I 70-20,502 MORI ARTY, B a n y Martin, 1932L0CATI0NAL PREFERENCES AND THE PATTERN OF RESIDENTIAL CHANGE IN THE LANSING-EAST LANSING, MICHIGAN, METROPOLITAN AREA. Michigan State University, Ph.D., 1970 Sociology, regional and city planning U n iversity M icro film s, A XEROX Com pany , A n n A rb o r, M ich ig an zj Copyright by BARRY MARTIN MORIARTY 1970 LOCATIONAL PREFERENCES AND THE PATTERN OF RESIDENTIAL CHANGE IN THE LANSING - EAST LANSING, MICHIGAN, METROPOLITAN AREA By Barry Martin Moriarty A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Geography 1970 PLEASE NOTE: Some pages have small and indistinct type Filmed as received* University Microfilms ABSTRACT LOCATIONAL PREFERENCES AND THE PATTERN OF RESIDENTIAL CHANGE IN THE LANSING - EAST LANSING, MICHIGAN, METROPOLITAN AREA by Barry M. Moriarty The central theme of this study is the investigation of the de­ gree to which the urban residential spatial pattern is structured by the different locational preferences of residential decision makers. The fundamental purpose of the study is to identify the relevant locational preferences revealed by the spatial behavior of household location seekers and to determine a normative spatial allocation model utilizing these preferences to distribute total household growth to residential areas of the urban region. Two alternative propositions are found to explain the different spatial distribution of social class groups residing in the urban re­ gion; the "economic competition" and "social choice" hypotheses. The more commonly applied "economic competition" hypothesis proposes that the different distribution is due to differences in the budget costs and income resources of the different groups. The lesser known "social choice" hypothesis proposes that the different distribution is due to differences in the values, needs, and desires of the social class groups. Studies are cited to show that different social class groups possessing similar income resources have different residential location­ al patterns, and that groups possessing different income resources have Barry M. Moriarty similar travel costs - the major budget cost determining locational be­ havior according to the "economic competition" hypothesis. Additional studies are cited to show that residential decision makers tend to choose household locations which are accessible to their employment and close to households of individuals possessing equal or more prestigeous social status than their own. The spatial behavior of residential decision makers suggests that the multi-attributes of alternative household opportunities are evaluated against four preference structures. Of these, the most im­ portant in dimensioning locational choices, for the urban population as a whole, are social distance preferences and employment accessibility preferences. To a lesser extent residential locational choices are di­ mensioned by racial or ethnic preferences and life style preferences, the study of which is beyond the scope of this investigation. Employ­ ment accessibility preferences are associated with the "economic compe­ tition" hypothesis since the "journey to work" is the most important factor determining a location seeker's travel costs. Social distance preferences are associated with the "social choice" hypothesis because they define the character of the social interaction that location seek­ ers belonging to different social class groups are willing to under­ take with decision makers in their own or other social class groups possessing perceived sets of values, needs, and desires. The investi­ gation of whether employment accessibility preferences or social dis­ tance preferences are the major factor dimensioning locational choices is tested in the Lansing - East Lansing, Michigan, Metropolitan Region: an area in which the income resources of the upper, middle, and working Barry M. Moriarty class groups do not differ too substantially from each other. Two hypotheses are quantified and tested for the urban popula­ tion as a whole. One proposes the existence of a relationship between the accessibility of an area to total employment opportunities and the potential intensity of residential growth in the area. The other pro­ poses the existence of a relationship between the accessibility of an area to total social opportunities weighted by the objective social dis­ tance preferences of location seekers and the potential intensity of residential growth in the area. Tests of the hypotheses show that while a significant correlation exists between both accessibility indices and potential residential growth, the relationship associated with social accessibility is significantly higher. Given that the income resources of the different social class groups are similar, the results tend to support the general proposition that "social choice" is a more impor­ tant factor than "economic competition" in dimensioning household lo­ cational choices for the urban population as a whole. Both relationships are used as the basis for a residential land use model to distribute total household growth to areas of the urban re­ gion. Tests of the spatial allocation model, while not producing esti­ mates of sufficient reliability for practical use, show that by combin­ ing both accessibility indices significant estimates of the pattern of residential development are achieved. Significant estimates are also achieved by use of the social accessibility index alone, but not by use of the employment accessibility index alone. The spatial allocation model renders better estimates for urban areas in which decision makers belonging to the working class group predominantly reside than it does Barry M. Moriarty for middle and upper class areas. The significantly better estimates for working class areas are due to the similarity of attractiveness rankings calculated for the working class group with the attractive­ ness rankings calculated for the urban population as a whole. Methods of refining the basic spatial allocation model through a consideration of individual social class group locational preferen­ ces are outlined for future research when adequate time-series and smal1-area data become available. ACKNOWLEDGMENTS Whatever contribution this study makes to the understanding and modeling of the spatial distribution of residential land in the urban area is, in part, attributable to my wife, Lorraine Chouinard Moriarty, my mother, Patricia Barry Moriarty, and to my associations with many individuals at Michigan State University and elsewhere. Of those academic individuals one deserves particular recog­ nition - my dissertation advisor, Dr. Gerard Rushton. Professor Rushton has constantly inspired my interest in theoretical geography and introduced me to relevant concepts by scholars in other disci­ plines, as well as directing my efforts on this study. His interest in this study, his assistance and encouragement, and mainly his per­ sistence that this study should contribute more than it does has made my work highly profitable and enjoyable. The fact that the study falls short of its expectations is no fault of his. The assistance of other individuals is also gratefully ac­ knowledged. I thank Dr. Lawrence M. Sommers for making the facilities of the Department of Geography at Michigan State University available to me. Dr. B&ruch Boxer provided many original ideas about urban land use development, and the late Dr. Paul C. Morrison was a constant source of inspiration. Finally, for their willingness to listen and to offer construc­ tive criticism I thank Dr. James 0. Wheeler, and Professor Myles Bolan iii and Professor Charles W. Barr of the Department of Urban Planning at Michigan State University. I also thank Dr. Richard Duke, Thomas Borton, Paul Ray, and Stewart Marquis of the Environmental Simulation Laboratory at the University of Michigan; and Dr. Bart J. Epstein, Dr. Saul B. Cohen, and Dr. Howard F. Hirt, all former faculty members of the Department of Geography at Boston University where I first be­ came interested in Urban Geography. iv table of contents Page ACKNOWLEDGMENTS ................................................... iii LIST OF T A B L E S ................................................... viii LIST OF ILLUSTRATIONS............................................ ix Chapter I. I N T R O D U C T I O N ............................................ 1 Central Questions of the Investigation and the Study Design . . ...................................... General Concepts and a Frame of Reference ............. Urban Land Use T h e o r y ................................ Subjective Preference as a Behavioral Factor in Residential Choice ................................ Subjective Accessibility and Residential Growth . . . II. THE PROBLEM OF ALLOCATING HOUSEHOLDS TO RESIDENTIAL AREAS OF THE C I T Y ........................................ Nature of the P r o b l e m .................................. Spatial Interaction and Land Development .......... Locational Preferences and the Attributes of Residential Areas ................................ The Distribution of Residential Growth in the Study A r e a .......................................... Outline of the Relevant Literature ................... Approaches to the Study of Residential Spatial S t r u c t u r e ............................................ Description of Residential Spatial Structure .. . . Alternative Explanations of Residential Spatial S t r u c t u r e ............................................ The "Economic Competition" Hypothesis ............. The "Social Choice" Hypothesis ................... Inconsistency of Locational Behavior with the "Economic Competition" Hypothesis ................. Major Assumptions and D e f i n i t i o n s .................. . . Locational Preferences .............................. Occupation as a Cue to Socio-Economic Status . . . . Limitations of the I n v e s t i g a t i o n ..................... v 3 6 6 10 14 16 16 17 19 22 22 22 24 25 26 27 29 30 31 33 34 Chapter III. DESCRIPTION OF RESIDENTIAL DECISION MAKERS' LOCATIONAL PREFERENCES .................................. Residential Preferences Revealed by the Classical Models of Urban Spatial Structure ..................... Residential Preferences Revealed by Social Area Analysis S t u d i e s ............................ Residential Preferences Revealed by Factorial Ecology M e t h o d s ......................................... Residential Preferences Revealed by Consumer B e h a v i o r ........................................ Residential Choice Revealed by Social-Distance P r e f e r e n c e s ............................................. Summary ......................................... IV. V. VI. MODELING OF RESIDENTIAL DECISION MAKERS' LOCATIONAL P R E F E R E N C E S ............................................... Page 35 36 40 43 44 49 56 60 The Development of Household Allocation Models . . . . Household Allocation Models ............................ The Accessibility Model .............................. The Regression M o d e l ................................ Intervening Opportunity Models ..................... Interactance Household Allocation Models ........... The Efficiency of Household Allocation Models ......... S u m m a r y .................................................. 62 68 68 71 73 76 80 82 A MODEL FOR PREDICTING THE SPATIAL ALLOCATION OF HOUSEHOLDS TO RESIDENTIAL AREAS OF THE CITY ........... 85 Basic Premises and the Choice of Significant V a r i a b l e s ............................................... Variables Related to Vacant Land and the Real Estate Developer .................................... Variables Related to Land Use Competition Between ........ Non-residential and Residential Developers Variables Related to the Attractiveness of Areas for Residential Location Seekers ................... The Spatial Allocation Model .......................... Derivation of the M o d e l .............................. Assumptions of the M o d e l ........................... Input Data Used in the Model ................. Description of the M o d e l ........................... Limitations of the Proposed Model .................... 95 99 99 105 106 107 108 EVALUATION OF THE SPATIAL ALLOCATION M O D E L ............. 110 Analysis of the Relationship Between the Pattern of Residential Development and Accessibility to Social and Employment Opportunities .......................... Residential Growth .................................. Residential Decline .................................. 114 115 119 vi 87 88 93 Chapter Page Analysis of the Relationship Between the Pattern of Residential Development and the Predicted Household Assignment for the Urban Region .......................... Household Density Distribution Analysis ............... ..................................... Deviation Analysis Correlation Analysis ................................... Analysis of the Relationship Between the Pattern of Residential Development and the Predicted Household Assignment by Social Class Group ........................ Analysis of the Relationship Between Residential Area Rankings and the Locational Preferences of Social Class Groups ....................................... Social-Distance Preferences ............................ Employment Accessibility Preferences ................. S u m m a r y .................................................... 131 132 134 138 CONCLUSIONS AND P R O S P E C T S .............................. 141 Summary of the Principal Findings ........................ Future Research Opportunities ............................ 143 150 Socio-Economic Status Types (or Residential Decision Maker T y p e s ) .................................................. 153 VII. 121 123 125 125 128 APPENDIXES A B Basic Data Used in the Allocation Model by Residential A r e a .......... ................................ CComputer Program of the Spatial Allocation Model ............. SELECTED BIBLIOGRAPHY . . . . . . . . . . . vii ........................ 158 168 171 LIST OF TABLES Table 111.1 111.2 VI.1 V I .2 VI.3 VI.4 VI.5 V I .6 Page Reference Group Identification by Occupation Category of Decision Makers (Hypothesized Level of Acceptance or A v o i d a n c e ) ............................ 47 Per Cent Distribution of Occupational Status of Next-Door Neighbors by Occupational Status of Decision M a k e r s ......................................... 48 Relative Significance and Explanatory Power of Accessibility to Employment and Social Opportunities on the Pattern of Residential Development ............. 113 Percentage Deviation of Predicted from Actual Resi­ dential Development in 45 Areas for the Employment, Social, and Combined Household Allocation Models . . . 124 Relative Significance and Explanatory Power of the Employment, Social, and Combined Household Allocation M o d e l s ....................................... 124 Relative Significance and Explanatory Power of the Employment, Social, and Combined Household Allocation Models by Social Class of Areas ........... 129 Average Ranking of Residential Areas Calculated for Social Class Groups by Social Status of Area 133 . . . Ranked Total Employment, Total Social Accessibility and Social Accessibility for Upper, Middle, and Working Class Groups by Social Status of 45 Residential Areas ....................................... 135 Characteristics of Socio-Economic Status Types in the Lansing - East Lansing Area, 1960 ...................... 156 B.l 1960 Land Use I n v e n t o r y .................................. 158 B.2 Employment and Social Opportunity Inventory ............ 160 B.3 1960 - 1965 Dwelling Unit Inventory and Land Use Change . 162 B.4 Land Use Z o n i n g .......................................... 164 B.5 1965 Travel-Time Between Areas 166 A.l viii ......................... LIST OF ILLUSTRATIONS Figure , 1.1 II.1 111.1 111.2 Page Lansing - East Lansing Standard Metropolitan Statistical Area 4 Lansing - East Lansing Area Dwelling Unit Changes 1960 to 1965 ........................................... Preferred Mean Social Distance for Males by Self-Identified Social Class for Seventeen Occupa­ tion Categories in a Sample of Cambridge and Belmont, Massachusetts, Residents ................. 23 . 52 Occupation Category of Next-Door Male Neighbors for Social Groups in a Sample of Cambridge and Belmont, Massachusetts, Residents ........ . 54 Residential Association of Male Occupation Categories for Social Class Groups in Chicago, Illinois, Census Tracts - 1950 ........................ 54 111.4 Residential Segregation by Occupation Category ........ 57 IV.1 Gravity or Impedance Model Distribution of Trip Attenuations with Distance from i ................... 69 Opportunity Model Distribution of Trip Attenuations with Distance from i . 74 Empirical Distribution of Work Trips by Distance of Workplace for Different Occupation Categories . . . . 77 111.3 IV.2 IV.3 V.1 VI.1 V I .2 VI.3 Relation Between Accessibility and the Ratio of Actual to Expected Residential Development . . . . . . 102 Relationship Between Development Ratio and Accessibility to Social Opportunities for 35 Areas Having an Increase in Dwelling Units ........... 117 Relationship Between Development Ratio and Accessibility to Employment Opportunities for 35 Areas Having an Increase in Dwelling Units . . . . 117 Relationship Between Development Ratio and Accessibility to Social Opportunities for 10 Areas Having a Decrease in Dwelling Units ........... 120 ix Figure V I -4 VI.5 VI.6 V I .7 VI.8 C.l Page Relationship Between Development Ratio and Accessibility to Employment Opportunities for 10 Areas Having a Decrease in Dwelling Units ........ 120 Actual 1965 and 1960 Dwelling Unit Density with Distance from the CBD:Lansing - East Lansing Area 122 . Actual 1965 Dwelling Unit Density with Distance from the CBD Compared with the 1965 Dwelling Unit Density Forecast by the Employment, Social, and Combined Accessibility Models: Lansing - East Lansing A r e a ..................... 122 Actual and Predicted Dwelling Unit Change by the Combined and Social Accessibility Models, 1960-65: Lansing - East Lansing A r e a .......................... 126 Actual and Predicted Dwelling Unit Change by Employment Accessibility Model, 1960-65: Lansing East Lansing A r e a .................................... 127 Flow Chart of One Iteration of the Subjective Accessibility M o d e l ............ 168 x CHAPTER I INTRODUCTION There is noscience but tells a different tale, when viewed as a portion of a whole, from what it is likely to suggest when taken by itself, without the safeguard, as I may call it, of others . . . John Henry Newman Residential land is the main space consumer in the urban area. As such,most locational and travel behavior in the urban area is based upon people making decisions dealing with residential locations.^These locational decisions affect public and private capital expend!- } ture programs, which in turn influence the locational decisions of all other land users. Consequently, residential location decisions con­ stitute a major force in determining the spatial structure of cities. The impact of this force has resulted in different land use configura­ tion patterns through time because of adjustments in spatial behavior to space-transforming technological innovations and social changes. Spatial behavior has two dimensions: behavior. locational behavior and travel Locational behavior is the action of a decision maker in . ^Frank B. Curran and Joseph T. Stegmeier, "Traffic Patterns in 50 Cities," Public Roads - A Journal of Highway Research, Vol. 30, No. 5 (December, 1958), (Washington, D. C.: U.S. Department of Commerce, Bureau of Public Roads, 1958). ^William L.C. Wheaton, "Public and Private Agents of Change in Urban Expansion," Explorations into Urban Structure, edited by M.M. Webber, e_t. al. (Philadelphia: University of Pennsylvania Press, 1964), pp. 154-196. 1 choosing a household site from which he can facilitate his interaction with those spatially distributed opportunities deemed necessary by him to satisfy his preferred set of values, needs, and desires. Travel b e ­ havior is the action of a decision maker in making a trip to overcome separation from the different types of spatially distributed opportuni­ ties deemed necessary by him to satisfy his preferred set of values, needs, and desires. The spatial behavior of decision makers and the spatial pro­ cesses which give rise to city spatial structure are of vital concern to geographers and others interested in explaining the spatial order­ liness of urban land uses.** First, an understanding of the spatial b e ­ havior of residential decision makers contributes to the problem of ex­ plaining the areal distribution of households in the city. Secondly, an understanding of the orderliness of urban residential land as re­ vealed by the spatial behavior of residential decision makers makes a contribution to the task of constructing a general land use model of urban spatial structure.^ However, in order to understand the spatial behavior of individuals it is first necessary to understand the rele­ vant preferences by which decision makers evaluate alternative house­ hold locations. 3 Donald L. Foley, "An Approach to Metropolitan Spatial Struc­ ture," Explorations into Urban Structure, edited by M.M.Webber, e_t. al. (Philadelphia: University of Pennsylvania Press, 1964), pp. 21-78. 4 Knowledge of consistent patterns of spatial behavior on the part of residential decision makers not only contributes to the problem of explaining the areal distribution of households in the city but also contributes to the problem of explaining the urban population density distribution and the distribution of social class groups within dif­ ferent subareas of the city. 3 CENTRAL QUESTIONS OF THE INVESTIGATION AND THE STUDY DESIGN This study seeks to examine the problem of explaining the pattern of residential land development related to the Lansing - East Lansing, Michigan, Metropolitan Area from 1960 to 1965. Figure 1.1 shows the location of the nine township study area in the Lansing - East Lansing Standard Metropolitan Statistical Area. The issues around which the investigation is organized may be formulated in three major questions as follows: 1) On the macroscopic level what is a good way to describe the locational preferences associated with the residential spatial structure of American cities? 2) Given a particular description of the locational preferences and the residential spatial structure, how ade­ quate are existing household allocation models in utilizing these behavioral principles in explaining this structure? 3) What is the relationship, if any, between the pattern of residential land development and the hypothesized locational preferences of residential decision makers? In addition to attempting to formulate at least partial answers to these substantive questions, the efficiency of a procedure designed to reveal the form of spatial orderliness in urban residential loca­ tions from a consideration of locational preferences will be examined. Specifically, the second chapter will examine the problems in­ volved in explaining the pattern of residential growth and development giving particular consideration to the empirical and theoretical litera­ ture in which the problems have been discussed. The proposition that residential location decisions constitute a major force in determining the spatial structure of cities suggests the first objective of this Investigation: to provide a description of the nature and impact of this force on the spatial structure. This C L I N T O N COUNTY r EATON COUNTY [_ 1 N T O ■ S INGHAM COUNTY S ‘____ I nils* Q B L a n s in g - E a s t Lansing Michigan M etro p o lita n Arso FIG.I•I Lansing - East Lansing S t a n d a r d M s tr o p o lit a n Statistical Area description should determine the essential behavioral properties asso­ ciated with the growth and development of the residential spatial pat­ tern. These behavioral properties relate to the relevant preferences exhibited by individuals in choosing household locations from among al­ ternative opportunities. This part of the study will be accomplished through an investigation of the literature to empirically derive gen­ eralizations about the locational preferences of residential decision makers. Given this description of residential locational behavior, Chapter Four will examine the problem of whether or not existing house­ hold allocation models possess the properties that are consistent with the revealed behavior of individuals to produce the residential spatial pattern of cities.-* The chapter will examine the manner in which loca­ tional behavior of different socio-economic groups is included in the models . Chapter Five will determine a household allocation model based upon the derived principles of behavior and Chapter Six will evaluate its efficiency in revealing the form of spatial orderliness in residen­ tial locations in two ways. The first test of its efficiency will be through a comparison of the model's predicted allocation of households to subareas of the city between 1960 and 1965 to the actual residential development in the subareas for the same period of time. 5 The second Two basic approaches have been employed in the development of land use models: 1) trend models and 2) logic models. Trend models use empirical data, defining what events have occurred in the past, and then extrapolates this information into the future. Such models utilize the techniques of a) graphic extrapolation, b) multiple regression equa­ tions, or c) simultaneous equations. Logic models mathematically de­ fine how events occur and serve as a means to calculate what will happen in the future. Two basic approaches are used: a) linear programming and b) simulation. 6 test will examine the efficiency of the model in comparison to alterna­ tive household allocation models. Chapter Seven will review the results that may be drawn from this approach to explain the location and distribution of residential land uses and indicate areas of future empirical research deemed neces­ sary for the further development of a set of general rules of spatial behavior from which relevant spatial theory can be formulated. GENERAL CONCEPTS AND A FRAME OF REFERENCE Urban Land Use Theory There is abundant evidence that land uses of a given type are consistently situated at similar locations in urban space with respect to other land use locations to produce corresponding spatial patterns in American cities as they evolve over time.*’ Theories that account for the empirical regularity of urban land use locations have been advanced since, at least, Marshall's analysis of urban land values.^ Ernest W. Burgess, "The Growth of the City," The City, edited by Robert E. Park, Ernest W. Burgess, and Roderick D. McKensie (Chicago: University of Chicago Press. 