American Institute of Planners 22:94-102.
Dacey, Michael F. 1960 Selection of an Initial Solution for the Traveling-salesman Problem. Operations Research 8:133-134.
Dacey, Michael F. 1963 The Status of Pattern Analysis: Identification of Problems in the Statistical Analysis of Spatial Arrangement. Regional Science Research Institute, Discussion Paper Series, No. 3. Unpublished manuscript.
Dacey, Michael F.; and Tung, Tse-hsiung 1962 The Identification of Randomness in Point Patterns. Journal of Regional Science 4:83-96.
Davidoff, Paul; and Reiner, Thomas A. 1962 A Choice Theory of Planning. Journal of the American Institute of Planners 28:103-115.
Deutsch, Karl W.; and Isard, Walter 1961 A Note on a Generalized Con of Effective Distonce. Behavioral Science 6:308-311.
Fisher, Joseph L. 1957 Potential Contributions of Regional Science to the Field of Economics. Regional Science Association, Papers and Proceedings 3:17-23.
Friedmann, John R. P. 1956 The Concept of a Planning Region. Land Economics 32:1-13.
Fuchs, Victor R. 1962 Statistical Explanations of the Relative Shift of Manufacturing Among Regions of the United States. Regional Science Association, Papers and Proceedings 8:105-126.
Garnsey, Morris E. 1956 The Dimensions of Regional Science. Regional Science Association, Papers and Proceedings 2:27-39.
Harris, Britton 1959 Urbanization Policy in India. Regional Science Association, Papers and Proceedings 5:181-203.
Heady, Earl O.; and Egbert, Alvin C. 1962 Programming Models of Interdependence Among Agricultural Sectors and Spatial Allocation of Crop Production. Journal of Regional Science 4, no. 2:1-20.
Henderson, James M. 1957 The Utilization of Agricultural Land: A Regional Approach. Regional Science Association, Papers and Proceedings 3:99-114.
Henderson, James M. 1958 The Efficiency of the Coal Industry: An Application of Linear Programming. Cambridge, Mass.: Harvard Univ. Press.
Herbert, John D.; and Stevens, Benjamin H. 1960 A Model for the Distribution of Residential Activity in Urban Areas. Journal of Regional Science 2, no. 2:21-36.
Hoover, Edgar M. 1948 Location of Economic Activity. New York: McGraw-Hill.
Hoover, Edgar M.; and Vernon, Raymond 1959 Anatomy of a Metropolis: The Changing Distribution of People and Jobs Within the New York Metropolitan Region. Cambridge, Mass.: Harvard Univ. Press. → A paperback edition was published in 1962 by Double-day.
Isard, Walter 1956a Location and Space-economy: A General Theory Relating to Industrial Location, Market Areas, Trade and Urban Structure. Cambridge, Mass.: M.I.T. Press; New York: Wiley.
Isard, Walter 1956b Regional Science, the Concept of Region, and Regional Structure. Regional Science Association, Papers and Proceedings 2:13-26.
Isard, Walter 1957 Potential Contribution of Regional Science to the Field of Political Science. Regional Science Association, Papers and Proceedings 3:29-32.
Isard, Walter; and Isard, Peter 1965 General Social, Political and Economic Equilibrium for a System of Regions. Parts 1 and 2. Regional Science Association, Papers and Proceedings 14:1-33; 15:7-25.
Isard, Walter; and Reiner, Thomas A. 1962 Regional Science and Planning. Regional Science Association, Papers and Proceedings 8:1-35.
Isard, Walter; Schooler, Eugene W.; and Vietorisz, Thomas 1959 Industrial Complex Analysis and Regional Development: A Case Study …. Cambridge, Mass.: M.I.T. Press.
Isard, Walter; and Smith, T. E. 1966a On the Resolution of Conflict Among Regions of a System. Regional Science Association, Papers and Proceedings 17:19-46.
Isard, Walter; and Smith, T. E. 1966b A Practical Application of Game Theoretical Approaches to Arms Reduction (and to Goal Determinations Among Regional Planning Authorities). Peace Research Society, Papers 14:85-98.
Isard, Walter; and Smith, T. E. Iy67 location Games: With Application to Classic Location Problems. Unpublished manuscript. → To appear in Volume 19 of Regional Science Association, Papers and Proceedings.