1925), pp. 47-62: Homer Hoyt, The Struc­ ture and Growth of Residential Neighborhoods in American Cities (Wash­ ington, D. C.: Federal Housing Administration, 1939); Chauncy Harris and Edward L. Uliman, "The Nature of Cities," The Annals of the Ameri­ can Academy of Political and Social Science, Vol. 242 (November, 1945), pp. 7-17. ^Marshall is perhaps the first economist to concern himself with urban land use which he viewed as being dependent on land values that allocated the uses to different areas of the city. He emphasized the importance of locational situation within the urban place in determin­ ing land value and, also, considered the attributes of the site as being important. He argued that the value of a piece of land is the sum of its situation value plus its site value where the site value is equal to the agricultural value of the land. Although Marshall did not ex­ tend his analysis to residential land, he was among the first to con­ sider intensive and extensive land use in the urban: area. He stated that "If land is cheap (the land user) will take much of it; if it is dear he will take less and build high." (p. 448). Alfred Marshall, Principles of Economics (7th ed., London: MacMillan Company, 1916). 7 The theories, largely the contribution of land economists, offer a reasonably coherent, though highly generalized, account of the process by which urban land is allocated to produce the characteristic spatial structure of cities. At the present time no alternative theory exists that could provide serious intellectual competition to the economic ex­ planation of the land use allocation process.® In the allocation process land economists emphasize the relation­ ship between land use and land value and argue that the location of all land use is dependent on land values that allocate the uses to different areas of the city. Almost without exception since Marshall's analysis, land use theories have postulated that the unit area value associated with any location is primarily a function of the location's accessibil­ ity to all other locations.^ While the validity of this fundamental g Ira S . Lowry, Seven Models of Urban Development: A Structural Comparison (Santa Monica: The RAND Corporation, 1967), p. 9. 9 Land economists such as Hurd and Haig ignore the influence of the size of the site in determining urban land values, and their analysis closely resembles that of von Thunen; i.e., they emphasize the friction of distance concept. Hurd asserts that as a city grows, more remote in­ terior land must be utilized producing higher land rents at the more ac ­ cessible sites. Richard M. Hurd, Principles of City Land Values (New York: The Record and Guide, 1924), pp 11-12. In keeping with the mainstream of economic thought concerning urban land values Haig asserts that rent appears as the charge which the owner of a relatively accessi­ ble site can impose because of the savings in transport costs which the use of the site makes possible. Robert M. Haig, "Towards an Understand­ ing of the Metropolis,." Quarterly Journal of Economics, Vol. 40 (May, 1926), p. 421. Neither Hurd nor Haig give detailed attention to resi­ dential land use. Present day theories by Wingo and Alonso conclude that residential land use is a function of land rent which, in turn, is determined by the transportation cost from the central core of the city, the quantity of land consumed by the household and all other costs (con­ ceived as a composite good) incurred by the household. With some modi­ fication between their theories Wingo and Alonso point out that any par­ ticular household can substitute between these variables so as to maxi­ mize its satisfaction by owning and consuming the goods it prefers and avoiding those it does not prefer within the limiting constraints of its budget. Lowdon Wingo, Jr. , Transportation and Urban Land (Washington, 8 postulate Is not being questioned in this investigation, the adequacy of the postulate as presently conceived in terms of transport costs or physical distance from the central core has been found deficient particularly in explaining the location and distribution of residential land.10 The contention in this study is that while accessibility is a primary variable in explaining the use that an urban site can be put to, it is incorrect to assert that accessibility, measured in terms of transport costs or mere physical distance, is the only attribute of a location considered by decision makers in their choice of a residential D.C.: Resources for the Future, 1961); William Alonso, Location and Land U s e : Toward a General Theory of Land Rent (Cambridge: Harvard University Press, 1965). ^Residential land use theory is based on the assumption that indi­ viduals with limited income exercise a preference for central locations in order to reduce their transportation costs to the various activity units with which they wish to interact. People with higher income, on the other hand, are more able to absorb higher transportation costs and therefore exercise a preference for noncongested locations near the peri­ phery. However, Tomazinis has revealed that there is no systematic re­ lationship between the trip lengths of people and the distance they reside from the central core in Detroit and Chicago. Instead he found that people living intermediately between the core and the periphery have the shortest trip lengths while trips originating at the center and the periphery have increased trip lengths. Anthony Tomazinis, "Trans­ portation Inputs of Urban Activities: An Investigation of the Basic A s ­ sumptions of Urban Location Theories Concerning the Transportation In­ puts of Urban Activities" (unpublished Ph.D. dissertation, University of Pennsylvania, 1963) . Tomazinis and Gabbour found that in small cities (Lexington and Madison) as the residential distance from the central core increased the trip length increased, but that in larger cities hav­ ing numerous spatially distributed sub-centers (Pittsburgh and Phila­ delphia) or cities having great geographic variation in land use (Puget Sound) no systematic pattern between trip length and distance from the central core could be found. Anthony Tomazinis and Iskandar Gabbour, "Trip Length Variations Within Urban Areas," (Institute of Environ­ mental Studies, University of Pennsylvania, 1966); Iskandar Gabbour, "Travel Cost Variations and the Size of Urban Areas" (unpublished P h D , dissertation, University of Pennsylvania, 1967); Robert C. Brown, "The Use and Misuse of Distance Variables in Land Use Analysis," The Pro­ fessional Geographer, Vol. 20 (September, 1968), pp. 337-340; James A. Quinn, Human Ecology (New York: Prentice-Hall, 1950), p. 282. 9 site. Urban land economists themselves recognize this deficiency in explaining household locations but continue to use transport costs and physical distance as measures of accessibility.*-*- Among them Alonso states that, Since distance can be measured directly, while accessi­ bility implies some subjective pattern of preference (or nuisance value of distance) which may vary from one individual to another, it is better . . . to use distance as the variable.12 The inconclusiveness of many attempts to model the location of residen­ tial land is a result of basing the attempts upon such deficient propo­ sitions to account for the subjective pattern of preferences exhibited by individual decision makers in exercising their choice of a residen­ tial location from among the array of available alternative opportun­ ities they perceive. Hurd, Haig, and Ratcliff each support the contention that there is more to the problem of determining residential locations than economic considerations alone. Hurd states that ,fthe basis of residential land is social and not economic - even though the land goes to the highest bidder." Hurd, op. cit., p, 77. Haig claims that the "theoretically perfect site for an activity is that which furnishes the desired degree of accessibility at the lowest cost of friction." Extending this idea to household locations he points out that "in choos­ ing a residence purely as a consumption proposition one buys accessibil­ ity precisely as one buys clothes or food. He considers how much he wants the contacts furnished by the central location weighing the costs of friction involved - the various combinations of site rent, time value, and transport costs; he compares this want with his other desires and resources, and he fits his scale of consumption and he buys." Haig, ££. cit.. pp. 422-423 Hence, Haig views the problem of explaining residential location as one of a location seeker's quest for accessibility to other locations within the framework of his individual subjective preferences and resources. Ratcliff indicates the importance of subjective influences in determining urban land values He states, "It is the total effect of the competition for sites to minimize the aggregate of inconvenience and frictions as evaluated in terms of the local value system." Richard U Ratcliff, Urban Land Economics (New York: McGraw-Hill Book Co., 1949), p. 385. 12 Alonso, _oj>. cit ■, p 27. 10 Subjective Preference as a Behavioral Factor in Residential Choice Geographers have shown a concern for developing spatial models that are structured on more general variables than such purely economic factors as land costs and transportation costs. Casetti, Curry, Rushton, and Wolpert argue that spatial theory formulated in these economic terms is not relevant to explaining diverse spatial systems in differ­ ent periods in t i m e . ^ They claim that variables in terms of the sub­ jective preferences that individuals exhibit by their residential loca­ tional behavior, such as a preference for a site having a central loca­ tion or for an uncongested site, rather than the economic costs and values associated with a site provide a more meaningful approach to the problem of deriving universal rules of spatial b e h a v i o r . ^ This sug­ gests that decision makers decompose the over-all utility of residen­ tial sites into separate attributes, each of which contributes some part-worth to their evaluation of the sites. However, because of the subjective preferences of decision makers the separate attributes are weighted differently by the location seekers. Urban land use theory considers the situation of the site with respect to the central urban functions and the spaciousness of the site to be the relevant attri­ butes evaluated by decision makers in their choice of a residence with the situation attribute contributing the greater part-worth to the over­ all evaluation. Shepard suggests that residential locational behavior may be better explained if the part-worth attributed to the locational situation was "further analyzed into two part-worths: one attributable 1^ Emilio Cassetti, personal communication, Michigan State Univer­ sity, East Lansing, Michigan, December 3, 1968. ^ E m i l i o Cassetti, "Urban Population Density Patterns: An Alterna­ tive Explanation," Canadian Geographer, Vol. 11, 1967, pp. 96-100. 11 to desirability of neighborhood per se and another attributable to dis­ tance from work."*-'* Shepard's suggestion is in keeping with Park's postulate that "social relations are so frequently and so inevitably correlated with spatial relations. The postulate is supported by a creditable amount of sociological literature which reveals that residential locational behavior exhibits a preference for sites access­ ible to compatible neighbors who share the same values, needs, and de­ sires and that possibly this subjective preference overrides other preferences. Consequently, there exists a framework within which an alternative theory of the household allocation process can be formu­ lated. This framework considers subjective preferences that lie in­ side the economic land use allocation mechanism to be more fundamental to an explanation of residential locational behavior. Such preferences are synonymously termed locational or space preferences in this in­ vestigation . Isard describes space preferences in terms of the social and psychic properties which induce location seekers to aggregate or dis­ perse over space. Marble suggests that the incorrectly predicted portion of household locational behavior can be explained in terms of variables representing space preferences that are exogenous to the ^ R . N . Shepard, "On Subjectively Optimum Selections Among MultiAttribute Alternatives," Decision Making, edited by Ward Edwards and Amos Tversky (Harmondsworth, England: Penguin Books Ltd., 1967), pp. 257-283. ^■^Robert E. Park, "The Urban Community as a Spatial Pattern and a Moral Order," The Urban Community, edited by Ernest W. Burgess (Chicago: University of Chicago Press, 1926), p. 18. *^Walter Isard, Location and Space Economy (Cambridge: Press, 1958), pp. 83-88. The MIT 12 spatial system under investigation and reflect the social and psycho­ logical characteristics of the location seekers irrespective of the spatial s y s t e m . R u s h t o n views space preferences in terms of a given level of social contact on the part o^ decision makers which are main­ tained regardless of their spatial situation.^ Social distance, the key subjective space preference in terms of which this analysis is formulated, is postulated to be such a social and psychological characteristic of residential location seekers. Social distance is defined as a subjective "attitude of ego toward a person (or group) with a particular status attribute (in this investi­ gation, occupation) on the part of a decision maker that broadly de­ fines the character of the interaction that the decision maker would be willing to undertake with the attitude object." 20 Social distance, while not possessing the metric characteristics in the rigorous sense that physical distance does, tends to order social interaction oppor­ tunities along an ordered continuum of the relative ease of access to the social opportunities ranging from a close intimacy to a distant probability that interaction will occur. The concept provides a measure of the worth that location seekers belonging to different socio-economic groups assign to social interaction opportunities regardless of their location in urban space. Since some urban 18Duane F. Marble, "Transport Inputs at Urban Residential Sites," Papers and Proceedings of the Regional Science Association, Vol. 5, 1959, p. 254. l^Gerard Rushton, "Spatial Pattern of Grocery Purchases by the Iowa Rural Population" (Studies in Business and Economics, Iowa City: University of Iowa, 1966), p. 64. Edward C. Laumann, Prestige and Association in an American Com­ munity: An Analysis of the Urban Stratification System (Indianapolis: The Bobbs-Merrill Co., Inc., 1966), p. 4. 13 neighborhoods provide greater opportunities for social interaction be­ cause of the status characteristics of the decision makers residing there, these neighborhoods will be preferred over others. Although space preferences are conceived to be exogenous to any particular spatial system, they help to determine the spatial behavior of decision makers when placed in the context of a given set of inter­ action opportunities. Typically an individual's opportunities for interaction are restricted to the limited environment which he per­ ceives and with which he has contact. The perceived environment within which the decision maker selects to move is his "opportunity space" or "action s p a c e . A decision maker's "action space" does not include all of the available opportunities in the urban space with which he could interact, and therefore, an individual's "action space" is a mental image of the real urban space. The "action space" consists of a number of spatially located opportunities from among which the decision maker might choose to interact. Since some of the spatially distributed opportunities he perceives are preferred over others, the decision maker assigns some measure of utility to all conceivable op­ portunities and ranks them according to his particular set of prefer­ ences. Social distance, then, provides a means of expressing the differential spatial behavior exhibited by individuals interacting with alternative spatially located social opportunities according to their ranked preference to interact or not to interact with decision ^ J u l i a n Wolpert, "Behavioral Aspects of the Decision to Migrate," Regional Science Association Papers, Vol. 15, 1965, pp. 159-172; Frank E. Horton and David R. Reynolds, "Urban Environmental Perception and Individual Travel Behavior" (Department of Geography, Special Publica­ tion No. 2, Iowa City: University of Iowa, 1968); Lawrence A. Brown and Eric G. Moore, "Intra-Urban Migration: An Actor Oriented Framework" (unpublished paper, University of Iowa, 1968). 14 makers possessing particular status attributes. Subjective Accessibility and Residential Growth The social distance preference ranking must be placed within the context of a given set of spatially distributed opportunities before it can affect the spatial behavior of location seekers. Presently there is only one method devised to orient decision makers to all spatially distributed opportunities. The method, termed subjective accessibility, provides a way of implementing the social distance rankings of location seekers. It is a formulation related to Stewart's index of accessibility and is a measure of the spatial dis­ tribution of different types of interaction opportunities about a location adjusted for the ability and desire of decision makers to overcome spatial separation. o2 The subjective accessibility measure ranks residential areas of the city in terms of their attractiveness to location seekers belong­ ing to different socio-economic groups. Areas of high attractiveness are those which provide a given group of location seekers the best chance to facilitate their interaction with those spatially distri­ buted opportunities deemed necessary by them to satisfy their preferred set of values, needs, and desires. Consequently, while some residential areas may have high subjective accessibility indices for some groups of location seekers, they may possess relatively lower indices for other groups. 22 This definition was advanced by Hansen. Walter G. Hansen, "How Accessibility Shapes Land Use," Journal of the American Institute of Planners. Vol. 25 (May, 1959), pp. 73-81. However, Hansen's measure­ ment in no way takes into account the desire of individuals to over­ come spatial separation. 15 Subjective accessibility differs from other indices of accessibil­ ity in that it measures the locational situation of a residential area with respect to: 1) its orientation to employment opportunities in the location seeker's occupation category and 2) its orientation to other decision makers with whom location seekers in particular occupa­ tion categories reveal a preference to confine their social relation­ ships . This investigation proposes that a significant relationship exists between a residential area's indices of subjective accessibil­ ity and the intensity of residential construction in the area. This general proposition expressed in terms of variables associated with the spatial behavioral characteristics of location seekers belonging to dif­ ferent socio-economic groups can be modeled in a probabilistic manner to generate a pattern of residential growth and development. A compar­ ison of the predicted land use pattern with the actual pattern can be made to test their degree of correspondence. The degree of corres­ pondence provides a means of ascertaining the validity of the loca­ tional preferences hypothesized as being relevant to the process of residential land use allocation. Close correspondence between the predicted and observed pattern would allow for changes in the magnitude of the variables on the bases of planned changes in order to ascertain the residential pattern under different policy considerations, under different social and economic conditions, and under different spacetransforming innovations. CHAPTER II THE PROBLEM OF ALLOCATING HOUSEHOLDS TO RESIDENTIAL AREAS OF THE CITY NATURE OF THE PROBLEM A problem exists in predicting the spatial allocation of house­ holds to different areas of the city because each residential location seeker can choose from among alternative opportunities, each of which possesses a number of subjectively disparate attributes related to its site and situation.^ To date a suitable method of quantifying these attributes or of expressing how they interact with one another to pre­ dict future household locations has not been determined. A major reason for this failure is that the process by which residential loca­ tion seekers weigh and combine or "trade off" the component attributes among the alternatives to arrive at household choices is not understood. Presumably, they evaluate the attributes associated with the alternatives Shepard, ££. cit., p. 257. Shepard points out that decision makers experience little difficulty in evaluating alternative oppor­ tunities with respect to any one component attribute by the process of elementary comparison. The difficulty occurs when evaluating a large number of component attributes (p. 257). The results of experiments by Shepard and others indicate that decision makers exhibit a consistent tendency to overcome this difficulty by making decisions "on the basis of one or two factors, while in effect, ignoring the significant con­ tribution of other factors." Consequently, "in making an evaluative judgment a subject can take account of only a very limited number of factors at any one time." (p. 266) This tendency toward factor reduc­ tion provides a basis for the determination of relevant variables considered by household location seekers in choosing residential sites. 16 17 against some unknown set of preferences. A major purpose of this in­ vestigation is to derive a set of common attributes revealed by the spatial behavior of residential decision makers and examine their relevance to the household allocation process. While the lack of success in predicting the spatial allocation of households seems to deny uniformity in the pattern of residential locational behavior, the actual urban land use pattern suggests that the behavior does, in fact, conform to some degree of order. There is ample evidence that land uses of a given type are situated at similar locations in space with respect to other land use locations so that corresponding spatial patterns are found in cities as they evolve over time. Certainly this orderly distribution of land use implies an or ­ derliness in the pattern of locational behavior on the part of location seekers- Confirmation and discussion of these patterns have been the subject of studies dealing with land use theory, social area analysis, tripmaking behavior, and social physics.^ Spatial Interaction and Land Development To explain the orderly pattern of land use, theorists postulate that the type and intensity of land use at a location is determined by, 2 See, for example: Harland Bartholomew, Land Uses in American Cities (Cambridge: Harvard University Press, 1955); Burgess, o£. ci t .; Hoyt, o£. c i t .; Harris and Ullman, o£. c i t .; Albert Z. Guttenberg, "Urban Structure and Urban Growth," Journal of the American Institute of Planners, Vol. 26 (May, 1960), pp. 104-110; Robert B. Mitchell and Chester Rapkin, Urban Traffic: A Function of Land Use (New York: Columbia University Press, 1954); Wingo, o£. c i t .; Alonso, 0£. c i t .; Theodore Anderson and Janice Egeland, "Spatial Aspects of Social Area Analysis," American Sociological Review, Vol. 26 (June, 1961), pp. 392398; Eshref Shevky and Wendell Bell, Social Area Analysis (Stanford, Calif.: Stanford University Press, 1955); John Q. Stewart and William Warntz, "Macrogeography and Social Science," Geographical Review, Vol.. 48 (April, 1958), pp. 167-184. 18 and in turn determines, the type and intensity of spatial interaction at that location. Quinn, for example, observes that the greater the frequency of interaction between a land use activity and complementary land use activities, the greater will be the tendency to maximize its accessibility to the complementary activities. 3 Hawley observes that land use activities having the largest accessibility requirements tend to occupy central locations while all other land use activities dis­ tribute themselves about central locations with distance away propor­ tional to their accessibility requirements.^ Wingo observes that the orderly spatial pattern is determined by the land market process of competition among land users for the most accessible site that will allow them a realization of profits or satisfactions over costs.^ support of these empirical generalizations, In land theorists propose the existence of a relationship between the accessibility of a location and the intensity of land use at that location so that the more accessible an area is to the various land use activities the greater will be its development potential.^ However, while the locational behavior of non-residential de­ cision makers tends to conform to this proposition, its presumed law­ fulness is not readily distinguished in the locational behavior of residential decision makers. Studies concerned with both the household o James Quinn, o p . cit., p. 286. 4 Amos Hawley, Human Ecology (New York: Chapter 12. ^Wingo, o£. c i t ., p. 26. ^Hansen, o£. c i t ., p. 73. Ronald Press, 1950), 19 and traffic assignment problems have found that residential decision makers exhibit geographical bias In both their locational and travel behavior which conflicts with the prescribed order.^ This inconsist­ ency results from the fact that even though no two residential decision makers exhibit the same spatial behavior, in the sense that each oc­ cupies different household locations from which he chooses to interact with different spatially distributed activities, many location theorists ascribe to each the same accessibility preferences they ascribe to nonresidential location seekers. However, the complementary activities that non-residential decision makers seek to be accessible to (that is, their consumers) are not the same complementary activities to which residential decision makers seek to increase their accessibility. Residential decision makers seek to be accessible to those activities deemed necessary by them to satisfy their day to day needs. The unique feature of the household allocation problem is that, not only do residential location seekers' accessibility preferences differ from those of non-residential location seekers, but they differ according to the socio-economic characteristics of the residential location seekers themselves. Locational Preferences and the Attributes of Residential Areas So far attempts to model these differences to arrive at a real­ istic assignment of residential growth to various areas of the city have been unsuccessful. One reason the attempts have failed is because 7 Richard J. Bouchard and Clyde E. Pyers, "Use of Gravity Model for Describing Urban Travel," Highway Research R e c o r d , No. 8 8 : Traffic Patterns (Washington: Highway Research Board, 1965), pp. 2-3, p. 32; Walter G. Hansen, "Accessibility and Residential Growth" (unpublished MCP thesis, Massachusetts Institute of Technology, Cambridge, 1959), p. 1. 20 of the problem of not being able to classify location seekers into groups exhibiting similar spatial preferences. A more important reason for the failure is that undue emphasis is placed on the situ­ ation attribute of residential areas, and the site attributes con­ tained in the models have proven to be insignificant to the household locational choice process. Certainly the determination of attributes found to be shared in common by most location seekers in their evalu­ ation of residential areas would contribute to the accuracy of house­ hold allocation models. This investigation contends that such common attributes are to be found in the tendency of residential location seekers to choose housing in compatible neighborhoods located where spatial separation from the activities in which they frequently par­ ticipate is at a minimum. If it was possible to find a large number of location seekers who exhibited similar locational preferences (such as a preference to be accessible to the central urban functions), then it would be possible to compare their household locational choices with those exhibiting different preferences (such as a preference to be accessible to com­ patible friends and neighbors). It would then be possible to make in­ ferences concerning the probability of the location seekers choosing alternative housing opportunities in areas possessing the combinations of these attributes that would attract them there. Total residential growth can also be distributed on the basis of these probabilities. Implied here is that the accuracy of the predicted allocation serves to test the significance of the attributes hypothesized as being relevant to the residential decision maker's choice of a household location. Nearly without exception household allocation models are based on 21 the proposition that residential decision makers substitute journey-towork expenditures for household site expenditures with the rate of substitution depending upon the location seeker's preference for "more o spacious living conditions," housing. Q "low density housing," or "better This investigation argues that the attributes of resi­ dential areas expressed in these terms have less significance to the location seeker's evaluation of alternative opportunities since the attributes have failed to account for the distribution of households in the city. Unlike models that transform the attributes of areas into transportation and household site expenditures, the present investi­ gation seeks to determine attributes associated with the social aspects of the areas and to weight these attributes by the revealed preferences of location seekers to generate a distribution of households to resi­ dential areas of the city. The proposition underlying this study holds that residential decision makers substitute accessibility to work for household sites with the direction of the substitution depending on the location seeker's preference to reside close to the households of compatible neighbors. g Lowdon Wingo, Jr., "An Economic Model of the Utilization of Urban Land for Residential Purposes," Papers and Proceedings of the Regional Science Association, Vol. 7, 1961, pp. 191-205; William Alonso, "A Theory of the Urban Land Market," Papers and Proceedings of the Regional Science Association, Vol. 6, 1960, pp. 149-158. 9 Cassetti, o£. c i t ., pp. 96-100. 10 Richard F. Muth, "The Spatial Structure of the Housing Market," Papers and Proceedings of the Regional Science Association, Vol. 7, 1961, pp. 207-220. 22 The Distribution of Residential Growth in the Study Area Figure II.1 illustrates the location and net change in the number of household units in the Lansing - East Lansing, Michigan, area be­ tween 1960 and 1965. Some areas, particularly those close to the central core, have declined in the number of dwelling units while other areas, more distant from the central core, have experienced an increase. A distinct directional bias is evident. The figure illustrates the ability of higher-ordered land uses (or those having high rent paying ability) to pre-empt accessible locations from lower-ordered uses and the importance of vacant land, or land at a lower use, to attract dwelling units. The problem with which this study is concerned is to determine site and situation attributes considered significant to the location seeker's evaluation of alternative household opportunities and formulate them into a model to simulate this pattern. The accuracy of this allocation mechanism in replicating the distribution in comparison with other models serves to test both the relevancy of the attributes to the residential decision making process and the preference structure by which the attributes are weighted. OUTLINE OF THE RELEVANT LITERATURE Approaches to the Study of Residential Spatial Structure In recent years two main avenues of research have characterized studies of residential spatial structure: 1) determination of the location and distribution of different types of residential areas de­ limited by the housing quality and the socio-economic characteristics of their inhabitants, and 2) development of general (macro) models designed to generate patterns of residential land use. The first group of studies have developed empirical instruments, by analytical 10 DwtUint anil* conttrwctM CS Dvalllfti anil* 4«a«M tM 4 N> U> FIG. Il l Loosing - E a s t Looting D w elling unil c tia n g s s I960 to 1969 Arsa 24 methods, that would be useful in describing different residential areas and in explaining the orderliness of the residential spatial pattern. The second group of studies have attempted to duplicate residential growth patterns, usually in terms of the land market mechanism whereby urban space is allocated to the most efficient user. Because of two limitations neither of these approaches has been successful in de­ termining the residential allocation process from which relevant spatial theory can be formulated. The first limitation is that the majority of studies ascribe the decision making behavior associated with an indi­ vidual non-residential location seeker to the collective behavior of all location seekers, both non-residential and residential. The result is that the residential decision maker's locational behavior is com­ pletely determined by the land market rather than merely constrained by the market. The second limitation is that the approaches tend to ignore that the residential location seeker's locational behavior, like any form of behavior, is a response conditioned by a set of preferences which grow out of past experience. The result is that the social goals and social processes underlying residential decision making behavior are not treated. While the approaches provide a great deal of insight into the residential allocation process, they lack the consideration that household choice is an individual matter insofar as it is the in­ dividual decision maker, not the aggregate, that is motivated (or not motivated) to reside in a particular area. Description of Residential Spatial Structure A basis for understanding the relevant preferences by which loca­ tion seekers make decisions in choosing from alternative residential opportunities was set forth by Park and Burgess as early as 1926 when 25 they observed that residential decision makers exhibited a tendency toward choosing households in areas containing groups belonging to their same social class. Park discovered, and subsequent investigations vali­ date, that the greater the similarity of the socio-economic character­ istics of decision makers the more intimately will they be related socially and the less will be the physical distance separating them residentially. Conversely, the less similar are the socio-economic characteristics the greater will be the physical distance separating their residences.^ At the same time Burgess revealed that the lower socio-economic groups tended to occupy centrally located sites while upper socio-economic groups were more conspicuously located near the periphery of the c i t y , ^ Some years later Hoyt qualified this observa­ tion by asserting that the actual distribution of socio-economic classes was confined to different sectors of the city although there was a general tendency for the groups to be distributed as Burgess had revealed.Hence, ecologists state that the city is divided into "natural social areas," each of which is made up of nearly-similar ethnic, racial, or social status groups living in dwelling units of similar quality. Alternative Explanations of Residential Spatial Structure There exists two alternative postulates to explain the process by which socio-economic equality is transformed into residential proximity ^ R o b e r t E. Park, Human Communities (Glencoe, 111.: Free Press, 1952), p. 231; see also Park, "The Urban Community as a Spatial Pattern and as a Moral Order," p. 18. 12 13 Burgess, o£. c i t ., Chapter 2. Hoyt, o£. c i t ., p. 23 and Chapter 6. 26 to achieve "natural social areas." Feldman and Tilly term these postulates the "economic competition" hypothesis and the "social choice" hypothesis.^ Present-day land use allocation models confine their attention to the purely economic postulate. The "Economic Competition" Hypothesis Advocates of the "economic competition" hypothesis, such as Wingo, Alonso, and Lowry, view the spatial distribution of all land uses as being governed by the impersonal process of competition for locations with fixed differential v a l u e s . ^ In this competition all land users, both residential and non-residential, possess similar preferences (because they share the same goals) but differ in their budget costs and available resources. The hypothesis stresses that ability to pay is the basic factor in explaining the distance of all land uses from the central core, and the quality, size, and location of households that different location seekers choose to occupy. Changes in the resi­ dential pattern occur when market competition forces the most access­ ible land, near the central core, out of residential use. At the same time transportation improvements can reduce the costs associated with peripheral locations so that even though the decision maker experiences higher travel costs, he realizes more spacious living conditions at a lower cost. Transportation and site costs are deemed to be the major determining factors considered by decision makers in their choice of a household so that at designated, market-determined prices a location 14 Arnold S. Feldman and Charles Tilly, "The Interaction of Social and Physical Space," American Sociological Review, Vol. 25 (December, 1960), p. 877. ^\,owry, o £ . cit. , pp. 5-13; Alonso, "A Theory of the Urban Land Market"; Wingo, Transportation and Urban L a n d , pp. 64-65. 27 seeker would be indifferent to his location. Residential decision makers, although they may possess similar housing preferences, compete for the locations of their choice on the basis of their financial re­ sources. Consequently, advocates of this hypothesis propose that the differences in the locational behavior of residential decision makers are due to differences in their available resources and that it is this difference that determines the spatial distribution of socio-economic groups in urban space. The "Social Choice" Hypothesis A substantial body of literature supports the "social choice" hypothesis which argues that residential decision makers choose to locate their households on the basis of more fundamental behavioral propositions than economic costs alone. Garrison and others point out that: from one urban site to another, differences in travel costs may be relatively small compared to differences in the costs of sites occasioned by other considera­ tions. A variety of neighborhood amenities plays a large role in residential site selection as do prestige, discrimination and other factors that may override loca­ tional travel questions.16 In support of this yiew numerous studies claim that residential loca­ tion seekers are not indifferent to their locations but are greatly in­ fluenced in their housing choices by the character of established residential areas. Advocates of the "social choice" hypothesis, such as Firey, the Duncans, and Feldman and Tilly, assert that the spatial distribution of land uses is an indicator of social values, governed in part by l^William L. Garrison, "Difficult Decisions in Land Use Model Construction," Highway Research Record No. 126: Land Use Forecasting Concepts (Washington: Highway Research Board, 1966), p. 21. 28 sentimental, non-economic, or cultural facto r s . ^ The hypothesis stresses that the selection of a household involves the conscious social choice of residential decision makers who vary in their preferences because of different values, needs, and desires. These choices tend to create relatively homogeneous "natural social areas" containing similar types of people because they seek to increase their accessibility to compatible neighbors who share the same values, needs, and desires. To explain this phenomenon Park proposed that household locational choices from alternative opportunities are a function of the social-distance preferences of decision makers toward different socio­ economic groups.*-® To date the use of the social-distance concept has not been attempted in a household allocation model. This investiga­ tion attempts to use this concept to explain the household assignment problem. Walter Firey, Land Uses in Central Boston (Cambridge: Harvard University Press, 1947); Walter Firey, "Sentiment and Symbolism as Ecological Variables." American Sociological Review, Vol. 10 (April, 1945), pp. 140-148; Otis Dudley Duncan and Beverly Duncan, "Resi­ dential Distribution and Occupational Stratification," The American Journal of Sociology. Vol. 60 (March, 1955),pp. 493-503; Feldman and Tilly, oj>. c i t . , pp. 877-884. *-®Park, "The Urban Community as a Spatial Pattern and as a Moral Order," p. 18. ^ C h a n g e s in the residential pattern of the city take place when changes occur either in: 1) the distribution of opportunities decision makers choose to interact with to meet their day to day needs or 2) the social structure of the city's inhabitants. The change in the dis­ tribution of opportunities is largely a response to changing technology, particularly space-transforming technology. Changes in social structure occur as a result of in-migration and social mobility, both ofwhich can greatly affect the proportion of decision makers belonging to any one socio-economic group. In this case even if the distribution of opportunities was to remain the same they would be evaluated different­ ly because the values, needs, and desires, and hence the locational preferences of the decision makers, would have changed in keeping with 29 Inconsistency of Locational Behavior with the "Economic Competition" Hypothesis While the "economic competition" hypothesis holds that It is the financial resources of individuals that determine the locational preferences of decision makers (in that as income increases the quanti­ ty of land they consume increases), the findings of the Duncans and Feldman and Tilly do not support this proposition. They found that blue collar workers and white collar workers at the same income levels have different residential patterns and by inference different loca­ tional preferences not explained by the "economic competition" hypothe­ sis.^ The apparent inconsistency of the "economic competition" hypo­ thesis with the locational behavior of residential decision makers has led investigators to consider more carefully the social goals of loca­ tion seekers in their choice of a household location.^ This approach to the problem of allocating households to residential areas of the their social status. The social distance concept used as a locational preference is a means of expressing this different evaluation. See: Barry M. Moriarty, "Human Iried paper presented at the Annual Meeting of the Association of American Geographers, Washington, D. C . , August, 1968), Abstract in Annals of the Association of American Geographers, Vol. 59 (March, 1969), p. 195. 20 tuDuncan and Duncan, o£. cit . ; Feldman and Tilly, o£. c i t . ^ W olpert has suggested that locational choice is conditioned by preferences which are a function of the family life style of the decision maker. See: Wolpert, o£. c i t ., pp. 159-172. Berry has in­ dicated the importance of social group intimacy which suggests social distance preferences. See: Brian J.L. Berry, "The Factorial Ecology of Calcutta," The American Journal of Sociology, Vol. 74 (March, 1969), pp. 445-491. Chapin has investigated the importance of frequently engaged in activities as factors related to the choices of location seekers. See: F. Stuart Chapin, Jr. and Henry C. Hightower, Household Activity Systems - A Pilot Investigation (Institute for Social Science Research, Chapel Hill: University of North Carolina Press, 1966); F. Stuart Chapin, Jr., Urban Land Use Planning (second edition; Urbana: University of Illinois Press, 1966), pp. 221-253. 30 city is adopted in this investigation. The nature of the problem dictates that the locational preferences revealed by the spatial behavior of different socio-economic classes must be measured and the relationship which exists between them and the growth and structure of residential areas examined. Specifically the study is concerned with understanding and quantifying locational behavioral patterns with respect to: 1) accessibility and social-dis­ tance preferences which may condition the behavior; 2) the available opportunities that exist in the urban space that may satisfy these preferences; and 3) the manner in which the site and situation attri­ butes associated with the opportunities may be weighed and combined or "traded off" in the light of these preferences. This information can then be used to estimate the allocation of households to residential areas and the hypothesis that a relationship exists between the sub­ jective accessibility of an area and the intensity of residential growth in that area may be tested. Such an approach would give cre­ dence to either the "economic competition" or "social choice" ex­ planations of Lue process by which socio-economic equality is trans­ formed into residential proximity. MAJOR ASSUMPTIONS AND DEFINITIONS A principal assumption underlying this investigation is that an understanding of how residential location seekers' social goals inter­ act with the spatial system in which they find themselves is of crucial importance in explaining their choice of a household location. The interaction of social goals and physical space results in what is termed a location or space preference. Subjective accessibility, or the measure of the spatial distribution of different types of 31 Interaction opportunities about a household location adjusted for the ability and desire of residential decision makers to overcome spatial separation from the particular opportunities they may be more likely to interact with, is hypothesized as being related to this space prefer­ ence. The measure provides a means by which decision makers weigh, combine, or "trade off" the site and situation attributes associated with alternative household opportunities to derive an index of the attractiveness of different residential areas. Locational Preferences A study of the literature describing different residential areas suggests that decision makers reveal space preferences in their choice of a household location. The preferences can be related to specific site and situation attributes of the areas, and by weighting or or­ dering the attributes by the preferences, a subjective accessibility index of its attractiveness for residential development for different types of location seekers can be made. Households may then be al­ located to areas on the basis of a total projected demand and the at­ tractiveness of the areas. The following preferences are revealed by residential decision makers in their choice of household locations: An accessibility preference is an attitude of desire or need on the part of residential decision makers to choose households at loca­ tions from which they may facilitate their interaction with the major urban functions (essentially work places). A social-distance preference is a non-spatial attitude of desire or need on the part of residential decision makers to confine their social relationships with others of approximately equal rank or status. 32 This preference helps to define the geographical bias exhibited by decision makers of different socio-economic classes in their locational and travel behavior by determining the neighborhood type that different decision makers will be more likely to locate their households wherever found in urban space. A family life-style preference is an attitude of desire or need on the part of residential decision makers to choose household types that are deemed necessary by them because of their life-style stage. This preference helps to define the housing type required by decision makers depending on their marital status and family size. Although a treatment of this preference is beyond the scope of this investigation, it serves to further limit the number of available household opportuni­ ties that decision makers perceive in the urban space. A racial or ethnic preference refers to an attitude of desire or need on the part of residential decision makers to confine their house­ hold choices to areas having people of approximately similar racial or ethnic backgrounds. Such a preference would be similar to social- distance p r e f e rences except that decision makers' attitudes toward people of different racial and ethnic backgrounds change regardless of their social rank or status. A treatment of this preference iB also outside the scope of this investigation. Subjective accessibility, therefore, results from an attitude of desire or need on the part of residential decision makers to choose a particular housing type in a compatible neighborhood located where spatial separation from their particular set of interaction opportuni­ ties is at a minimum. In this investigation these opportunities relate to employment and such social relationships as may be involved with friends, I^in, neighbors, and informal social groups. 33 Occupation as a Cue to Socio-Economic Status Research has shown that occupation is one of the most important determinants of socio-economic status in the American c o m m u n i t y . ^ 2 Moynihan points out, "In the United States, what you do is what you are. By and large, status is distributed through the occupational structure. Although not denying the importance of other variables, it appears that occupation provides an easily obtained symbol by which people may define their attitudes toward and expectations of a decision maker or his family. In discovering the close correlation between oc­ cupation, and income, housing, and education, investigators point out that both income and education to a great extent determine the prefer­ ences which govern the manner by which people behave and live their lives. Laumann and others conclude that occupation serves as a realistic condensed cue by which other decision makers categorize a person or family according to their own set of preferences and elect to imitate or not to imitate the behavior associated with the person or f a m ily.^ Whether it be spatial behavior, or any other type of 22 Lloyd W. Warner, jet. al. , Social Classes in America (New York: Harper Torchbooks, 1960). Warner found that occupation alone had a coefficient of correlation of .91 with all the characteristics of families used to establish social classes. Investigators have con­ cluded that from the standpoint of scientific parsimony, occupation yields an index of stratification sufficiently accurate for most practical purposes. 23 Daniel F. Moynihan, A statement attributed to Moynihan by Joseph Alsop. The statement appears in Moynihan's forthcoming book, the publication of which has been reportedly deferred until he leaves his position as chief United States advisor on urban problems. Lansing, Michigan, State Journal, February 10, 1969. o/ Laumann, o£. c i t . 34 behavior, people will tend to imitate what they consider to be success­ ful behavior and not to imitate what they consider unsuccessful be­ havior. Relative high status occupations are associated with successful behavior and low status occupations with less successful behavior. For this reason this investigation will use occupation as a cue to the dif­ ferential spatial behavior exhibited by residential decision makers. A description of the socio-economic status types (or household types) found in the Lansing - East Lansing, Michigan, area is presented in Appendix A. LIMITATIONS OF THE i n v e s t i g a t i o n It is recognized that all the characteristics associated with the behavior of residential location seekers cannot be defined, objectively measured, or evaluated because of the present state of the knowledge about the behavior. Even if the behavior were understood, it would be difficult to simulate the behavior to generate a household development pattern because of the lack of compatible time-series data. Conse­ quently while researchers interested in the household allocation problem would like to be more concerned with the spatial behavior of individuals, their studies deal only with aggregates of population. While the same is true of this investigation, an attempt is made to disaggregate the decision makers into major social class groups and to isolate particular hypotheses about the locational preferences ex­ hibited by the spatial behavior of these groups. In this way it is possible to determine whether or not the specified preferences are systematically related to the process of allocating households in urban space. CHAPTER III DESCRIPTION OF RESIDENTIAL DECISION MAKERS' LOCATIONAL PREFERENCES The previous chapter argued that residential allocation models lose their power to predict the assignment of dwelling units to differ­ ent areas of the city by aggregating all household location seekers in­ to one homogeneous category and by ascribing to the residential decision makers the same locational preferences as non-residential decision makers. The effect of the aggregation is to attribute to the individual residential decision maker a wider range of locational opportunities from which to choose a household than he actually perceives in his action space. In models which disaggregate types of decision makers, the effect of ascribing similar locational preferences creates rela­ tively large areas of homogeneous land use, despiLe important e x ­ ceptions to the contrary. It was suggested that a more efficient household allocation model must possess at least two qualities to eliminate these problems. First, it must divorce itself from the loca­ tional preferences revealed by the spatial behavior of non-residential decision makers and, instead, treat the locational preferences re­ vealed by the spatial behavior of residential decision makers. Secondly, it must not deny the individual decision maker the chance to exercise his conditioned preference to choose or not to choose a compatible location that is consistent with his values, needs, and 36 desires from among the alternative household opportunities he is likely to perceive in his action space. The present chapter is designed to describe the locational preferences of different groups of household decision makers as revealed by their spatial behavior. Subsequent chapters will be concerned with both modeling and testing specific hypotheses considered relevant to the household allocation process. RESIDENTIAL PREFERENCES REVEALED BY THE CLASSICAL MODELS OF URBAN SPATIAL STRUCTURE A classic approach to the problem of explaining the growth of residential land has been to describe the location of residential areas by household type and to formulate generalizations about the nature of future development patterns. Hurd, as early as 1903, introduced the idea that urban growth proceeded according to two patterns: growth and axial growth. central He observed that, A continual contest exists between axial growth, pushing out from the centre (sic) along transpor­ tation lines, and central growth, constantly fol­ lowing and obliterating it, while new projections are being made further out the various axes.l Hence, urban growth tends to occur in all directions outward from the center of the city, and it occurs more rapidly along major transpor­ tation routes. Although Hurd discusses the location of different types of residential areas, paying particular attention to the income and rent payments of household decision makers, he does not use the concept of central and axial growth to systematically generate patterns of residential growth and development. ^Hurd, o£. c i t ., p. 59. He does, however, say, 37 The main consideration in the individual selection of a residence location is the desire to live among one's friends or among those whom one desires to have for friends . . . and the ultimate aim in residence location is to be as close as possible to those of the highest social position.2 A few decades later, in the early '20s, the compact circular urban form led Burgess to stress the central growth aspect of Hurd's concept to generate the well-known concentric zone distribution of residential areas by household type.