Isard, Walter; and Tung, Tse-hsiung 1964 Selected Non-economic Commodities: Definitions and Speculations on Supply and Demand, Measurement, and Utility. Regional Science Association, Papers and Proceedings 13:71-91.
Isard, Walter et al. 1960 Methods of Regional Analysis: An Introduction to Regional Science. New York: Wiley; Cambridge, Mass.: M.I.T. Press.
McLaughlin, Glen E.; and Robock, Stefan 1949 Why Industry Moves South: A Study of Factors Influencing the Recent Location of Manufacturing Plants in the South. Washington: National Planning Association, Committee of the South.
Manne, Alan S. 1958 A Linear Programming Model of the U.S. Petroleum Refining Industry. Econometrica 26:67-106.
Marble, D. F.; and Nystuen, John D. 1963 An Approach to the Direct Measurement of Community Mean Information Fields. Regional Science Association, Papers and Proceedings 11:99-111.
Marschak, Thomas A. 1958 A Spatial Model of U.S. Petroleum Refining. RAND Research Memorandum No. RM-2205. Santa Monica, Calif.: RAND Corp.
Meyer, John R. 1963 Regional Economics: A Survey. American Economic Review 53:19-54.
Miller, Ronald E. 1963 Domestic Airline Efficiency: An Application of Linear Programming. Cambridge, Mass.: M.I.T. Press.
Mitchell, Robert B. 1961 On Paradigms and Paradiddles: A City Planner Looks at Regional Science. Regional Science Association, Papers and Proceedings 7:7-15.
Neft, David S. 1965 Statistical Analysis for Areal Distributions. Philadelphia: Regional Science Research Institute.
Nystuen, John D.; and Dacey, Michael F. 1961 A Graph Theory Interpretation of Nodal Regions. Regional Science Association, Papers and Proceedings 7:29-42.
Olsson, Gunnar 1965 Distance and Human Interaction: A Review and Bibliography. Regional Science Research Institute. Bibliographic Series, No. 2. Philadelphia: The Institute.
Perloff, Harvey S. et al. 1960 Regions, Resources and Economic Growth. Baltimore: Johns Hopkins Press.
Stevens, Benjamin H. 1961 Linear Programming and Location Rent. Journal of Regional Science 3, no. 2:15-26.
Stouffer, S. A. 1960 Intervening Opportunities and Competing Migrants. Journal of Regional Science 2, no. 1:1-20.
Teitz, Michael B. 1962 Regional Theory and Regional Models. Regional Science Association, Papers and Proceedings 9:35-50.
Teitz, Michael B. 1964 Regional Theory and Regional Models. Ph.D. dissertation, Univ. of Pennsylvania.
Tobler, Waldo R. 1963 Geographical Area and Map Projections. Geographical Review 53:59-78.
Vietorisz, Thomas 1956 Regional Programming Models and the Case Study of a Refinery-Petrochemical-Synthetic Fiber Industrial Complex for Puerto Rico. Ph.D. dissertation, M.I.T.
Warntz, William 1959 Toward a Geography of Price: A Study in Geo-econometrics. Philadelphia: Univ. of Pennsylvania Press.
Warntz, William; and Neft, David S. 1960 Contributions to a Statistical Methodology for Areal Distributions. Journal of Regional Science 2, no. 1:47-66.
Webber, Melvin M. 1964 Culture, Territory, and the Elastic Mile. Regional Science Association, Papers and Proceedings 13:59-69.
Whitney, Vincent H. 1957 Potential Contributions of Regional Science to the Field of Sociology. Regional Science Association, Papers and Proceedings 3:24-28.
Whittlesey, Derwent S. 1954 The Regional Concept and the Regional Method. Pages 19-68 in Preston E. James and Clarence F. Jones (editors), American Geography: Inventory and Prospects. Syracuse Univ. Press.
Wingo, Lowdon Jr. 1961 Transportation and Urban Land. Washington: Resources for the Future.
Zipf, George K. 1949 Human Behavior and the Principle of Least Effort: An Introduction to Human Ecology. Cambridge, Mass.: Addison-Wesley.