-* He described residential areas according to the density of dwelling units and the socio-economic characteristics of the residential decision makers. Burgess found that the two vari­ ables were inversely related to each other so that proceeding outward from the central core the zones become progressively less dense in dwelling units but increase in the socio-economic status of the decision makers. He and Park argued that the dominance of a particular land use in a zone, and consequently urban form, was the result of competition among different interest groups. Land use changes take place through the Invasion into the zone by a competing interest group which may lead to the ultimate succession to dominance of the new land use in the area. This process receives its energy from popu­ lation increase and generally manifests itself in older population groups migrating from original sites near the congested central core ^Ibid., p . 78. 3 Ernest W. Burgess, "The Growth of the City: An Introduction to a Research Project," Publications of the American Sociological Society, Vol. 18, 1924, pp. 85-97. Reprinted in Robert E. Park, Ernest W. Burgess and Roderick D. McKenzie (eds.), The City (Chicago: University of Chicago Press, 1925), pp. 47-62. See also: Ernest W. Burgess, "Urban Areas," Chicago: An Experiment in Social Science Research, T. V. Smith and L. D. White, editors (Chicago: University of Chicago Press, 1929), pp. 114-123. 38 toward less-crowded sites closer to the periphery as their socio­ economic status improves. Residential location seekers, new to the city or with little or no family responsibilities, are able to find more numerous household opportunities in the vacated residences near the central core that have not been demolished for higher-order, nonres idential land uses.^ The spatial behavior of residential decision makers, as viewed by Burgess, reveals a locational preference for a more spacious living environment away from the overcrowded conditions near the central core as real income increases. Residential decision makers having lower in­ comes choose sites that exhibit a locational preference to be more centrally situated with respect to the major urban functions, par­ ticularly employment opportunities. Nearly t w o decades later, in the late '30s, the star-shaped urban form caused Hoyt to emphasize the axial pattern of Hurd's con­ cept. In an empirical study of sixty-eight cities he found, the highest rental area is in every case located in one or more sectors on the side of the city (and) 4 Ernest W. Burgess, "Residential Succession in American Cities," Annals, Vol. 140 (November, 1928), p. 112. See also: Lawrence A. Brown and David B. Longbrake, Migration Flows in Intra-Urban Space: Place Utility Considerations (paper presented at the Annual Meeting of the Association of American Geographers, Washington, D.C., August, 1968). Additional studies dealing with residential succession are: Paul F. Cressy, "Population Succession in Chicago: 1898-1930," American Journal of Sociology, Vol. 44 (July, 1938), pp. 59-68; Richard G. Ford, "Population Succession in Chicago," American Journal of Sociology, Vol. 56 (September, 1950), pp. 156-160; Otis Dudley Duncan and Stanley Lieberman, "Ethnic Segregation and Assimilation," American Journal of Sociology, Vol. 64 (January, 1959), pp. 368-369; Stanley Lieberman, "Suburbs and Ethnic Residential Patterns," American Journal of Sociology. Vol. 67 (May, 1962), pp. 673-682; Stanley Lieberman, Ethnic Patterns in American Cities (New York: The Free Press of Glencoe, 1963). I 39 low rent areas, extending from the center to the edge of the city are found in practically every city.5 In contrast to Burgess, Hoyt showed that both high and low rent neighborhoods occupy distinct sub-areas of the city and that they are not aligned concentrically about the central core but are distributed in a sector fashion. He proposed that city form is determined by the residential location seekers who can best afford to pay the highest rent. These location seekers pre-empt land along "the best existing transportation lines," "high ground... free from the risk of floods," and "land along lakes, bay, river and ocean fronts where such water­ fronts are not used for i n d u s t r y . T h e y also tend to locate on the side of the CBD where office buildings, banks, and stores are located and away from the side where warehouses and industries are found. In turn all other socio-economic groups tend to locate their households in the direction of the residences of these "community leaders" or other prestigeous leaders with whom they "identify" themselves.^ This development continues in the same direction for a long period of time. The spatial behavior of residential decision makers^ as viewed by Hoyt, reveals a social attraction to prestigeous neighborhoods as -*Hoyt, o£. c i t ., pp. 75-76. ^Ibid., p. 117. Lloyd Rodwin, in applying Hoyt's model to the City of Boston in 1950, found that most of the characteristics dealing with residential growth and form proved generally to be true although he discovered inconsistencies with some of Hoyt's propositions. See: Lloyd Rodwin, "The Theory of Residential Growth and Structure," Appraisal journal, Vol. 18, 1950, pp. 295-315. See also: Homer Hoyt, "Residential Sectors Revisited," Appraisal Journal, Vol. 18, 1950, pp. 445-450; Walter Firey, "Residential Sectors Re-Examined," Appraisal Journal, Vol. 18, 1950, pp. 451-453. ^Hoyt, Structure and Growth of Residential Neighborhoods in American Cities, p. 120. 40 the fundamental locational preference, with accessibility to employment opportunities a secondary locational preference. No new, significant descriptive model of the residential spatial structure has been formulated since those of Burgess and Hoyt, except that Firey has argued that residential location seekers who share the same cultural values and goals will place the same significance on g the various sub-areas of the city and tend to cluster in urban space. Polynuclei descriptions of urban spatial structure are considered to be metropolitan forms rather than city forms and tend to over-emphasize the role of transportation technology on changing land use patterns. Q It is important to note that advocates of the "economic competition" hypothesis tend to support the Burgess model, while the Hoyt model is closely associated with the "social choice" hypothesis. RESIDENTIAL PREFERENCES REVEALED BY SOCIAL AREA ANALYSIS STUDIES It was not until another two decades had passed, in 1960, that the zonal and sector models were compared and tested statistically. Prior to this comparison, however, a means of empirically measuring the relevant characteristics of residential areas needed to be de ­ termined. Such empirical instruments were developed by Shevky and ®Firey, "Residential Sectors Re-Examined," work is of limited value in predicting patterns and development. He argues that urban land use 1) past and present cultural systems and social munity goals, values, and sentiments. pp. 451-453. Firey's of residential growth is both a function of systems; and 2) com­ 9 Harris and Ullman, oj>. c i t . , pp. 7-17. See also: Edward L. Ullman, "The Nature of Cities Reconsidered," The Regional Science Association Papers and Proceedings, Vol. 9, 1962, pp. 7-23. 41 Williams in 1949 and consist of three indices designed to describe the population and housing characteristics of urban census tracts using from one to three census variables each.*^ The indices allow the census tracts to be differentiated into "social areas" and measure: 1) social rank or socio-economic status; 2) urbanization or family life-style; and 3) segregation or minority group isolation. Subsequent investi­ gations have shown that the indices effectively summarize the resi­ dential characteristics of census tracts and possess a relatively stable structure from city to c i t y . H The first to statistically test and compare the zonal and sector models in terms of the Shevky-Williams-Bell indices were Anderson and Egeland in 1960. 12 Using analysis of variance they found that both the Burgess and Hoyt descriptions of the residential spatial structure have validity; that is, census tracts, when classified according to their degree of urbanization or family life-style status are distributed Eshref Shevky and Marilyn Williams, The Social Areas of Los Angeles, Analysis and Typology (Berkeley and Los Angeles: University of California Press, 1949); Eshref Shevky and Wendell Bell, op. cit., pp. vi and 70. It must be remembered that these indices pertain to spatial units rather than individuals and describe the environmental attributes of census tracts. Socio-economic variables deal with the level of occupation, education, income, rental value, and the quality of housing. Family life style variables are the number of women in the labor force, family size, and age of household members. The ethnic variables deal with minority group concentrations. For a dis­ cussion of these indices and an extensive bibliography of work done using them, see: Wendell Bell, "Social Areas: Typology of Urban Neighborhoods," Community Structure and Analysis, Marvin B. Sussman, editor (New York: Thomas Y. Crowell Company, 1959), pp. 61-92. ^ M a u r i c e D. Van Arsdol, Jr., Santo F. Camilleri and Calvin F. Schmid, "The Generality of Urban Social Area Indexes," American Sociological Review, Vol. 23 (June, 1958), pp. 277-284. 12 Anderson and Egeland, o£. cit. , pp. 392-398. 42 in a zonal pattern, whereas when differentiated on the basis of their social rank or socio-economic status they are distributed in a sector pattern.^ Other investigations support this finding and suggest that each "social area" index captures the essential features of one of the classical spatial models: 1) Burgess' concentric zone model de­ picts family life-style patterns; 2) Hoyt's sector model is charac­ terized by socio-economic status patterns; and 3) Firey's social cluster model portrays ethnic status patterns. Several researchers have argued that the three models are independent additive contributors to the total structuring of urban neighborhoods. Berry, for example, states, This basic triad of spatially arranged social dimensions can be super-imposed to form, at the intersections of sectors, zones, and segregated areas communities of similar social, family and ethnic status. 14While the social area analysis methodology provides some knowl­ edge about the relevant locational preferences of residential decision makers, it lacks the ability to determine the degree of importance that can be placed on them; that is, the relative weight attached to each preference. It is essential to be able to determine whether household decision makers are more influenced in their choice of a residential location because of 1) a preference to be close to the households of Anderson and Egeland found that areas characterized by high dwelling density, families with few children, and families in which the wife works are more frequently near the city center and gradually de ­ cline in relative frequency toward the periphery; i.e., these variables were distributed in a concentric zone fashion. Variables associated with the prestige of residential areas, as measured by occupation and education of the inhabitants, were distributed in sector fashion, although the position of the sectors varied from city to city. 14_ Berry, o£. cit. 43 equal or more prestigeous decision makers vis-a-vis a preference to be accessible to the major urban functions as postulated by Hoyt's socio­ economic, sector model or 2) a preference for more spacious living con­ ditions vis-a-vis a preference for a location that is accessible to the major urban functions as postulated by Burgess in his family life­ style, zonal model. RESIDENTIAL PREFERENCES REVEALED BY FACTORIAL ECOLOGY METHODS Studies using factor analysis to determine the validity of the social area analysis indices as measures of the socio-economic, family life-style and segregation attributes of census tracts indicate that socio-economic status is the more important factor in dimensioning the preferences by which household choices are made to produce the resi­ dential spatial s t r u c t u r e . S h e v k y and Bell have shown that, in the United States, factorial ecology revealed that the relevant preferences which result in household choices are dimensioned cumulatively: first, by socio-economic status; second, by family life-style status; and third, by the constraints of race and ethnicity .^ An investigation of ^**It is the pattern of preferences described by Burgess that is being treated in the work of Muth, Cassetti, Wingo, and Alonso. ^Factorial ecology is the term used to characterize studies in­ volving the application of factor analysis to ecological investiga­ tions . ^ F o r a review of this literature see: Berry, o£. c i t . A slightly different factor structure exists between cities of the United States' North and South, and between cities of developed and underdeveloped countries. The different factor structures generally occur with respect to the race and ethnicity dimensions. In several cities of the South no factor separation can be found between indicators of socio­ economic status and indicators of ethnic status. This not unexpected 44 the locational behavior revealed by the residential spatial structure of Pittsburgh in 1950 and 1960 cause Lee to conclude, if one factor alone could be used to predict the locations of households the single dimension ... which we would choose ... would be socio-economic status. 1® This study will focus on Hoyt's socio-economic status description of the urban residential spatial structure while, at the same time, recognizing that accessibility to the major urban functions, family life-style, and ethnic considerations have a part in dimensioning the locational preferences that result in household choices. RESIDENTIAL PREFERENCES REVEALED BY CONSUMER BEHAVIOR The "economic competition" hypothesis argues that household decision makers budget their expenses on the basis of satisfactions they receive from the consumption of goods and services within the constraints of their income. However, these satisfactions cannot be measured purely in terms of the quantity of land they utilize at some factor linking indicates that a decision maker's ethnic status (Negro or non-Negro) is fundamental in defining the socio-economic preferences he exhibits by his household locational behavior in southern cities. Despite similar situations that appear in Montreal, Chicago, and Toledo, all studies show socio-economic status as the more important factor in structuring residential choice than family life-style status. l®Douglass B. Lee, Jr., "Household Disaggregation in Urban Models," (paper presented at the Annual Meeting of the Regional Science Asso­ ciation, Cambridge, Massachusetts, November, 1968). Lee found that socio-economic status explained 37 per cent of the total variance in 1960 and 35 per cent in 1950; family life-style status explained 18 per cent of the variance in 1960 and 23 per cent in 1950; and ethnic status explained 14 per cent of the variance in both 1960 and 1950. His analysis also suggests that socio-economic status is becoming more of a factor in dimensioning residential spatial structure and family life-style status less of a factor as time goes on. See Appendix B. 45 distance from the central core as evidenced in the work of Muth, Cassetti, Wingo, and A l o n s o . ^ Veblen, as early as 1899, and more recently Galbraith, have pointed out that the consumption habits of household decision makers reveal that they consume their total range 20 of goods and services for two purposes. w First, they need the product or service to satisfy the basic requirements of food, clothing, and shelter. Second, they need the product or service for the purpose of identifying themselves with a particular reference group. Advocates of the "social choice" hypothesis argue that it is the satisfactions that decision makers receive from identifying with their reference group that determine h o w they choose to budget their income and where they choose to locate their households. Several conclusions can be derived from consumer identification patterns that are applicable to the problem of allocating households to residential areas of the city. First, there must be other people near­ by belonging to the reference group that establish the pattern of con­ sumption to be followed by decision makers desiring to identify with the group. Second, since every act of consumption is at least status conserving, and in favorable circumstances status enhancing, there must be people around from the reference group to validate the status sought by the decision maker. Consequently, decision makers seeking, or ^ M u t h , o£. c i t ., pp. 207-220; Cassetti, o£. c i t ., pp. 96-100; Wingo, "An Economic Model of the Utilization of Urban Land for Residential Purposes," pp. 191-205; see also: Wingo, Transpor tat ion and Urban Land; Alonso, "A Theory of the Urban Land Market," pp. 149158; Alonso, Location and Land U s e . 20 Thorstein Veblen, The Theory of the Leisure Class, An Economic Study of Institutions (New York: The Macmillan Company, cl912). John K. Galbraith, The Affluent Society (Boston: Houghton Mifflin Co., 1958) . 46 attempting to maintain a particular social status, tend to cluster together to establish and verify particular consumption patterns. 91 *■ The concentric zone proposition based on "economic competition" does not allow for the reference group conduct revealed by consumer b e ­ havior as well as does the sector pattern based on "social choice." Hoyt's proposition is consistent with consumer reference group identi­ fication behavior in that it shows how decision makers of lower socio­ economic status either identify with people in the same strata or, in the case of the upwardly mobile decision makers, identify with people in higher status groups. Hoyt's proposition is also consistent with the invasion-succession process of land use change and with the avoidance-acceptance proposition found in psychological studies of inter­ personal and intergroup behavior. Consumer reference group identifica­ tion behavior can be stated in terms of whether a decision maker be­ longing to one socio-economic group wishes to avoid decision makers of other socio-economic groups or wishes to be accepted by them. The probable level of acceptance or avoidance between the different status groups is hypothesized in Table III.l. Laumann1p findings of the pro­ pensity of a decision maker in a particular occupation category is in remarkable agreement with this hypothesized level (See Table III.2).22 Finally Hoyt's proposition is consistent with the results of moving behavior and residential choice as found recently in a survey by Butler, at. a_l. The survey revealed that household decision makers 21 Chapter Five will treat the probable level of interaction between different status groups to generate household locations. 22 Laumann, ££. c i t ., p. 71. table III.l REFERENCE GROUP IDENTIFICATION BY OCCUPATION CATEGORY OF DECISION MAKERS (HYPOTHESIZED LEVEL OF ACCEPTANCE OR AVOIDANCE) OF REFERENCE GROUP LEVEL OF IDENTIFI­ CATION DESIRED: Occupation Category Profession­ al & Tech, BY VIEWING GROUP Professional, Technical Proprietors, Managers & Salesmen Clerical Workers Skilled Workers XX X — Props., Mans., & Salesmen XXX XX X Clerical Workers XX XXX XX X X XX XXX XX X X XX XXX XX Skilled Workers Semi & Unskilled XXX * Great Acceptance XX ■ Acceptance X « Some Acceptance - Some Avoidance Avoidance Great Avoidance __ Semi & Unskilled ... - TABLE III.2 PER CENT DISTRIBUTION OF OCCUPATIONAL STATUS OF NEXT-DOOR NEIGHBORS BY OCCUPATIONAL STATUS OF DECISION MAKERS NEXT-DOOR NEIGHBORS' OCCUPATIONAL STATUS Professional Top Business Semi-Prof. Mid. Bus. Clerical Small Bus. Skilled Workers Semi & Unskilled Profess. Top Bus. 48.6 23.9 14.5 6.5 6.5 Semi-Prof. Mid. Bus. 24.5 26.6 25.9 10.8 12.2 23.0 14.8 32.6 Occupation Category Source: Clerical Small Bus. t—4 CO DECISION MAKERS' OCCUPATIONAL STATUS 21.5 Skilled Workers 7.6 12.6 18.5 19.3 42.0 Semi & Unskilled 4.9 4.1 12.2 22.8 56.1 Laumann, Prestige and Association in an American Community, p. 71. 49 overwhelmingly preferred better neighborhood quality with either a less desirable housing unit or less accessible loca­ tion over a less desirable neighborhood with either a better housing unit or better access­ ibility (70 per cent to 27 per cent).23 The study also revealed that location seekers prefer a place that has a very nice appearance inside and less desirable outside appearance to a place that presents a very nice outside appearance but less desirable inside appearance (80.4 per cent to 14.2 per cent).24 The lack of concern for outside appearance coupled with a concern for neighborhood quality suggests that the major preference of decision makers in choosing a household location is not spacious living con­ ditions as proposed by the "economic competition" hypothesis. Rather, neighborhood prestige is a more critical preference in choosing house­ hold locations from among alternative opportunities. RESIDENTIAL CHOICE REVEALED BY SOCIAL-DISTANCE PREFERENCES Hoyt's proposition that residential choice is a function of the differential prestige associated with socio-economic groups has been confirmed by a number of studies. Duncan and Duncan found a close relationship between what they termed spatial and social distance in an investigation of the differential distribution of occupations by 23 Edgar A. Butler, e_t. _al. , Moving Behavior and Residential Choice: A National Survey (Center for Urban and Regional Studies, Institute for Social Science Research, Chapel Hill: University of North Carolina Press, March, 1968), p. 5; see also: Michael A. Stegman, "Accessibility Models and Residential Location," Journal of the American Institute of Planners, Vol. 35 (January, 1969), pp. 22-29. 24 Butler, et. a l ., l o c . cit . 50 census tracts in Chicago. They concurred with Park's proposition that residential proximity is associated with social equality so that the physical distance at which socio-economic groups of different status reside from each other is positively related to their social distance a p a r t . ^ Feldman and Tilly attempted to assess the relative importance of income and education in explaining the differential dis­ tribution of occupations in Hartford, Connecticut's census tracts. They found that the residential distribution of occupations was comparable to the pattern observed by the Duncans in Chicago and also concurred with Park's formulation. 2A ° Both studies lend support to Hoyt's proposition that non-economic factors are important in explain­ ing the differential residential distribution of the socio-economic classes. Additional support for Hoyt's proposition can be found in Laumann's investigation of the social-distance relationships of a sample of Belmont and Cambridge, Massachusetts, residents. Laumann defines social distance in terms of subjective preferences and objec­ tive choices. Subjective social distance is an attitude of ego toward a person (alter) with a particular status attribute (in this and Laumann's study occupation) that broadly defines the character of the inter­ action that ego would undertake with the attitude object.27 Objective social distance is the actually observed likelihood of having a friend, neighbor or kin 25 2A Duncan and Duncan, oj>. c i t ., pp. 493-503. Feldman and Tilly, oj>. cit., pp. 877-884. ^Laumann, £j>, c i t . , p. 29. 51 relation with persons of different occupational background.28 Two facts are evident from Laumann's study of social— distance prefer­ ences: 1) each socio-economic class differs with respect to their social-distance preferences from the different occupations, and 2) social-distance preferences are highly correlated with the prestige ordering of the different occupations regardless of the socio-economic class of the decision maker; that is, the expressed preference of all classes is to associate intimately with the higher status types and to maintain relatively greater distance from lower status types. OQ The general trend of the three lines in Figure III.l from the upper left-hand corner to the lower right-hand corner dramatically illustrates this finding with respect to the social-distance preferences of the up­ per, middle, and working class groups. Laumann has found a nearly total diffusion of occupational prestige as a factor which dimensions the preferences by which decision makers choose their friends, spouses, and household locations.**® 28Ibid., 29 30 Although not considering the spatial relations p. 30. Ibid., p. 45. Studies by Centers and Hunt, although not concerned with the residential spatial structure, tend to support the hypothesis that there is a higher rate of intimate interaction and communication within the same occupational category and a lower rate between different oc­ cupational categories. See: Richard Centers, "Marital Selection and Occupational Strata," American Journal of Sociology, Vol. 54 (May, 1949), pp. 530-535; and T. C. Hunt, "Occupational Strata and Marriage Selection," American Sociological Review, Vol. 5 (August, 1940), pp. 495-504. Several other studies find support for Hoyt's proposition that decision makers choosing others for close social relationships tend to name those of somewhat higher socio-economic status than themselves. For example, Loomis and Beegle reported that rural church members of upper or middle class in a small Michigan community chose persons of the same or higher socio-economic status more often than 52 PROFESSIONALS TECHNICALS UPPER 0 L A 8 9 M ID D L E CLASS PROPRIETORS WORKINS OLASS OCCUPATI ON CATEGORY MANASKRS CRAFTSMEN FOREMEN OPERATIVES LABORERS SSO SOO 4S0 400 PREFERRED F I 0. Preferred social SOURCE! SO C IA L for by s e l f - i d e n t i f i e d seventeen oc c u p a tio n a l c a t e g o r i e s C a m b r id g e and Belmont O. L i u m i n n , DISTANCE I m ean social d is t a n c e f o r m ale s class sample o f IN MEAN Rrattl(a an4 A i»«ci a tlo n SOO BOO tn an UrSan in a Massachusetts r e s i d e n t s C om m unity a 42 53 of friend and kin associations, he found the likelihood of having nextdoor neighbors belonging to the same occupation category to be generally quite high although there was a slight tendency for the respondent's O1 occupation to be a little more prestigeous than that of his neighbor.->JDespite significant methodological differences between Laumann's study and that of the Duncans, both found that the residential dis­ tribution of occupations exhibits a pattern which corresponds to the different social-distance preferences of each socio-economic class, but only in the case of the upper class is the distribution highly corre­ lated to the prestige ranking revealed by their social-distance prefero0 ences. A comparison o f .the line trend in Figures III.2 and,III.3 would be expected by chance, and those beneath them in socio-economic status less often than would be expected by chance. See: C. P. Loomis and J. A. Beegle, Rural Social Systems (New York: Prentice Hall, 1950). For a review of additional studies which support the prestige hypo­ thesis see: Henry W. Riecken and George C. Homans, "Psychological Aspects of Social Structure," Handbook of Social Psychology, G. Lindzey, editor, Vol. 2 (Reading, Massachusetts: Addison-Wesley, 1954), pp. 786832; see also: Morton B. King, "Socio-economic Status and Sociometric Choice," Social Forces, Vol. 39 (March, 1961), pp. 199-206. 31 Laumann, oj>. c i t ., p. 72. Tt must be remembered that a house­ hold location seeker, regardless of his class, exercises only a passive control over the status of the next-door neighbor because of the limited chance that an available household opportunity will present itself next-door to the neighbor of his choice. In order to reduce the possibility of getting an obnoxious neighbor, location seekers exercise an active control over their choice of a neighborhood and take their chance on what the status of the neighbor is. By their selection of a slightly more prestigeous neighborhood the chance of locating next to an undesirable neighbor is lessened. In comparing Figures III.l, III.2, and III.3 it is of less im­ portance to be concerned with the particular values noted than with the slope and character of the curves. For example, the relationship of the lines to each other in all three figures is approximately the same, and they cross each other at approximately the same occupation categories in each figure. TOP P R O F E S S IO N A L S P R O F E S S IO N A L S T C C H N IC A L S A N D B U S IN E S S P R O P R IE T O R S S M A L L B U S IN E S S CLASS S K ILLE D M ID D L E C L A S S WORKERS OF UPPER SEM I S K IL L E D , U N S K IL L E D CATEGORY W O R K IN G C L A S S CATEGORY C L E R IC A L ANO M ANAG ERS C L E R IC A L S A L E S W ORKERS UPPER C LA SS OCCUPATI ON M ID D L E B U S IN E S S NEIGHBORS S E M I- P R O F E S S IO N A L S N IO O L E C L A S S — — ------ W O R K IN G Ui ■l>O P E R A T IV E S 60 SO I I--- 1 40 30 EO OCCUPATI ON W ORKERS I I I 10 0 0 tdwB’l0lia»«aa to SO 40 SO 60 L OW ASSOCIATION INDEX OF RESIDENTIAL OCCUPATION ASSOCIATION FIG . IM -3 F IS . HI - 2 o f nast door m ala n o io tilio rs social groups in a samplo of CuiTiDnoyu jn d 10 HlOK ASSOCIATION PER CENT RESIDENTIAL OCCUPATION ASSOCIATION O ccupation c a ta a o rj fo r Boimoot.lNoss. its id a n fa IM AB«»tr*tiBR *•Ur»«* 9 VI CLASS social class groups in Ctiicago, IRinois cansws tra cts-IS S O MuflCI Ot»s t Bwugga osl ••agrlp Butte*. *’Beg•!gtftei «•« tHfapgiigagi 55 shows that only the upper class group reveals residential locational behavior that is highly consistent with their social-distance prefer­ ences shown in Figure III.l; that is, the line trend in all three figures is from the upper left-hand corner to the lower right-hand cotHer. The residential distribution of the middle class indicates a progressively poorer correlation with the prestige ranking and that of the lower class an inverse correlation with the prestige ranking. For the working class group a comparison of the two lines in Figure III.2 and III.3 representing their actual residential association with the line on Figure III.l representing their social-distance preferences reveals that the residential association is in sharp conflict with their preferences as the lines cross to resemble an X. This conflict of preference with the actual residential association could be con­ strued to mean that it is totally the economic circumstances of each class that determines the household locations, but the fact that the Duncans found blue and white collar workers at the same income level to have different residential patterns suggests that the prestige of neighbors is less of a locational preference Lo lower status decision makers than is a preference to be accessible to the major urban func­ tions.-^ Arguments to the effect that blue collar workers reside close to the major urban functions because of lower trip costs while white collar workers can more easily absorb trip costs are likewise invalid since studies by Tomazinis and Gabbour found that decision makers re­ siding close to the central core have similar costs to those residing at the periphery. 33 They found that decision makers residing at Duncan and Duncan, loc. c i t . 56 intermediate locations realized lower trip c o s t s . ^ While not denying that economic circumstances are involved, the findings of Laumann, Feldman and Tilly, Tomazinis and Gabbour, and the Duncans tend to present a good case for Hoyt's "social choice" proposition that the socio-economic classes are highly influenced in their choice of a house­ hold location because of a preference to be close to the residences of equal or more prestigeous decision makers rather than they are because of a preference for more spacious living conditions. Figure III.4 shows the residential location by census tracts of the upper, middle, and working class groups in the Lansing - East Lansing Area for 1960 as determined by the location quotient technique. In accordance with the findings of Laumann and the Duncans the technique reveals that the upper class group is the more highly segregated, followed by the work­ ing class group. The middle class group shows less of a tendency to be segregated. SUMMARY This chapter has provided an overview of the literature which describes the household locational preferences of residential decision makers as revealed by their spatial behavior. The discussion centered on the "social choice" hypothesis as an alternative approach to the "economic competition" hypothesis in explaining the differential resi­ dential distribution of socio-economic groups in urban space. The analysis found that occupations assume a kind of hierarchical arrange­ ment in terms of the prestige they may confer on decision makers, so that certain occupations become identified with particular socio-economic ■^Tomazinis, loc. cit.; Gabbour, loc. c i t .; Tomazinis and Gabbour, loc. cit. (!„■«. ,# ^o^ iaqCIosi croHl«in,foriitA i i r v i c t tfo rk tri) Niddif C13S1.3' N 0S . UPptr 9r [p r o p r itto fs .s d ld . o ffic io i* , d t n c e l , . g V . 0n CIOI* ( p r o ft iS M n o l* t.chnicolo) in "J #1 ,*wf F I G. Ill* 4 R ESID EN TIA L SEGREGATION BY OCCUPATION CATEGORY classes. It was pointed out that the actual existence of socio­ economic status groups and the houses in which they choose to live de­ pends in considerable measure upon the level of income derived from these occupations so that the occupations themselves have considerable status significance. However, it was stressed that the location of the households in urban space nas considerably less to do with the in­ come level of the decision makers. Instead it was found that resi­ dential location seekers exhibit a primary preference to reside close to the households of equal or more prestigeous decision makers and con­ sequently they are not indifferent to their locations at designated prices as purported by the "economic competition" hypothesis. This social-distance preference was more obvious with respect to the loca­ tional behavior of upper socio-economic classes. Secondly, the loca­ tional behavior of decision makers revealed a preference to choose locations with respect to the various urban functions deemed neces­ sary by them to maintain their livelihood. This accessibility prefer­ ence was more obvious with respect to the locational behavior of the lower socio-economic groups and more affluent young ma rri ed couples, recent arrivals to the city, and older couples beyond the child rearing stage. Thirdly and to a much lesser extent, location seekers reveal a preference for more spacious living conditions. This life-style preference was more obvious with respect to decision makers in the family rearing stages. Finally, residential locational behavior re­ vealed a preference for decision makers to reside close to households of similar minority group status. This ethnic or racial preference was found to be more significant in cities where the proportion of ethnics to the total population was generally quite high and consequently 59 relegated the ethnics to very low status. Decision makers belonging to similar socio-economic, family life-style and ethnic status weigh these preferences in the order described above, and the additive weights contribute to the total structuring of residential spatial patterns. It is significant to note that this notion is a possible aid to the problem of allocating residential growth to different areas of the city. The residential locational preferences described in the chapter differ in both quality and degree from those exhibited by the spatial behavior of non-residential location seekers. The locational prefer­ ences of non-residential location seekers are primarily, accessibility and secondly, spaciousness. However, in some cases, it can be seen that non-residential location seekers also exhibit a preference to be close to prestigeous neighborhoods and ethnic groups. CHAPTER IV MODELING OF RESIDENTIAL DECISION MAKERS 1 LOCATIONAL PREFERENCES The principal findings of urban residential location studies have yielded two guiding hypotheses to account for the distribution of households in urban space: choice" propositions. the "economic competition" and "social In the previous chapter it was argued that residential decision makers' locational preferences appeared to be more a matter of "social choice," whereas "economic competition" character­ ized the locational preferences of non-residential decision makers. Both hypotheses propose that decision makers will choose locations that provide maximum geographic utility or satisfaction from among the alternative opportunities perceived in their action spaces. (Geogra­ phic utility refers to both site or place utility and situation utility.) Land use allocation models in their purest form are devices that attempt to measure the geographic utility of urban areas in terms of decision makers' locational preferences and assign site choices on the basis of the attractiveness of the opportunities in the areas. Resi­ dential decision makers' locational choices reveal that they make selections by evaluating a number of attributes associated with alternative residential opportunities. Household allocation models attempt to account for the many attributes which are assumed to be the 60 61 most relevant to this decision-making process. To predict locational choices the various attributes are weighted and combined in what is presumed to be the manner in which the decision makers trade off the attributes between alternative opportunities according to their preferences. Both the determination of the part-worth that each at­ tribute contributes to the over-all evaluation of alternative house­ hold opportunities and the appropriate rules for combining the component attributes are determined by experimentation or by observation of the residential decision making process.^ The part-worth that each attri­ bute contributes is differentially associated with the locational preferences of the decision makers and is presumably independent of the particular set of opportunities, the particular spatial system, or the particular moment in time. The household allocation model imple­ ments the decision makers' locational preferences by placing them in the context of a particular set of opportunities located in a par­ ticular spatial system at a particular moment in time.2 While in theory the models seek to represent the household decision making process, in practice the process is not known and households are allocated to locations on the basis of mathematical parameters determined by the aggregate past spatial behavior of decision makers as determined by multiple regression procedures, for example. The short history of household allocation models reflects an effort to overcome this problem. The present chapter will examine these models to determine the ^Shepard, oj>. cit. , pp. 269-270. 2Ibid., p. 279. 62 attributes considered by model builders to be relevant to the decision makers' evaluation of alternative household opportunities, particularly with respect to whether they inherently ascribe to the "social choice" or "economic competition" preference structures. The examination will be confined to operational and quasi-operational models whose ef­ ficiency has been evaluated to some degree and which are used, in •a original or modified form, by a great number of planning agencies. THE DEVELOPMENT OF HOUSEHOLD ALLOCATION MODELS Household allocation models developed as an adjunct to traffic assignment models about a decade ago, and with few exceptions this in­ fluence has persisted until the present time. This transportation orientation of household allocation models developed as a result of both a need for land use growth models in transportation studies and the mathematical modeling capability of transportation planners - a capability not found among land use planners until recently.^ 3 Since household allocation models hardly existed prior to 1960 and many of them are in a more or less continuous state of revision, the documentation of particular models is incomplete. Consequently, this investigation may not be aware of recent innovations in their development. 4 For an excellent compendium of methods utilized in predicting urban transportation demand including land use forecasting methods up until 1960 see: Brian V. Martin, Frederick W. Memmott, III, and Alexander J. Bone, Principles and Techniques of Predicting Future Demand for Urban Area Transportation (M.I.T. Report No. 3, Cambridge: M.I.T. Press, June, 1961). Particularly see pp. 75-96 for land use forecasting methods. Exceptions to the transportation based modeling efforts are the San Francisco CRP Urban Renewal Model developed by Arthur D. Little, Inc., the recent research models used at the University of North Carolina, and the Pittsburgh CRP Model designed by the University of Pittsburgh and CONSAD Research Corporation. 63 Transportation planners, concerned with the problem of predicting intraurban traffic patterns during the period 1945 - 1955, found that the gravity-type model introduced by Reilly in 1929 to describe the interurban movements of tripmakers could be used to approximate intraurban travel behavior as well.-’ Although principally devoid of behavioral considerations, this methodological advance became the basis of land use growth models. The approach is based directly on the em­ pirical observation that the type and intensity of land use at a loca­ tion is reciprocally related to the type and intensity of travel be­ havior at that location so that as the city grows the pattern of land use is modified; that is, the more accessible a location becomes in relation to all other locations, the more likely it will be more in£ tensively used. This observation has been explained by the "principle of least effort" which proposes that individuals prefer to choose house­ hold locations whose orientation in space or distance separation from other major urban activity locations is at a minimum. Curry points out that "the only method for describing the orientation of an individual in space is that due to Stewart ."7 Stewart's method renders an index of accessibility for any location and is expressed: E n c o 1 (IV1) j-i j ^William J. Reilly, "Methods of Studying Retail Relationships," University of Texas Bulletin, No. 2944, 1929. ^Hansen, "Accessibility and Residential Growth" and "How Accessibility Shapes Land Use." ^Leslie Curry, "Central Places in the Random Spatial Economy," Journal of Regional Science. Vol. 7, No. 2 (Supplement, 1967), p. 236. 64 where: Si * a measure of the intensity of opportunities in an area j (employment, social, or retail trade opportunities although most household allocation procedures use employment oppor­ tunities only); djj = the distance between area i and area j usually expressed as minutes of travel time to take into account spacetransforming technological innovations between different periods of time; and b ■ an empirically determined exponent describing the effect of distance on the attenuation of trips for the particular urban region under study.8 Accessibility measures of this type implicitly relate the household allocation process to the spatial behavior of location seekers by in­ dicating the relative intensity of travel observed between spatially distributed opportunities and household locations as decision makers attempt to satisfy their day to day demands for goods, services, and social intercourse. In all household allocation models except one some index of accessibility is either the only attribute or the major at­ tribute contributing to a location's level of geographic utility. Investigations of the variation between travel patterns generated by gravity-type formulations and actually observed travel behavior reveal that the distance-decay exponent (b) varies according to: 1) the size and density of the opportunities located in different areas, 2) the purpose of the trip, 3) the distance of the area from the CBD in which the trip originates, 4) the socio-economic character­ istics of the tripmaker, 5) the distance between areas, and 6) the level ®john Q. Stewart, "Empirical Mathematical Rules Concerning the Distribution and Equilibrium of Population," Geographical Review. Vol. 37, 1947, pp. 461-483. See also: Stewart and Warntz, "Macro­ geography and Social Science." 65 of technological development of the society. The result of these in­ vestigations shows that gravity-type formulations require considerable manipulation of proportionality factors in order to achieve results comparable to observed travel behavior. This finding has generated considerable debate among land use forecasters in their search for an efficient residential allocation procedure. Much of this controversy has centered on the appropriateness of accessibility measures used in the procedures. Most transportation oriented investigators such as Marble feel, that a significant relationship exists between desired movement patterns and the amount indi­ viduals are willing to bid for specific locations as permanent residential sites.10 Consequently, improvement in predicting residential spatial patterns is believed to be related to improving the method of determining the desired movement patterns of decision makers. In general transporta­ tion oriented investigators agree that the gravity-type model has Q A review of these findings can be found in the following sources: Gerald A.P. Carrothers, "An Historical Review of the Gravity and Potential Concepts of Human Interaction," Journal of the American Institute of Planners, Vol. 22, 1956, pp. 94-102; Walter Isard, £t. a l . , Methods of Regional Analysis: An Introduction to Regional Science (Cambridge: M.I.T. Press, 1960), pp. 493-568; Fred Lukermann and Philip W. Porter, "Gravity and Potential Models in Economic Geography," Annals of the Association of American Geographers, Vol. 50, 1960, pp. 493-504; Gunner Olsson, "Central Place Systems, Spatial Interaction and Stochastic processes," Regional Science Association Papers, Vol. 18, 1966, pp. 13-45. ^ D u a n e F. Marble, "A Theoretical Exploration of Individual Travel Behavior," Quantitative Geography, edited by William L. Garrison and Duane F. Marble (Part I: Economic and Cultural Topics, Evanston: Northwestern University, Department of Geography, 1967), p. 36. See also: John D. Nystuen, "A Theory and Simulation of Intraurban Travel," Quantitative Geography, edited by William L. Garrison and Duane F. Marble (Part I: Economic and Cultural Topics, Evanston: Northwestern University, Department of Geography, 1967), pp. 54-83. * 66 little to do with describing desired movement patterns but argue that it is able to approximate aggregate travel patterns, or at least render an index of the latent accessibility of a location with respect to spatially distributed opportunities. Demographic oriented investigators, although agreeing that resi­ dential choices are related to desired movement patterns, maintain that gravity-type accessibility measures are so intertwined with the spatial system of land uses that it is too much a part of the dependent vari­ able to be treated as an ordinary variable. Chapin and Weiss have argued that complex variables such as the gravity measure of accessibil­ ity tend "to obscure or average out the influence of otherwise dominant variables" relevant to the household allocation p r o c e s s . H Harris, recognizing the importance of other attributes in the choice of a resi­ dential location, suggests that "quite possibly an arbitrary definition of tripmaking could be applied which would produce usable results" for household allocation models. 12 Wolpert claims that the distance-decay exponent in the gravity model is the major factor hindering progress in the development of principles of spatial behavior because it 13 provides little explanation of the actual behavior involved. J Schneider claims that a fundamental procedure more related to human . Stuart Chapin, Jr. and Shirley F. Weiss, Factors Influencing Land Development (Institute for Research in Social Science, Chapel Hill: University of North Carolina Press, 1962), pp. 8-9. 12 Britton Harris, "Note on Residential Location in a Subnucleated Region" (unpublished paper, Institute for Environmental Studies, University of Pennsylvania, Philadelphia, March, 1966), p. 2. 13 r Wolpert, o£. c i t ., p. 159. 67 behavior than the gravity model is n e e d e d . ^ Curry and Rushton suggest that a more fundamental behavioral approach, not tied to the particular spatial system under study, but possessing greater long-run stability than the distance-decay exponents in gravity models is n e e d e d . ^ Stewart and Warntz, the major proponents of the gravity-type measure of accessibility, have themselves long contended that the distance-decay exponent should be unity and the spatially distributed opportunities be manipulated by behavioral fa c t o r s . ^ This investigation suggests that social distance is such a fundamental behavioral factor. Each of these criticisms has generated research endeavors de­ signed to produce more efficient household allocation procedures. While only a few of these are operational or quasi-operational, many of the procedures exhibit 1) locational attributes, disaggregation of either travel behavior or 2) a consideration of additional household locational behavior, and in the case of a few 3) the abandonment of the gravity measure of accessibility in favor of other methods, princi­ pally the opportunity measure. In essence the research endeavors recognize the importance of accessibility to complementary activities as an important attribute in residential choice but are seeking to understand and model attributes that render equally accessible ^^Morton Schneider, "Gravity Models and Trip Distribution Theory," Papers and Proceedings of the Regional Science Association, Vol. 5, 1959, pp. 51-58. ^ C u r r y , _oj>. c i t ., p. 219; Gerard Rushton, "The Scaling of Loca­ tional Preferences," (Research Report, Computer Institute for Social Science Research, Michigan State University, April, 1969), p. 1. ^ W i l l i a m Warntz, "The Topology of Socio-Economic Terrain and Spatial Flows," Regional Science Association Papers, Vol. 17, 1966, p . 48. 68 locations (under present measurement methods) different levels of utility for different residential decision makers. HOUSEHOLD ALLOCATION MODELS The Accessibility Model The first person to use the gravity-type model as an index of accessibility to predict the spatial allocation of households was Hansen in 1959.*^ His model is based on findings by Carroll and Bevis and Voorhees that the journey from employment locations to household locations decays in an orderly manner which can be described for any city by varying the exponential value (b) in the gravity model.1® Figure IV.1 illustrates the influence of distance on the attenuation of trips for different exponential values. Hansen, using this evidence, reasoned that as a city grows new dwelling units would be constructed on vacant land within the most accessible distance to employment opportunities described by the distance-decay exponent. Hansen's household allocation procedure proposes that the proportion of the total urban increase in dwelling units that can be expected to be located in a given residential area is equal to the product of the vacant land in the area times the accessibility of that area divided by the summation of the similar products for all ^Hansen, "Accessibility and Residential Growth." 18 Alan M. Voorhees, "A General Theory of Traffic Movement," Proceedings of the Institute of Traffic Engineers, 1955, pp. 4656; J. Douglas Carroll and Edward W. Bevis, "Predicting Local Travel in Urban Regions," Papers and Proceedings of the Regional Science Association, Vol. 3, 1957, pp. 183-197. Gravity PER CENT OF I ’S DECISION-MAKERS NOT REACHING THEIR N O DESTINATION o o o or impedance a to *n H > Z © O m O model D o O' VO dis trib ution 3 of trip A *« areas in the urban r e g i o n . ^ The generalized form of the model is: i=l where: A H £ = the predicted proportion of total urban growth (in dwelling units) in a given residential area i; H t * the total urban increase in dwelling units; *> the available vacant land in area i; = the accessibility index to employment opportunities for area i; and c = an empirically determined exponent describing the effect of the accessibility index on the allocation of households for the particular urban region under study. Lowry's "Model of Metropolis" allocates households in much the same manner as Hansen's procedure except that Stewart's index of access­ ibility only is used, and households are allocated to residential areas subject to a maximum density constraint (zoning restrictions) which sets a ceiling on the number of households that can be assigned to an area 19 number serves models can be Hansen's model does not contain any constraint to limit the total of household units that can be assigned to an area. The model as a probability function similar to the competing accessibility used in retail trade allocation studies. The probability model stated: 71 during any period thus forcing surplus households to be allocated else­ where.^® The constraint may be expressed: Ht (V± + Dj.) (IV.3) where: H. ■ the total number of dwelling units that can be contained in a residential area i during an iteration period; V = the amount of usable tion period; vacant land in an area i inanitera­ D^ ■= the amount of developed residential land in an area i in an iteration period; and « the number tial space of dwelling units permitted per unit ofresiden­ in area i in an iteration period. Tests of the Hansen model show that the two attributes, accessibil­ ity to employment opportunites and availability of vacant land, do not produce estimates of sufficient reliability for practical purposes. The same may be said for the Lowry model. 21 Despite this, a number of theorists believe that the model possesses several creditable features which can provide a structure for a more accurate allocation proce22 dure. The Regression Model The vast majority of household allocation models developed are linear formulations which estimate residential growth by regression techniques utilizing a number of attributes found relevant to land development as viewed by those making locational decisions. 20 Ira S. Lowry, A Model of Metropolis (Santa Monica: Corporation, August, 1964). The RAND ^Hansen, "Accessibility and Residential Growth," p. 1. 22 Wingo, Transportation and Urban L a n d , pp. 15-16. The 72 procedure is to determine the attributes and their weights which in a linear combination can be related to the amount of growth which has been observed to take place over a past period. The model can be ap­ plied to the individual analysis areas to forecast the magnitude of growth. Such models have the form: A H i » aQ + b ^ (IV. 4) + b2X 2 + . . . + bnXn where: X ^ , X2 , . . . X * the attributes found relevant to the household allocation process; and a , b, , . . . b * proportionality coefficients to be determined o l n . by regression. Lakshmanan, for example, found that the residential growth that could be expected in a given area was a function of the following attributes: 1) the available (vacant) residential land in the area; 2) the accessibility of the area to employment opportunities; 3) the per cent of the area possessing sewer and water services; 4) the value of housing in the area, measured by the product of the median housing value and prevailing density of housing in the area; 5) the prestige level of the area, measured by the income of families and unrelated individuals; 6) the per cent of new houses built in the area during the previous decade; and 7) the per cent of non-whites living in the area. 23 Chapin, WeisB, and others have found essentially the same factors to be significant attributes in the residential decision making process although they argue that greater disaggregation of decision makers and 23 T. R. Lakshmanan, "An Approach to the Analysis of Intraurban Location Applied to the Baltimore Region," Economic Geography, Vol. 40, No. 4, 1964, pp. 348-370. 73 more specific attributes of residential areas are needed to filter out the distinct preferences of location seekers.^ They contend that con­ centration on the desired movement patterns of decision makers as they go about their daily activities will contribute to solving the prob­ lem . Intervening Opportunity Models The evidence that the single exponent gravity-type model of accessibility produced poor estimates of the actual distribution of household growth led to the development of new procedures to determine the accessibility attributes of different residential areas. The procedures are based on a firm conviction that accessibility is the most important attribute considered by decision makers in their choice of a household location but that gravity-type measures are inappro­ priate for determining the level of accessibility of an area. Al­ though more commonly applied to the traffic assignment problem, inter­ vening opportunity measures of accessibility have been used in regres­ sion and other formulations to simulate residential growth. Figure IV.2 illustrates the manner in which trips attenuate over distance using this measure of accessibility. Both the Stouffer and Schneider formulations have been used although the latter has received greater attention.^ The general notion behind both formulations is that the probabil­ ity that a vacant household location opportunity will be chosen by a ^ C h a p i n and Weiss, oj>. cit., p. 9; Chapin and Hightower, op. c i t . , pp. 54-74. ^S a m u e l A. Stouffer, "Intervening Opportunities: A Theory R e ­ lating Mobility and Distance," American Sociological Review. Vol. 5, No. 6 , (December, 1940). Opportunity PER CENT OF I'S DECI SI ON- MAKERS THEI R NOT REACHING DESTINATION N O modil N distribution *n (0 o z -I > -J o < of IN) PI II 3 o trip o attenuations 75 decision maker decreases as some function of the number of inter­ vening household opportunities ranked in order from the central core of the city. As applied to the distribution of residential growth by Hamburg and Lathrop, the Schneider model can be expressed: Ht 1 - H j_x - Ht e -LVi i J "1 ( I V . 5) where: H fc ■ the total urban increase in dwelling units to be allocated; Hj_ 2. * the cumulative number of dwelling units already allocated ranked from the central zone i up to zone j ; Vj_^ » the cumulative number of vacant household sites available for residential purposes ranked from the central zone i up to zone j ; e « the base of the natural logarithm; and L * an empirically determined model parameter expressing the probability that vacant household opportunities will be chosen for residences.26 Consequently the growth (AH^) to be allocated to any one zone is: - . c i t ., pp. 63-87. The procedure renders 99 an index of the attractiveness of an area based on the probability of having a neighbor in a particular occupation category. Such a proce­ dure can serve to clarify the problem of which occupational cate­ gories reveal the greater tendency toward a social-distance space preference. The index of accessibility to social opportunities is expressed by the following constructed variable: ASki - n . > I m > I SOCo 1P ok °J■ (V.9) where: A S ^ ® an index of the potential accessibility of an area i to all decision makers belonging to different occupation categories o based on the propensity for interaction w i t h individuals in the occupation category by a loca­ tion seeker in occupation category k; SOC . «* the number of male decision makers residing in an area J j employed in category o (o = 1 , 2 , . . . m ) ; PQk ■ a reference group identification matrix expressing the probability of a male location seeker in occupa­ tion category k having a male neighbor in employment category o; and ^ij ’ T an<* ^k “ t^e ti“*e-distance, terminal-time, and distance-decay exponent explained previously. THE SPATIAL ALLOCATION MODEL Derivation of the Model The attractiveness of an urban area for residential development is hypothesized to be related to the amount of available vacant land in the area and its proximity to employment opportunities and compatible neighbors for different groups of location seekers .11 To develop 11Through experimentation with the land use allocation model, it has been found that vacant land refers to woodland, cropland, and pasture land as well as unused land. It does not treat unusable land such as water bodies. 100 this proposition into a model which determines the number of new house­ holds which will be allocated to a particular residential area, a sim­ ple probability point of view may be taken. region in which Suppose there is an urban A.Ht new household units are to be allocated. region be divided into numerous residential areas j. Let the Let there also be known the amount of vacant land VAC^ In the urban region, so that each residential area differs in its amount of vacant land. Further, let there be no significant differences among residential areas in the socio-economic status, family life-style, or racial and ethnic prefer­ ences of the resident population. In addition assume that there are no costs involved for the household unit or in interacting with other areas of the urban region either in terms of money, time, or effort. Hence, the friction of distance is zero, and the resident population does not exhibit any difference in their accessibility preferences. On the basis of these assumptions it may be expected that for a representative household unit the probability that it will be located in a particular residential area i will be equal to the ratio VAC^/ VACt, which is the amount of vacant land in the residential area 1 divided by the total amount of vacant land in the urban region. For example, if the total amount of vacant land in the urban region is 1,000,000 acres and that of the residential area 50,000 acres, it may be expected that the housing unit will have a five per cent chance of being allocated to the residential area. However, this reasoning ap­ plies to the probability that a single housing unit will be allocated to the residential area. Since there are A H t new housing units to be allocated in the urban region, the expected number of new housing units P(H^) that will be allocated to a residential area will be A H t 101 times the probability that a single household will be allocated to the"' area so that VAC^ P (H f) - ------------- A H ,. (V .10) VACt In like manner the expected total number of household units that will be allocated to each jth residential area can be estimated. Thus, for the urban region a set of hypothetical allocations based on the availability of vacant land for each residential area can be obtained. The next step is to determine the possible effect of a residential area's accessibility on the number of household units constructed in the area. By dividing the actual number of dwelling units DU^ con­ structed in an area for a given period of time by the expected or probable residential growth in the area P(H^) for the same period of time, a residential development ratio Dt is obtained; that is, DU 4 — = (V.ll) P (H i) Plotting both the development ratio and accessibility index Ai for each residential area on a graph with a logarithmic scale along each axis as illustrated in Figure V.l suggests a linear relationship. A straight line may be fitted to the plotted data by the least squares method. Since the development ratio (the dependent variable) and ac­ cessibility index (the independent variable) are transformed logarith­ mic variables, the equation of the line is: log DUi ------P 1 1 1 O Ci t 2.1 2.0 1.0 1.7 SOCIAL ACCESSIBILITY 1.9 1.4 ASf 2.3 2.1 1.0 1.7 1.0 1.4 EMPLOYMENT ACCESSIBILITY AEj FIG. V I I FIG .VI-2 RELATIONSHIP BETWEEN DEVELOPMENT RATIO RELATIONSHIP BETWEEN AND ACCESSIBILITY TO SOCIAL OPPORTUNITIES ACCESSIBILITY TO EMPLOYMENT OPPORTUNITIES (35 AREAS HAVING AN INCREASE IN DWELLING UNITS) DEVELOPMENT RATIO AND 118 Considering that the inter-correlation coefficient between the two accessibility indices is 0.060 the substantial difference in the partial correlation coefficients requires explanation. The high inter-correlation is due to the presence of a single factor that both indices share in common and which produces an internal consistency between the measures. This factor is the time-distance each residential area is from all other areas in the urban region appearing in the de­ nominator of both indices. An examination of the accessibility measures calculated for each residential area reveals that both in­ dices have their highest values in areas close to the central core of the city and gradually decline toward the periphery. This similarity accounts for the high inter-correlation between the accessibility in­ dices. The gradient of the employment accessibility index, however, is steeper than that of the social accessibility index. This is under­ standable since the major employment areas are located close to the central core, whereas the major residential areas are located inter­ mediately between the central core and the periphery. In the process of o c t a g o n a l izing the two accessibility indices with the development ratios in the multiple regression equation, the difference in the em­ ployment and social accessibility gradients becomes significant. When employment accessibility is allowed to explain all of the variance in the development ratios that it can, a direct partial correlation only regularity that could be found was that in nine out of the ten residential areas in which decision makers belonging to the working class group predominate, negative residuals were obtained. This im­ plies that for working class areas the model can be expected to over­ predict residential growth. Of interest also, although no regularity exists, is that for two upper class areas extremely high positive residuals were obtained. This means that the amount of residential growth allocated to these areas will be highly underestimated. 119 coefficient (0.337) exists between social accessibility and tbe ratios. Relatively speaking residential development is high in areas of high social accessibility and low in areas of low social accessibility. On the other hand, when social accessibility is allowed to explain all of the variance in the development ratios that it can, an inverse partial correlation coefficient (-0.055) exists between employment accessibility and the ratios. Considering that residential develop­ ment in the Lansing - East Lansing region is low in areas of high em­ ployment accessibility, the inverse partial correlation is understand­ able. The reason for the inverse relationship becomes more apparent when areas in which the number of dwelling units declined during the period are considered. Most of the decline occurred in areas having the highest employment accessibility measures. The evidence shows a relationship between residential development and both accessibility in­ dices but suggests the existence of a causal relationship between the more moderate gradient produced by the social accessibility index rather than to the steep gradient generated by the employment accessi­ bility index. Residential Decline Table VI.1 also shows that of the variables tested accessibility to total employment is the best single indicator of the decline in the number of dwelling units in residential areas. The correlation coef­ ficient of 0.851 can be interpreted to mean that accessibility to em­ ployment opportunities explains seventy-two per cent of the variation found in the residential development ratios. Accessibility to social opportunities appears as less of a factor with a coefficient of correla tion equal to 0.769 so as to explain only fifty-nine per cent of the Q. \ 3 O ?>fl.67+7.99 LOO ASj n < £E o o r -8.22+ 5.37Lo»AEi w n 6 >- D E V E L O P M E N T R A T I O D U j / P ( H j ) « Z III z > in a 120 a o _j in O l ■ 1.8 1.7 1.6 i.e 1.3 SOCIAL ACCESSIBILITY 1 . 4 ASj 1.9 1.6 1.6 1.3 I4 EMPLOYMENT ACCESSIBILITY 1.2 AE; FIG.VI 4 F I G . V I - 3 RELATIONSHIP BETWEEN DEVELOPMENT RATIO RELATIONSHIP BETWEEN DEVELOPMENT RATIO ANO AND ACCESSIBILITY TO SOCIAL OPPORTUNITIES ACCESSIBILITY TO EMPLOYMENT OPPORTUNITIES (10 AREAS HAVING A DECREASE IN DWELLING UNITS) 121 variance. Figures VI.3 and VI.4 illustrate the relationship between the development ratios and accessibility to social and employment op­ portunities for ten areas in which there was a decrease in the number of dwelling units during the forecast period. When both accessibility indices are combined in a multiple re­ gression format the coefficient of correlation is raised slightly to 0.871 so as to explain seventy-six per cent of the ratio's variance. The partial correlation coefficients show that both accessibility to employment (0.640) and accessibility to social opportunities (-0.358) contribute to explaining the variance although the contribution made by accessibility to social opportunities is only three and one-half per cent. The significance probability of the F statistics, however, show that both the multiple and partial correlations may be due to chance factors since they are greater than the 0.05 level. In all like­ lihood the small number of residential areas in which there was a net decline in dwelling units contributes to these inconclusive results. None the less, residential areas in the Lansing - East Lansing region that have experienced a net loss in dwelling units are highly accessible to expanding employment opportunities. ANALYSIS OF THE RELATIONSHIP BETWEEN THE PATTERN OF RESIDENTIAL DEVELOPMENT AND THE PREDICTED HOUSEHOLD ASSIGNMENT FOR THE URBAN REGION Three methods were used to evaluate the efficiency of the models to forecast the spatial distribution of household units: 1) a ^ T h e inter-correlation coefficient between the two accessibility indices is 0.964. 70 60 1969 ACTUAL I9 6 0 ACTUAL (O U V /A C R E ) (D U W * C B £ S ) 60 UNIT D W E L L IN G D W E L L IN G 30 20 1969 ACTUAL 1969 E MODEL FORECAST 1 9 6 9 C O M B IN E D 40 AND S O C IA L M O D E L 30 20 10 00 00 0 2 4 6 B 10 0I9TANCE FROM 12 I* C6 D 1C IB 20 22 24 (MINUTES) Fie.VI-9 0 2 4 6 I DISTA N C E 10 FROM 12 14 C9D 16 19 20 22 24 (M IN U T E S ) Fie.vi-e Actual 1969 and I960 dwelling unit density with d i i t a n c a tra m Actual 1969 dwelling unit density with distance tram tha CBO t h a CBO: {.am in g-C o it Lanaing compared with tha 1969 dwelling unit density fo rec ast by tha area FORECAST 122 UNIT D E N S ITY D E N S IT Y 40 SO employment, social end com bin ed a c c e s s ib ility m odels: 123 comparison of the actual and predicted household density distribution profiles for the urban region with distance from the CBD; 2) an analysis of the degree to which the predicted patterns of development deviated from the actual pattern in each residential area; and 3) cor­ relation analysis of the degree of correspondence between the actual and predicted pattern. Household Density Distribution Analysis Figure V I .5 shows a profile of the actual distribution of house­ holds per acre of total usable land in each residential area for 1960 and 1965 with distance from the Lansing - East Lansing Central Business District. During the five year period there was a substantial decline in the number of dwelling units located within five minutes travel time of the CBD. The residential pattern remained stable between five and eight minutes of travel time during the period. The major zone of residential growth was from eight to twelve minutes distance from the CBD with only insignificant increase beyond. Figure V I .6 profiles the 1965 actual distribution of dwelling units compared to the 1965 forecast generated by the employment, social, and combined accessibility models. comparison of the profiles. Two features are apparent from a First, the spatial allocation of households generated by the three models do not differ substantially from each other. As a point of fact the allocations forecast by the social and combined accessibility models are nearly the same for each residential area although the combined model renders a slightly better approxima­ tion of the actual distribution. The employment accessibility model is the least accurate of the three. Second, a comparison of the three predicted distributions with the actual 1960 distribution illustrated 124 TABLE VI.2 percentage deviation of predicted from actual residential DEVELOPMENT IN 45 AREAS FOR THE EMPLOYMENT, SOCIAL, AND COMBINED HOUSEHOLD ALLOCATION MODELS Deviation Ratio Tolerance Limit Spatial Allocation Model +20% +30% +40% +50% +60% Per Cent of Residential Areas Within Deviation Ratio Tolerance Limit .07 .09 .11 .18 CM « Social Model .04 .09 .18 .24 .36 Combined Model .07 .09 .18 .24 .36 o Employment Model TABLE VI.3 RELATIVE SIGNIFICANCE AND EXPLANATORY POWER OF THE EMPLOYMENT, SOCIAL, AND COMBINED HOUSEHOLD ALLOCATION MODELS Std. Er. of Est. F Stat. r«b *0 Sign. Prob. of F St a t . 0.071 176.294 3.279 0.08 N.S. 0.336 0.113 178.215 5.480 0.02 0.336 0.113 177.789 5.475 0.02 Spatial Allocation Model Exp. Employment Model 2.40 0.266 Social Model 4.00 Combined Model 2.40 4.00 Coef. of Cor. (r) Coef. of Det. (r2) 125 in Figure VI.5 shows Chat the forecasts do not differ substantially from the earlier actual pattern. All three models produce a decline in the number of dwelling units within four minutes of travel time from the CBD, maintain the relatively stable residential pattern in the five to eight minute distance zone, and generate growth beyond the eight minute distance zone. Not enough growth, though, is al­ located to the eight to twelve minute zone and not enough decline is generated within five minutes of the CBD. Deviation Analysis The degree to which the predicted household allocation deviated from the actual residential development pattern in each area was calculated to determine the over-all accuracy of each model. The results presented in Table VI.2 show that in only seven per cent of the residential areas are the models accurate within -±-20 per cent of the actual residential development and in only twenty-four per cent of the areas are they accurate within -t-50 per cent. The combined model is slightly more efficient than the social model, and the em­ ployment model is the least efficient. One major conclusion that can be reached at this point is that, at the present time, none of the models is useful for practical purposes. Correlation Analysis This conclusion is borne out by a statistical analysis of the actual and predicted pattern of residential development. The summary found in Table VI.3 shows that both the social and combined models are best able to approximate the actual assignment of dwelling units with a coefficient of correlation equal to 0.336 so as to explain eleven per (IO M f# ri,o p » rO ti*# » , t/o rk tn g C lo ii c ro fH w tn , fo rp **n , s p rv ic * H id d l* C la **,o r No S * v tg o iio n 1 v w p rM r*) (p ro p n *to r» ,» a l» 5 .. . . . o ff ic ia l* , c ip rie o i, m on os* '*) (p rp fp ttio M ll. UPpdf C IO ** to c tin ic d l*) cu FIG VI-7 a c t u a l m o mcoicreo o v e l l i m GMANOE SOCIAL OV THE COH OINED A CC ESSIBILITY l9 S 0 -e 3 ; u n it AND H O OELS, L A N S IN S .C A S T L A N SlftC a r e a a c tu a l-p '*d 'C t*C (10 P o r t ' S , O p * r a t i « * « , u fo 'iu ftg C ia tt c ro fl» *» t« , f o r t » * n , s trtn c * * t r M r s ) :: (p ro p rtfle rs ,3 0 1 *)^ Mi ddl e c n M . i r No s « T r t g i f - o n o fftc to it, c itr ic e l, monoptrf) (p ro tf» l**n o l3 v :; : t*cftfMC«lt| ii! iii;:!;mi ************iiiiiii>1iiiii' FIG. VI 8 ACTUAL AND M CDlCTCO DWELUNS UNIT CHANCE > T CNPLOYNCNT a c c e s s i b i l i t y n o e c L , i » e o - 63 i LAMSINS, CAST LANSINS AREA 4 00 200 0 t QCtuOl- prMictfd o »• 'I L E S 128 cent of the variance found in the actual pattern of development. The employment model produces even less satisfactory results having a coef­ ficient of correlation of 0.226 and an ability to explain only seven per cent of the variance. The household allocations generated by the social and combined models, however, prove to be significant at the 0.02 level, whereas those generated by the employment model fall outside the ac­ cepted 0.05 level. Two conclusions appear valid at this juncture. First, for the urban population as a whole social accessibility weighted by the social distance preferences of decision makers is a more important factor in the household allocation process than is accessibility to employment. Second, factors other than accessibility to social and employment oppor­ tunities, and the amount of land that may be developed for urban use are relevant to the household allocation process. ANALYSIS OF THE RELATIONSHIP BETWEEN THE PATTERN OF RESIDENTIAL DEVELOPMENT AND THE PREDICTED HOUSEHOLD ASSIGNMENT BY SOCIAL CLASS GROUP Correlation analysis was used to evaluate the efficiency of the three models to predict residential development for major social class groups. Although the proportion of the actual residential development variance explained is only significant in the case of residential areas where decision makers belonging to the working class predominate, there is substantial evidence that disaggregation of social class groups will increase the accuracy of household allocation models. Figures VI.7 and V I .8 illustrate the actual and predicted pattern of development generated by each of the models for the urban region. 129 TABLE VI.4 RELATIVE SIGNIFICANCE AND EXPLANATORY POWER OF THE EMPLOYMENT, SOCIAL, AND COMBINED HOUSEHOLD ALLOCATION MODELS BY SOCIAL CLASS OF AREAS Spatial Allocation Model Employment Model Social Model Combined Model Social S tatus of Areas Coef. of Cor. (r) Upper -0.203 Middle Working Std. Err. of Est. F Stat. r*b "0 Sign. Prob. of F Stat. 0.041 187.781 0.344 0.57 N.S. 0.220 0.048 176.112 0.914 0.35 N.S. 0.728 0.529 140.476 14.623 Upper -0.182 0.033 252.047 0.273 0.62 N.S. Middle 0.386 0.149 173.345 3.157 0.09 N.S. Working 0.735 0.540 115.108 15.277 Upper -0.182 0.033 251.071 0.275 0.61 N . S . Middle 0.386 0.149 172.601 3.153 0.09 N.S. Working 0.736 0.541 115.246 15.320 Coef. of Det. (r^) 0.002 0.002 0.002 130 Each residential area Is classified by its major social class group, and while there exist few areas completely dominated by any one group, location quotients were calculated to determine the predominant one residing in each area. For the fifteen areas where the quotients are high for the working class group each of the three models exhibit their highest reliability and are significant at the 0.002 per cent level. Table VI.4 shows that the best result is obtained with the combined ac­ cessibility model although that obtained by the social accessibility model is nearly as efficient. The coefficient of correlation of 0.736 produced by the combined model can be interpreted to mean that for residential areas in which working class decision makers predominate the combined accessibility index and the amount of land at a lowerorder use explain fifty-four per cent of the variance in the actual residential development pattern. The efficiency of the models to predict residential development in twenty middle class areas is substantially lower than their ability to forecast development in working class areas. Still, all the models are slightly more reliable In predicting residential development in middle class areas than they are for the urban region as a whole. For the middle class areas both the social and combined models are able to explain nearly fifteen per cent of the variance in the actual pattern of development, whereas for the urban region as a whole they explain only eleven per cent of the variance. With a significance level of 0.09, though, the correlations of 0.386 are in all probability due to chance factors. In the ten areas where the location quotients were high for upper class groups each of the three models showed an inverse relationship 131 between the patterns of actual and predicted residential development. The strongest inverse relationship is produced by the employment model which obtained a coefficient of correlation equal to -0.203. The in­ verse correlations could be taken to mean that accessibility and the amount of land at a lower-order use are not important factors in upper class decision makers' choice of household locations. pretation is premature for two reasons: Such an inter­ first, the correlations are highly suspect since their significance values are greater than 0.57; and second, the inverse relationship actually attests to the need for disaggregating social class groups to account for their individual locational preferences. The results of the analysis shown in Table VI.4 indicate that the relative importance of accessibility to social and employment opportunities varies with the social status of the decision makers. It must be remembered, however, that the household forecasts are based upon aggregated variables rather than by social class group. That is, the models compute an index of accessibility to total em­ ployment opportunities, a combined index of accessibility to social opportunities, and allocate the total number of household units to residential areas without respect to social class. ANALYSIS OF THE RELATIONSHIP BETWEEN RESIDENTIAL AREA RANKINGS AND THE LOCATIONAL PREFERENCES OF SOCIAL CLASS GROUPS An examination of the rankings of residential areas calculated by the models reveals the influence of locational preferences on the pattern of residential development. It is evident that if the 132 Individual social class group rankings for each area could have been used to allocate dwelling units, the accuracy of the predicted dis­ tribution would be greatly improved. The examination shows that most residential areas are differently ranked with respect to their ac­ cessibility to total employment opportunities and to total social op­ portunities. Only ten of the forty-five areas possess the same rank although the trend of both sets is similar so as to account for the high inter-correlation coefficient between the two Indices. Social-Distance Preferences Table V I .5 shows the existence of a relationship between the average residential area rankings calculated for each social class group and the predominant class of decision makers residing In the area. For the ten areas where the upper class group predominantly resides the average rankings are highest for the upper class (1.4) and lowest for the working class (2.6). The average rankings of fifteen working class residential areas are relatively low for the upper class (2.1), but the same areas have a higher average rank for the working class (1.3). An exception can be found in the average rankings of twenty middle class areas where a different order than expected is obtained. Working class areas have the highest average rank (1.5) rather than middle class areas which are ranked second (1.7). third (2.0). Upper class areas are ranked The range between the average rankings for middle class areas is small, however, and working class areas possess the same average ranking for the middle class group (1.7) as do middle class areas. In all likelihood the closeness of the average rankings is due to the difference in the methods used by Laumann and the U.S. Census 133 TABLE VI.5 AVERAGE RANKING OF RESIDENTIAL AREAS CALCULATED FOR SOCIAL CLASS GROUPS BY SOCIAL STATUS OF A R E A a Social Status of Residential Area Social Class Group Social Status of Ranking Group Upper Middle Working Upper 1.4 2.0 2.1 Middle 2.0 1.7 1.7 Working 2.6 1.5 1.3 ^ o w values have highest rank - lowest value « 3.0, highest value - 1.0. 134 Bureau to classify occupational g r o u p s . ^ It Is of considerable Importance to note that the pattern of average rankings shown In Table VI.5 has essentially the same relation­ ship as the pattern of values found In Laumann's table expressing the probable residential association of occupational groups presented In Chapter Three. It is apparent that the social-distance preference weights used in the model are able to rank residential areas in close accord with those revealed by the different class groups. The model fails as a predictive instrument since for nearly two-thirds of the areas the aggregate total social accessibility rankings are the same as the rankings calculated for the working and middle class groups. This is shown in Table VI.6 along with the fact that in only two areas are the aggregate social rankings the same as calculated for the upper class group. This fact alone helps to explain the poor performance of the model to predict the spatial allocation of dwelling units to upper class areas and its higher reliability in predicting the assignment in middle and working class areas. Employment Accessibility Preferences Even though the model does not rank residential areas according to accessibility to employment opportunities in the occupation cate­ gories associated with the different social class groups, a limited analysis can be made of the influence of employment accessibility ^ I t should be remembered that in the investigation the number of decision makers in an area belonging to a social class group is based on 1960 U.S. Census data, but the social-distance preference weighting is based on Laumann's findings in a sample of Cambridge and Belmont, Massachusetts, residents and that his classification of social class groups differs somewhat from the Census Occupational classifications. 135 TABLE VI.6 RANKED TOTAL EMPLOYMENT ACCESSIBILITY, TOTAL SOCIAL ACCESSIBILITY AND SOCIAL ACCESSIBILITY FOR UPPER, MIDDLE, AND WORKING CLASS GROUPS BY SOCIAL STATUS OF 45 RESIDENTIAL AREAS Area ID 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 Major Social Status Group Middle Working Upper Working Middle Working Working Middle Middle Middle Middle Middle Working Working Middle Middle Middle Upper Middle Middle Working Working Working Middle Upper Upper Upper Upper Upper Working Upper Working Upper Upper Middle Middle Middle Working Ranked Total Employment Access. 20 17 5 2 3 7 18 23 12 9 8 15 1 4 16 14 13 26 24 27 25 32 36 22 21 11 10 6 19 29 33 35 40 34 43 37 30 28 Ranked Total Social Access. 23 24 13 4 3 5 20 27 14 7 8 15 2 6 17 1 12 26 16 21 o < ■—q 28 34 22 19 9 11 10 18 29 36 37 39 35 43 33 31 30 Ranked Upper Class 21 25 15 9 4 8 23 24 16 6 10 17 7 12 18 5 13 28 20 22 27 30 28 19 14 3 2 1 11 26 31 33 36 29 41 35 34 32 Social Accessibility By Class Middle Working Class Class 23 24 13 4 2 5 20 26 14 7 9 15 3 6 17 1 12 27 16 22 25 29 35 21 18 8 11 10 19 28 36 37 38 33 43 34 31 30 23 21 10 4 3 5 18 27 11 8 7 12 2 6 17 1 9 26 16 19 25 28 34 24 20 13 15 14 22 30 37 36 40 35 43 33 31 29 136 TABLE VI.6 (cont'd.) Area ID 39 40 41 42 43 44 45 Major Social Status Group Middle Working Working Working Middle Middle Middle Ranked Total Employment Access. 38 31 41 39 45 42 44 Ranked Total Social Access. 41 32 42 38 45 40 44 Ranked Social Aeees-sibility By Class Upper Middle Working Class Class Class 40 37 43 39 45 42 44 41 32 42 39 45 40 44 41 32 42 38 45 39 44 137 preferences on household locational choices. The conclusions are based upon personal knowledge of the location of employment opportun­ ities In the Lansing - East Lansing Area. The analysis reveals that residential areas that are ranked high In terms of their accessibility to social opportunities for each social class group are also areas In close proximity to employment op­ portunities In the group's occupation category. cited to Illustrate this fact. Several cases may be Table VI.6 shows residential areas ranked by the social accessibility model as first, second, and third for the upper class group are all areas in which upper class decision makers predominate. The three areas are located in East Lansing adjacent to Michigan State University, the major employment area in the urban region for upper class groups. Two of the three residential areas having the highest social rankings for the middle class group contain a concentration of middle class decision makers. All three top ranked areas are in the City of Lansing and either contain or are adjacent to large shopping facilities and office complexes. Four of the six areas having the highest social accessibility rankings for the working class are areas in which working class decision makers pre­ dominate. All six areas are located in Lansing either close to the CBD or close to manufacturing plants. It appears that while decision makers exhibit a preference to reside close to compatible friends and neighbors, many are able to do so and at the same time* reside close to their employment. The analysis aids in explaining the ability of the employment accessibility model to predict residential development in working and middle class areas, and its poor reliability in upper class areas. When a comparison is made between the total employment 138 accessibility rankings and the individual social accessibility rankings for each social class group, it is found that the rankings are approximately similar for the higher ranked middle and working class area. On the other hand there is a wide range between the employ­ ment and social accessibility rankings for areas in which upper class decision makers concentrate. At the same time the employment accessibil­ ity rankings do not correspond with the areas in which the majority of upper class employment opportunities are found. Rather, the rankings closely match areas in which middle and working class opportunities are found. Since the middle and working class employment accessibility preferences are best met in the employment model and the upper class preferences are not met, it can be expected that the employment model should render its best performance in predicting residential growth in middle and working class areas. Disaggregation of the employment and social accessibility components to capture the individual social class group locational preferences should aid in correcting this problem. SUMMARY The analysis of the spatial allocation model finds that urban residential location seekers as a whole exhibit a preference to be accessible to compatible friends and neighbors, more so than they prefer to reside close to employment opportunities. The relationship between social accessibility and the residential development ratios proves to be more significant than the relationship between employment accessibility and the ratios. Similarly, the relationship between the actual pattern of residential development and the pattern predicted by the social accessibility household allocation model is more signifi­ cant than the pattern forecast by the employment accessibility model. 139 These findings are significant since no researcher to this investiga­ tor's knowledge has found accessibility to social opportunities to be more significant than accessibility to employment opportunities in ex­ plaining the spatial distribution of residential land in the urban region. These findings may be interpreted in the light of the "economic competition" and "social choice" hypotheses. Proponents of the "economic competition" hypothesis argue that the different residential locational patterns of the various social class groups are due to their economic stratification given that a high correlation exists between each group's occupation category and median family income. However, such is not the case in the Lansing - East Lansing area since the in­ come levels of the various occupation groups do not differ significant­ ly from each other. "There is a substantial overlap of Income levels for all status groupings, hence income is a weak indicator for characterizlng households in census tracts." 12 It is not unreasonable to claim, as have the Duncans and Feldman and Tilly in other cities, that in the Lansing - East Lansing area blue collar and white collar workers have similar income levels but possess different residential locational patterns. It may be argued that since little economic stratification exists in the study area, it is the social distance between decision makers belonging to different occupation categories that is the important factor in creating the different residential locational patterns of the status groups. 12 , The results of the analysis thus render Tri-County Regional Planning Commission, M.E.T.R.O.: A Gaming Simulation, M.E.T.R.O. Project Technical Report #5, (Lansing, Michigan, January, 1966). See Appendix V - Household Types. 140 support for the "social choice" hypothesis that residential decision makers tend to choose household locations that are close to compatible friends and neighbors. The fact that the spatial allocation models are best able to predict residential development in middle and working class areas and render their poorest forecast for upper class areas attests to the need for disaggregating decision makers belonging to different social class groups in order to capture their particular locational preferences. The analysis finds that the social-distance preference structure used in the social accessibility variable renders a satisfactory ranking of the attractiveness of each residential area for different social class groups. However, the spatial assignment of households is based on a combined social accessibility ranking that is nearly equivalent to the individual rankings calculated for the middle and working class groups. The social model allocates residential development with its poorest reliability to upper class areas because the combined social accessibil­ ity rankings differ substantially from those calculated for the upper class group. The employment accessibility model suffers from the same lack of being able to meet the locational preferences of residential decision m a k e r s . O p e r a t i o n a l i z a t i o n of the subjective accessibility model should rectify this matter to further the understanding of how locational preferences structure the spatial pattern of urban resi­ dential land. ^ A n important feature of the social accessibility model that should be noted is that it requires little exogenously projected data such as is necessary in the employment accessibility model. The em­ ployment model is based on the location of t+1 employment opportunities, but the social model is based on the location of t+0 decision makers belonging to the different social class groups. CHAPTER VII CONCLUSIONS AND PROSPECTS This study has been concerned with the household locational preferences of residential decision makers and the pattern of urban residential growth. In summarizing the investigation it may be profitable to recall the three fundamental questions posed in the in­ troductory chapter: 1) On the macroscopic level what is a good way to describe the locational preferences associated with the residential spatial structure of American cities? 2) Given a particular description of the locational preferences and the residential spatial structure, how adequate are existing household allocation models in utilizing these behavioral principles in explaining this structure? 3) What is the relationship, if any, between the pat­ tern of residential land development and the hypothesized locational preferences of residential decision makers? The summary will not attempt to answer each of these questions in turn. Rather it seeks first to outline the conceptual scheme and the assumptions upon which the research rests, and second, to highlight the more important findings, all of which relate in some manner to these central questions. The central theme of this study has been the investigation of the degree to which the urban residential spatial pattern is structured by the different locational preferences of residential decision makers. Its fundamental purpose has been to identify the relevant locational 141 142 preferences in terms of which the residential spatial pattern can be described and to determine a normative spatial allocation model utiliz­ ing these preferences to assign household units to residential areas of the city. A decision maker's occupational category has been assumed to be a crucial factor influencing his spatial behavior; that is, occupation is regarded from two points of view in this investigation: first, as the principal determinant of a residential decision maker's values, needs, and desires that give rise to his particular set of locational preferences, and second, as a condensed piece of information or cue by which the decision maker perceives social and employment opportunities and ranks them into positions of relative attraction or repulsion based on his set of preferences. It has been further assumed that the dif­ ferent rankings placed on all conceivable spatially distributed op­ portunities by decision makers belonging to different occupation catei gorles give rise to different patterns of spatial behavior by deter­ mining the decision maker's "action space" or mental image of the real urban space. Within this general frame of reference subjective accessibility was hypothesized as playing an exceedingly important role. Subjective accessibility was defined as a measure of the spatial distribution of different types of interaction opportunities about a location adjusted for the ability and desire of decision makers to overcome spatial sepa­ ration. By expressing in mathematical format the objective preference rankings revealed by decision makers belonging to different occupational groups the subjective accessibility measure orders residential areas of the city in terms of their relative attractiveness for the different 143 groups. Areas of high attractiveness are those which provide a given group of decision makers the best chance to facilitate their inter­ action with their preferred set of spatially distributed opportunities. It was hypothesized that decision makers would choose household sites in residential areas offering them maximum attractiveness provided that household sites were available in the areas. Consequently, as the city grows, new household units would be constructed on vacant land in areas providing maximum attractiveness for the different groups. From the macroscopic standpoint the residential spatial structure was investigated to determine the relevant locational preferences manifested by household decision m a k e r s . Operational household allo­ cation mechanisms were investigated to determine the manner in which the revealed locational preferences of decision makers were considered and modeled to distribute new dwelling units to residential areas of the city. Next, a household allocation model which considers the more im­ portant locational preferences was derived. Because of data limitations a modified version of this model was used to determine the attractive­ ness of areas and predict the spatial assignment of households to ihe urban region. The influence of the hypothesized locational preferences on the pattern of residential development was examined, and the dwelling unit assignment predicted by the model was tested against the actual household growth in each residential area. With these general considerations in mind the following is a brief synopsis of the principal findings of the investigation. SUMMARY OF THE PRINCIPAL FINDINGS The first main set of findings concerns the relative impact of the "economic competition" and "social choice" hypotheses on the 144 residential spatial structure. The "economic competition" hypothesis proposes that the different distribution of social class groups in the urban region is due to differences in their budget costs and income resources. The "social choice" hypothesis proposes that the different distribution is due to differences in the values, needs, and desires of the social class groups. It was proposed that "economic competition" is the major process involved in spatially distributing broad categories of land use to various areas of the city. For example, higher-ordered, non-residential land users, because of their higher rent-paying ability, are able to pre-empt desirable locations from lower-ordered users. However, contrary to Alonso, Wingo, and others, it was proposed that "social choice" is the major process determining the residential dis­ tribution of different social class groups in the city. Studies were cited to show that different social classes possessing similar income resources had different residential locations and that groups possess­ ing different income resources had similar travel costs, this latter cost being the major determinant of locational behavior according to the "economic competition" hypothesis. it was proposed that budget costs and income resources are important considerations determining the quality of the housing unit only and that social factors are re­ sponsible for the locational distribution of the social class groups in the city.* An important related finding is that both hypotheses propose that residential as well as non-residential decision makers tend to *The different spatial distribution of social class groups in pre-industrial cities and in cities of "underdeveloped" countries as compared to the distribution found in cities of "developed" countries also supports this view. 145 choose locations that are accessible to their complementary set of In­ teraction opportunities. In other words non-residential decision makers tend to choose locations that are accessible to their consumers, and residential decision makers tend to choose locations which are ac­ cessible to their employment and compatible friends and neighbors. This latter fact, however, has received little attention in the modeling of residential locations. The second major set of findings relates to the analysis of the literature from which generalizations concerning the locational prefer­ ences of residential decision makers were drawn. It was found that the locational behavior of the decision makers revealed a fundamental pref­ erence to reside close to the households of individuals possessing equal or more prestigeous social status than their own, even more so than they preferred to reside close to their employment. However, this t social-distance preference was more obvious with respect to the loca­ tional behavior of the upper class groups, and the work place accessi­ bility preference was more obvious with respect to the locational b e ­ havior of the lower or working class group, young marrieda, and new arrivals to the city. It was argued that social distance, or the at ­ titude of ego that defines the character of the social interaction that location seekers belonging to different social class groups are willing to undertake with decision makers in their own or other social class groups, provides a means of explaining how decision makers evaluate alternative household locations. That is, decision makers choose households in residential areas where they perceive the existence of individuals who are their friends or whom they would like to have as their friends and avoid residential areas in which they perceive 146 Incompatible friends and neighbors. Consequently, for most residential location seekers it does not matter whether their household location is accessible to employment opportunities or not as long as it is in an area possessing individuals with whom the location seeker would like to identify himself. Social distance provides a means of ranking residential areas according to the subjective preferences revealed by decision makers belonging to different social class groups. The analysis of the literature revealed to a much lesser extent that location seekers exhibited a preference for more spacious living conditions. This life-style preference was more obvious with respect to decision makers in the family rearing stages regardless of their social class. To an even lesser extent decision makers revealed a preference to reside close to the households of individuals belonging to similar minority groups. This racial or ethnic preference was found to be more significant in cities where the proportion of ethnics to the total population was generally quite high, and consequently relegated the ethnics to very low status. Although consideration of the latter two preferences is beyond the scope of this investigation, the attributes of alternative household opportunities are evaluated against the four preference structures to determine the part-worth that each attribute contributes to the over-all evaluation of the opportunities. The third main group of findings concerns the adequacy of operational household assignment models in utilizing behavioral principles to allocate residential growth to areas of the urban re­ gion. The models were shown to deal mainly with the gross locational characteristics of the aggregate urban population and consequently, do 147 not provide an adequate basis for residential locational decision making. The models rest heavily upon economic speculations which regard the negative exponential decline of urban land values and population density with distance from the central core of the city to be explained by locational preferences for increased space, better housing, or less-dense housing vis-a-vis the convenience of increased accessibility to employment locations on the part of residential decision makers. Previous chapters have shown that such economic speculations are almost wholly unsupported by empirical studies of residential locational behavior. None of the first generation models consider the basic social phenomenon that different status groups exhibit a preference to segre­ gate themselves both socially and geographically from other status groups. More recent quasi-operational models attempt to solve this problem by disaggregating the residential population into groups of decision makers believed to exhibit similar social goals. While these expanded models reflect an increasing sensitivity to both social processes and decision making behavior, none the less they remain oriented to the "economic competition" hypothesis. No model was found that incorporated a treatment of the different locational be­ havior of status groups from a consideration of the social-distance preferences revealed by decision makers in interacting selectively with decision makers belonging to different status groups. In practice nearly all residential allocation models rely on descriptive methods dealing with aggregates of the urban population or subgroups of the population, are highly transportation oriented (employment - residence locations) and are based on economic behavior. It was also found that 148 the most substantial obstacle to the operationalizing of models is the assembling and organizing of relevant data. Data restrictions in the Lansing - East Lansing area made it impossible to operationalize the subjective accessibility model derived in Chapter Five. Small area data are not available on the number of household units located in each residential area by the occupation cate­ gory of the decision makers residing in them for 1965, or for the num­ ber of employment opportunities existing in each area by their occupa­ tion category for any period of time. An alternative spatial allocation model was used and its efficiency to predict residential development was analyzed. The alternative model calculated an index of attractiveness for each residential area in terms of its accessibility to total employ­ ment opportunities and its accessibility to total social interaction opportunities. The social opportunities in each residential area were weighted by the objective social-distance preferences determined by Laumann for five social class groups and an index of attractiveness was calculated for each area for each group. The total social accessi­ bility index for the area was determined by summing the five separate index rankings. Of the two variables, accessibility to total employ­ ment opportunities and accessibility to total social opportunities, social accessibility was able to explain nearly sixty per cent of the variation found in the ratio of actual to predicted residential growth. Employment accessibility was only able to explain fifty-five per cent of the variation. When the two indices are combined in a multiple regression format, social accessibility was again found to be more significant than employment accessibility. The results provide 149 evidence in support of the "social choice" hypothesis and attest to the validity of the proposition that the social-distance preference is a more important factor in household locational choice than is the accessibility to employment preference, at least for the urban popu­ lation as a whole. On the other hand, a decline in the number of house­ hold units in a residential area was best explained by its accessibility to employment opportunities, a not too surprising outcome. Three different spatial allocation models were used to predict residential development: 1) an employment accessibility model, 2) a social accessibility model, and 3) a combined employment and social accessibility model. The combined accessibility model proved to be slightly more efficient than the social accessibility model, and both were significantly more accurate than the employment accessibility model. However, the accuracy of even the combined model is too poor to be useful for practical purposes. Each of the three models was evaluated in terms of its effi­ ciency to predict household development for residential areas where decision makers belonging to major social class groups predominate. All three models were significantly able to predict residential de­ velopment for working class areas, but substantially less so for the middle and upper class areas. The combined model was again the most reliable allocation mechanism although the social model was nearly as accurate. All three models allocated households to upper class areas in an inverse manner; the highest inverse relationship was produced by the employment accessibility model, again showing that accessibility to employment is a less significant factor in household locational choice. 150 The differing ability of the models to predict the spatial as­ signment of households for different social class groups attests to the need for disaggregating the groups to account for the variation in the manner in which they rank similar residential areas. An analysis of the separate index rankings for each residential area calculated by the social or combined accessibility models revealed a high degree of correspondence between the predominant social class group residing in each area and the manner in which it was relatively ranked by each group. Thus, social-distance preferences appear to be adequately ex­ pressed by the model but must await the availability of additional data for further testing. FUTURE RESEARCH OPPORTUNITIES The subjective accessibility spatial allocation model presented in this investigation is in a developmental stage. Adjustments to the model at this time without additional theoretical refinements and a more adequate data base would appear to have only marginal payoffs. This early version, however, demonstrates an approach to the modeling of residential site selection. Further efforts in its improvement should focus on three main objectives. First and foremost, priate data base. there is a need to develop a more appro­ The 1970 Census provides small area data by residence-work place pairs and household unit data by the occupation category of the decision makers residing in the area. This informa­ tion should make a substantial contribution to improvement of the model. The tabulation of recently collected land use inventory data in the study area should also contribute to its improvement. 151 Secondly, substantial theoretical and methodological refinements are needed to determine the creation and release of vacant land for developmental purposes in the various areas of the urban region. Theoretical refinements must take into account speculative holdings, zoning restrictions, and the deterioration of existing structures. Methodological refinements are needed in the method used to determine a more accurate land use inventory for the forecast period, par­ ticularly the amount of projected higher-ordered, non-residential land use acreage to be allocated to each area prior to the household assign­ ment. The present simple arithmetic method of estimating the 1960 and 1965 land use inventory by using 1955 and 1962 data has proven to be quite inadequate. More recently collected land use data, when it be ­ comes available, will allow more refined regression estimation methods to be used. Thirdly is the problem of determining the rules by which decision makers evaluate and "trade off" the various attributes associated with alternative residential sites in their choice of a household location. Considerably more knowledge is needed about the employment accessibil­ ity and social-distance preferences of decision makers. A major direction for this research to take is to determine the subjective and objective preference structures of decision makers classified according to the Census occupation categories. Such a classification is neces­ sary to be able to cross-classify the preferences with Census compiled information. Laumann's categories are possibly too inconsistent with the Census categories. Within this same objective is a need to know more about the "real" part-worth that the different attributes of residential areas contribute to the over-all evaluation of alternative 152 household opportunities at the same time the location seekers are doing the evaluating, rather than the "felt" part-worth claimed by respondents to questionnaires years after they have made their choices. There is evidence from a survey conducted in the study area that the part-worth assigned to the different attributes change with time after the household site has been selected. This is believed to be a psycho­ logical mechanism whereby decision makers rationalize their household choices to themselves. An additional requirement for organizing this information for use in a spatial allocation model is that it be col­ lected by the occupational category of the responding location seekers. With these objectives realized, one way in which the spatial allocation model can be improved would be to make an initial assign­ ment of household units to residential areas on the basis of their accessibility to employment opportunities in the occupation categories of the location seekers and then allocate or reallocate the household units on the basis of the social distance probability that the location seekers belonging to the various categories interact socially with the inhabitants already residing there. If dwelling units of the type re­ quired by the life-style preferences of the location seekers are not to be found in the area, the household units are further reassigned to the next most accessible area in which can be found the preferred friends and neighbors and the preferred dwelling unit type. In addition the iteration period used in the model may be reduced so that small incre­ ments of growth may be distributed to areas at any one time. Several short forecast periods rather than a single extended period will pro­ duce a different allocation pattern and may produce a more realistic assignment of growth to residential areas of the urban region. APPENDIX A SOCIO-ECONOMIC STATUS TYPES (OR RESIDENTIAL DECISION MAKER TYPES) APPENDIX A SOCIO-ECONOMIC STATUS TYPES (OR RESIDENTIAL DECISION MAKER TYPES) It is essential to determine the locational preferences of resi­ dential decision makers who live within relatively homogeneous areas in order to quantify the subjective accessibility household allocation model. This investigation supports the proposition that these preferen­ ces vary along a dimension characterized by the socio-economic status positions of the decision makers. As many sociologists have noted, the most important unit of social class analysis is the family or the household. Investigators of social class position, such as Barber, point out that all members of a famil­ ial unit share the same socio-economic status as the head of the house­ hold.^- With some modifications sociologists generally agree that the easiest way to classify the socio-economic characteristics of house­ hold decision makers is to utilize the objective indicators containing the essential features that characterize American social class posi­ tion. Lenski, for example, found that there are at least four main indicators that reveal the status position of the decision maker: ^■Bernard Barber, "Family Status, Local Community Status, and Social Stratification: Three Types of Social Ranking," Pacific Sociological Review, Vol. 4 (Spring, 1961), pp. 3-10. 154 education, occupation, Income, and ethnic background. 2 Household type has been substituted for ethnic background In several studies, and since the present study's concern is the spatial allocation of households, this practice will be followed in this investigation.^ Using any one of these status indicators or a combination of the indicators to determine the socio-economic class of a decision maker or to determine the characteristics of an analysis area, rests on ex­ plicit assumptions about each indicator. These assumptions relate to the preferences which govern the manner by which people behave and live their lives. First, the amount of formal education undergone by the head of the household influences his preferences for goods and services, social relationships, and general style of living. Second, the occupation of the head of the household indicates the skill, power, and prestige he is likely to possess because of his level of formal education. Third, the income realized from the decision maker's oc­ cupation determines his ability to purchase the goods and services he prefers. Finally, the household type reflects or mirrors the decision maker's preferences and his ability to satisfy them. ble, conspicuous evidence of his status position. 1c is the tangi­ Holllngshead and Myers found that the social class position of decision makers could be predicted with a probability of .93 by use of a combined index of oc­ cupation and household type. When education was added to the index 2 Gerhard Lenski, "Status Crystallization: A Non-vertical Dimension of Social Status," American Sociological Review, Vol. 19 (August, 1954), pp. 405-464. 3 Laumann, oj>. cit., p. 117; Tri-County Regional Planning Com­ mission, loc. cit. 155 the predictive ability rose to .94.^ Income has been found to be a progressively poorer indicator of social class position as time goes on. Warner found that occupation alone is the best indicator of the socio-economic status of a decision maker, having a predictive probabil­ ity of .91.^ Laumann, likewise, concluded that occupation serves as the most easily obtained attribute by which other people can categorize the social class position of a decision maker.^ For these reasons the present investigation employs occupation as an indicator of the dif­ ferential locational preferences of household location seekers. The stereotype characteristics that pertain to the indicators of the relative socio-economic status of Lansing - East Lansing area house­ hold decision makers are shown in Table A.I. The table is derived from data obtained from the U . S . Census of Population and Housing: 1960 (Final Report PHC 1-73) and was developed by the M.E.T.R.O. staff. In presenting the five category classification the M.E.T.R.O. researchers state that there are too few households in the area that could be identified as "upper class" in the classic sense and consequently this class was included in the next lowest class.^ The table shows that the professional and managerial groups clearly have the highest socio­ economic rank, while semi-skilled operatives and service workers and ^August Hollingshead and Fredrick Redlich, e d s ., Social Class and Mental Illness: A Community Study (New York: John Wiley & S o n s , I nc., 1958). ^Warner, e_t. al. , loc. cit . ^Laumann, o£. c i t ., p. 3. ^Tri-County Regional Planning Commission, loc. c i t . TABLE A-l CHARACTERISTICS OF SOCIO-ECONOMIC STATUS TYPES IN THE LANSING-EAST LANSING AREA - 1960 Socio-Economic Status Group Occupation Upper Class Professional Technical Workers Business Managers Officials Proprietors Clerical Clerical Workers Sales Sales Workers Middle Class Education Over 16- years Value over $20,000 Rent over $150.00/month $10,000-$15,000 16 years Value over $20,000 Rent over $150.00/month $10,000-$15,000 Value - $15,000-$20,000 $10,000-$15,000 12-16 years Rent - $100.00-$149.00/month Craftsmen Skilled W 10 Of H o 00 •r) Household Type \ k Above $15,000 Professional Semi-Profess ional Income Value - $10,000-$15,000 $7,000-$10,000 12 years Foremen Rent - $80.00-$100.00/month Operatives Value - $5,000-$9,900 , c M U a & Semiskilled Nonhousehold Service Workers $4,000-$7,000 8-11 years Rent - $20.00-$59.00/month Laborers Unskilled Value less than $5,000 Less than $4,000 Household Service Workers Less than 8 years Rent Less than $20.00/month 157 unskilled laborers and household service workers are clearly lowest In socio-economic status. This ranking corresponds to schemes used In several studies, and despite the fact that there Is disagreement on the placement of skilled craftsmen, clerical workers and salesmen, this g scheme would probably be agreed on by most social scientists. Q °Otis Dudley Duncan and Beverly Duncan, "Residential Segregation and Occupational Stratification," pp. 493-503; Otis Dudley Duncan, "A Sociometric Index for All Occupations" and "Characteristics of the Sociometric Index," Occupations and Social Status. edited by A.J. Reiss, Jr. (New York: Free Press, 1961), pp. 109-161; National Opinion Research Center, "Jobs and Occupations: A Popular Evaluation," Opinion News. Vol. 9 (September, 1947), pp. 3-13. APPENDIX B BASIC DATA USED IN THE ALLOCATION MODEL BY RESIDENTIAL AREA I TABLE B-l 1960 LAND USE INVENTORY1 Total 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 1692 986 600 518 635 932 1265 1557 709 591 421 693 653 397 398 787 663 1082 1811 957 2924 4934 3732 1136 1003 Vacant 721 237 40 18 29 297 458 885 59 77 48 96 38 27 15 110 33 425 888 240 1022 1048 593 702 410 Unusable 19 0 2 18 10 50 44 23 0 9 0 17 34 19 43 0 7 85 0 2 258 43 0 0 0 Residential Commercial Industrial 495 292 243 209 291 227 332 180 231 162 141 225 96 137 107 354 320 195 552 359 220 561 498 106 274 52 46 27 42 36 53 40 36 18 80 31 10 135 26 2 29 13 1 62 37 15 59 19 3 17 54 46 44 76 4 72 53 43 44 9 1 40 42 58 1 28 41 0 18 12 151 30 5 0 2 Cropland Woodland 10 67 0 0 0 0 65 241 5 0 0 0 0 2 2 5 0 0 0 4 729 2663 2271 261 74 blic 341 298 244 155 255 233 273 149 352 254 200 305 308 128 228 261 249 376 285 303 529 480 346 64 226 TABLE B-l (cont'd.) Area ID Total 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 394 250 2369 419 960 2542 2304 2925 4041 8247 14473 4045 3843 16128 7608 15497 23908 23690 22824 22735 Vacant 43 11 485 64 661 285 116 208 957 250 437 546 219 344 832 450 52 168 409 448 estimated acreage Unusable Residential Commercial Industrial Cropland VToodlan blic 0 0 23 2 1 300 643 236 169 72 142 79 188 104 417 337 2511 556 594 817 239 130 252 248 140 286 172 248 596 235 743 594 238 440 553 416 537 406 588 451 6 22 97 13 3 19 34 6 35 20 44 50 11 5 77 19 16 2 23 9 0 1 59 0 1 54 0 2 13 6 172 4 52 1 55 140 40 1 50 32 0 0 682 2 19 1279 1224 2000 1859 7336 12173 2130 2491 14430 4437 13481 19416 21944 20096 20122 106 86 771 90 135 319 115 225 412 327 761 642 644 804 .237 654 .336 613 l064 856 TABLE B-2 EMPLOYMENT AND SOCIAL OPPORTUNITY INVENTORY AREA ID 1965 EMPLOYMENT Total 1960 MALE RESIDENTS BY OCCUPATION ■ Total Professional & Technical Managers, Officials, Proprietors, Salesmen Clerical Craftsmen 6c Foremen Laborers Operatives 6i Service Workers r, 3498 3010 5837 14856 3087 2859 2773 1549 4301 2771 2682 998 17179 4176 102 2412 1804 198 1297 968 762 976 326 1952 928 1673 1701 2139 1404 1374 438 1758 1083 1338 1640 1201 1248 637 2012 2219 431 1804 1.761 463 1097 1015 381 80 413 113 341 139 52 79 150 91 159 116 114 86 131 239 349 108 205 222 33 61 45 537 114 394 128 495 214 135 148 314 254 288 276 168 198 160 498 624 179 250 274 79 173 113 148 51 137 134 238 92 80 23 143 119 178 171 72 97 55 170 215 24 168 123 26 77 83 483 289 244 296 411 308 415 95 431 235 291 370 155 298 125 459 499 78 492 516 119 364 289 403 394 485 1030 654 651 692 93 720 384 422 707 692 569 166 646 532 42 689 626 206 422 485 160 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 TABLE B-2 (cont'd.) AREA ID 1960 MALE RESIDENTS BY OCCUPATION Total Total Professional & Technical Managers, Officials, Proprietors, Salesmen 123 244 279 1749 14508 296 8 395 100 243 439 232 829 361 194 93 220 269 219 71 194 69 99556 i 529 895 1075 1366 2529 924 285 558 416 654 1091 334 1729 1010 400 516 1115 658 759 591 839 507 50096 104 407 469 518 1088 434 16 151 45 160 339 50 141 120 17 55 83 39 46 55 65 38 8147 175 330 341 213 294 224 28 134 54 186 499 87 335 253 80 114 192 122 113 194 200 140 10321 Clerical Craftsmen & Foremen Laborers Operatives & Service Workers 47 30 75 126 246 46 22 34 40 59 47 17 128 80 15 15 65 22 41 11 64 18 3872 75 33 39 68 151 62 47 111 113 106 96 78 468 247 112 169 364 211 208 126 189 177 10512 128 95 151 441 750 158 172 128 164 143 110 102 657 310 176 163 411 264 351 205 321 134 17244 161 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 TOTAL 1965 EMPLOYMENT TABLE B-3 1960-1965 DWELLING UNIT INVENTORY AND LAND USE CHANGE AREA ID 1960 1965 1960-1965 Change 2013 1184 2006 2929 3855 1945 1698 576 2287 1478 1970 2225 2066 2014 769 2465 2688 555 2086 2128 490 1343 1316 620 2439 1226 2011 2292 3409 1725 1727 1475 2118 1415 2011 2223 1675 1916 785 2513 2696 907 3434 2232 515 1576 1398 480 426 42 5 -637 -446 -220 29 899 -169 -63 41 -2 -391 -98 16 48 8 352 1348 104 25 233 82 -140 LAND USE CHANGE2 Total Ind., Comm., and Industrial Pub. Land Land 57 26 23 35 23 63 48 68 13 -6 24 18 68 9 -1 -2 28 35 85 68 280 270 160 10 2 0 -3 0 -3 8 35 35 -2 0 0 5 8 2 0 10 3 0 2 0 30 0 2 0 Commercial Land 2 5 18 30 13 33 20 25 10 5 7 5 52 9 0 8 5 0 15 25 3 10 5 0 Public Land 53 21 13 5 13 22 -7 8 5 -11 17 8 8 -2 -1 -20 20 35 68 43 247 260 153 10 162 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 DWELLING UNITS1 TABLE B-3 (cont'd.) AREA ID 1960 1965 1960-1965 Change 962 1218 1245 2454 1055 324 732 599 713 1226 405 2069 1170 441 593 1358 826 1025 575 944 592 1310 1269 1460 2535 1365 357 867 580 797 1454 472 2239 1560 482 929 1521 1021 1209 707 1032 690 348 51 215 81 310 33 135 -19 84 228 67 170 410 41 336 163 195 184 132 88 98 LAND USE CHANGE Total Ind., Comm., and Industrial Pub. Land Land 38 0 15 180 10 3 98 54 64 30 59 23 175 93 288 293 130 183 263 205 218 0 0 0 60 0 0 33 0 0 0 0 8 2 35 0 35 18 73 0 10 25 ^U.S. Bureau of the Census 2 Estimated Acreage Calculated from 1955 and 1962 Land Use Inventories Commercial Land 5 2 0 0 0 0 2 0 0 12 0 13 33 3 0 13 5 0 0 0 0 Public Land 33 -2 15 120 10 3 63 54 64 18 59 2 140 55 288 245 107 110 263 195 193 163 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 DWELLING UNITSJ TABLE B-4 LAND USE ZONING1 Residential 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 1422 701 510 357 546 419 1037 1309 581 478 345 596 170 258 398 700 590 1071 1702 843 1848 3947 3284 375 702 371 Commercial 68 59 18 36 70 37 51 62 43 106 76 14 209 36 0 63 20 11 54 67 29 99 0 0 50 8 Industrial 203 227 72 124 19 475 177 93 85 6 0 83 274 105 0 24 53 0 54 48 88 888 448 11 0 0 Public Agricultural Unzoned 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 93 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 968 0 0 738 130 0 0 0 0 0 0 0 0 0 0 0 0 0 11 120 8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 64 Area ID TABLE B-4 (cont'd.) Area ID 172 190 381 951 2313 2142 2720 3710 * 7835 3908 1658 538 1774 1598 1085 6455 474 5021 2501 ^Acres of land Commercial Industrial Public Agricultural Unzoned 73 24 29 96 51 69 88 242 82 5066 121 77 161 2054 0 263 474 1141 227 0 24 0 0 152 92 117 81 330 1158 0 884 1290 0 4184 2630 0 0 2728 5 1990 8 0 25 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 142 0 0 0 0 0 0 0 4342 2265 2344 12902 3956 10228 14584 22742 15749 17279 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 913 0 165 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Residential 166 TABLE 1965 1 AREA ID 17 2 3 IB 19 1 2 XjJLO____ 9 . 9 0 1 . Ou 3 8.20 4 . 5 U .00 6 8 . 8 0 5 . 5 a .80 1 5 5 . 8 0 5 . u 0 7 . 5 u 9 . 3 a .4 0 4 .50 6 7_ . 5 U 1 2 7 8 9 10 11 12 l l T a 0 1 5 _ .5 0 1 0 .j5 0 1 1 .j8 0 ' 1 2 ^ 8 0 1 1 . 0 U 8 . 6 0 P 90 9 . 0 0 7 70 7 .7 0 4 . 4 . 6U .1 U 1 0 .5 0 " 1 3 . 3 0 1 0 . 20 1 1 -J 4 0 7 .2 0 1 ,, 0 0 16 .8012 .6 u 21. 1 6 . 1 0 1 5 . 7 U 1 2 . .9 0 23 1 ?T! ^ t! 12 19 . H016 . 6Uli 25 7 . 3 0 8 80 20 6 50 . 1.00 5 . 9 0 T.TO 4 5 . 6 U 5 . 0 0 7 40 6 . 5 0 6 . 9 0 5 . 3 0 7 . 8 0 9 8 0 1 1 . 70 8.80 8.8 0 8 . 4 0 7 8 ._ 4 0 7 . 2 0 5 . 8 0 6 . 5 0 4 . 3 0 - 1 . 0 0 8 7 J10 5 . 9 0 3 . 4 0 4 . 0 0 1 . 0 0 8 . 6 0 6 . 1 0 4 , 3 0 5 . 5 0 4 , 5 0 . V i) 8.60 9 . 0 0 1 1 8 70 9 , 0 IJ1 7 . 9 0 9 . 6 U 1 2 . 1 . 0011 . 7 0 4 . 6 0 4 . 2 0 .001 ?.9 0 1 0 • 0 0 ‘ 9 . 6 0 . 4 0 1 4 . 3 0 1 1 . 4 0 1 1 . 1 0 1 .00 .50 1.00 6 . 1 0 9 . 5 0 1 2 . 1 U 1 2 . 4 0 1 4 . 5 0 1 5 l. J > 0 72T 5T90-'T.00 . 4 0 9 . 8 0 . 8 0 1 2 . 5 0 1 5 . 3 0 1 7 . 2 0 1 4 . 3 0 1 4 . 0 0 1 2 . , 7 0 1 1 1.00 . 5 0 . 9 . 1 0 7 . ? 0 8 . 4 0 . 9 0 1 5 . 5 0 1 4 , 3 0 1 3 . 0 0 1 1 . 0 0 8 . 7 0 9 . 9 0 7 . 5 0 - 0 . 0 0 . 1 0 1 3 . 5 0 1 0 . 8 0 1 1 . _ 9 0 i n . 60 9 . 4 0 6 . 0 0 7 . 5 0 5 . 4 0 9 . 2 0 9 .9 0 0 . 6 0 0 1 5 9 . 0 0 1 1 . 6 0 1 1 ^ 8 0 1 3 . 8_____ . 3 0 1 1 . 5 0 8 . 7 Q 6 . 9 0 4 . 5 0 6 , 6 0 . 4 0 1 4 . UO 9 ^ 6 011 . 8 0 1 2 . 1 0 1 2 . 7 0 1 0 . 2 0 8 . 4 0 8 . 7 0 . 7H 12 . .2019 .3 0 6 . 9 U 9 , 0 0 6 . 9 0 1 0 . 9 0 13 . 2' U 1 / 9 . . 5 U10. ,50 9 . 7 0 .5 0 . 1U 1011. in .0 0 1 4 .1 uj 2 .5 0 1 3 .4 0 1 . 4 0 1 8 . _0 0 1 3_. 0 0 1 5 . 8 0 1 6 . 0 0 1 4 . 2 0 1 1 . 6 0 1 3 . 0 0 1 0 . 6 0 1 . 0 0 7 . 4CI 7 . 2 0 7 . 4 0 9 . 4 0 9 . 7 II1 0 . 7 0 1 4 . 7 0 .8 0 . 7 0 1 0 . 1 0 1 1 . 7 0 1 1 . 6 0 1 2 . 9 0 • 7 0 _ 5 .3 0 7 . 0 0 0 . 9 0 1 0 . 3 0 1 1 . 9 0 1 1 . 8 0 1 3 . 1 0 6.30 4 .5 0 .00 5 . 1 0 6 . 2 0 7 . 8 0 7 .7 0 . 5 0 6 7 . 7 0 9 . 3 0 1 0 . 7 0 1 0 . 1 0 1 1 . 9 0 9 . 2 0 1 0 . 6 0 0, 1 TOO 8 .10 4 .10 7 . 50 5 . 10 6 . BO J7 ' . 0U 2 . 7 0 9 . 2 0 26 1 1 . 7 0 1 5 27 l‘l .? C i 4 .8016 .. 3 0 1 1 1 3 . 1 0 1 4 . 9 U1 1 ., 8 0 1 1 28 '1 3 ~ . 4 b i ' s ' . T L i 7 , 8 0 1 1 3 . 2 0 1 5 . 0 a l l .. 9 0 1 1 . 1 0 B . / O 8 . H O I D . 90 7 .00 40 4 . 1 0 .50 7 . 6 0 1 3 . '1 0 1 6 . 0 0 1 7 . , 5 0 1 9 . 0 0 9 . 8 0 1 2 . 1 0 1 6 . 1 0 5 .10 90 T . 4 o' . 9 0 1 . 0 0 >20 6.00 . 4 0 9 . 5 0 1 1 . 1 0 1 1 . 7 0 29 31 32 8 . 6 0 8 . 7 0 1 0 . 8 0 . 7 U 1 3 ., 7 0 1 3 . 0 0 1 0 . 4 1 1 1 0 . 6 0 1 2 . 6 0 9 . 6 U 1 2 . 9 0 1 8 . 00 ,, 5 0 1 3 . 7 0 1 2 . 9 0 1 5 . 8 0 1 7 30 .2 0 ?*.9tJ1 $ ."3U 1 2 .20 1 6 .2 0 ~ *?.00 J^4 . 9 0 1 6 1 5 . / 0 1 7 . 6 7 . 4 0 v a l Ty a i 50 " * .. 7 0 1 3 . 3 1) 1 5 . 5 U 1 5 . 4 0 1 7 . 5 0 1 4 0 . 7 0 1 2 . 7 0 • 5 0 1 0 . 8 0 1 0 ., 3 0 8 . 7 U 1 3 . 2 0 1 5 . 1 0 1 2 . , (|U 1 9 .n o 01 3 . 9 0 1U 4 • 3 0 17 . 2 0 1 8 ' 1 3 . 4 0 1 5 . 3 0 1 2 , '1 4 * 6 0 1 7 .5 U 1 V . 16 32 5 . 5 0 .00 * . 3 0 6 . 7 U i . 0 0 1 1 . 3 0 1 2 . 2 0 1 1 .. 7 0 _ 1 8 . II 0 1 4 . 1 U 1 4 ., 3 0 1 ? , 1 0 1 5 . 9 0 1 7 , 6 0 2 0 . 0 0 1 0 “l 15 31 7 . 6 0 10 24 14 3' 60 0 6 7 • Ou . 7 0 J 1.0a 7.70 8 .5u 8 . 1{ S ♦U C 22 , l l .7 0 13 29 00 7 . 5 0 4 . 70 .. 9 0 1 5 . 0 0 1 4 . 4 U 1 6 . 9 0 1 8 . 8 0 1 5 20 12 28 8 . 5 0 1 0 7.001 0 8 3 . 4 0 11 27 8 . 5 0 9 . 5 0 7 . 2 0 10 26 6 . 00 * . 5 U 4 . 1 a 9 . 7 0 1 1 . 2 a 25 5 . 6 0 9 .4 0 6.50 ~ 1.00 _ J i i 5 0 _ 8_. 6 0 5 . 3 0 1 . 0 u 9 24 4 , 2 U 30 .5 U ,1U , .90 *.5 0 8 1 3 . 3 0 1 1 19 .60 5 .5 0 15 16 18 .60 5 .1 0 1 1 . 8 0 . 8. . 9 ______ 7 .4 0 14 7 23 00 .20 1 . 0 0 .4 0 9 6 22 . 30 .7 i) 1 0 . 6 U1 P . 70 . 8 0 4 .9 0 " 1 1 . 6 0 1 3 . 2 0 1 0 5 . riu 9.a011.9u d 13 17 6u .20 A T R A V E L -T IM E . B E TW E E N A R E A S 21 4 20 B .5 9 .5 0 6 .00 u 14, . 5 0 1 4 . 6 0 1 1 . 2 U 1 1 . 4 U 1 3 . 5 0 1 0 9 . 2 0 1 0 . 5 0 .70 TYDO 4 . 2 0 .6 0 .2b y.3(T TTSo .305.20 , 4 0 1 0 . 0 0 7 . 6 0 9 . 6 0 1 1 . 7 0 9 . 3 0 1 1 .4 ,0 . 6 0 3 . 7 0 1 . 0 0 .3 0 1 1 . 3 0 1 2 . 6 0 1 4 . 3 0 1 4 . 0 0 1 5 . 5 0 16 >9010.Sail , 3 0 1 6 . 5 0 1 4 " . 5 0 1 5 . 8 0 1 9 . 7 0 T TSo- 2.v f'T TTTff 6 . 6 0 7 . 2 0 6 . 6 0 1 . 0 0 18.5020 .4ul7 . 5 0 1 7 . 4 0 1 4 . 0 0 1 4 . 2 0 1 6 . 2 0 1 3 .2012 . 8 0 1 0 . 4 0 1 2 . 1 0 1 4 . 0 0 1 5 . 4 0 1 6 . 0 0 1 5 . 6 0 1 5 . 7 0 "lT.1020.Ou?i , 6 5 i p T T T a i T V V u T 7 7 1 0 2 l . t T S u fff. 5 u i £ 7 3 0 1 7 111 8 . 7 0 7 . 2 J d . 3 0 1 5 . 2 J 1 3 . l u 1 2 . 6 0 1 0 . U 0 1 7 . 5 0 1 7 ^ / 0 1 5 ^ 3 0 1 6 . 8 0 j • 4 U 7 0 1 1 . 3 0 1 3 . 9 0 1 3 . 2 0 l . U l l 7 . 4 u i l . 2 . 1U 20 * 2 0 1 ? . 6 0 2 0 . 1 0 2 0 . 0 0 . 701 7 . 8 0 1 9 .5 b S n j 7 4 0 2 3 . 0 0 2 2 . 9 0 3 • U U 9 . 2 J 1 1 . 3 U H JTJ1 5 . 0 0 1 3 . 3 0 1 1 . 7 0 » . e J 1 U . 9 u l 2 . 6 0 1 5 . 3 0 1 P . 1 0 1 5 . 2 0 1 5 . £ 0 1 4 . 2 0 1 3 . 4 0 1 0 . . 7 0 1 4 ; ' 9 u i 6 . 2 . 9 u 1 5 . 8 0 1 7 . 7 0 1 4 . 8 0 1 4 . 5 0 1 3 . 2 0 1 ? . 0 0 7 . 3 0 6 9 . 5 0 9 . 9 0 1 2 . 5 0 1 1 . 7 0 . 9 0 1 6 . 6 0 2 0 7 2 0 1 5 . 6 0 2 0 . 8 0 2 1 . 0 0 1 $ . 3 0 2 0 • 70 20 T S O M 2 7 . 5 0 2 6 . 4 J 2 5 . 8)119 .Ou 6 . 3 U 1 U .O O 1 . 0 IJ 1 1 . 7 0 1 5 . 8 U 1 2 . 7 U 1 ? . B U 8 6 . b O 1 3 . BO 1 0 . 2 0 1 2 . 7 0 1 1 . 1 0 1 4 . 0 0 1 0 . .0 0 2 2 . 7 0 2 2 . 6 0 1 . 0 0 . 9 U 1 3 . 9 0 1 3 . 6 0 1 8 . 3 0 2 1 . 1 0 1 8 . 2 P 1 8 . 6 Q 1 7 . 2 0 1 6 . 5 0 1 3 . 9 0 1 4 8U 7 .4 0 . . 4 0 1 7 . 0 0 1 6 . 3 0 . 5 0 2 3 . 4 Q26 . 0 0 2 5 , 9 0 _ 4 014 . 8 0 1 5 • 0 0 1 3 ^3 014^ . 6 0 1 4 . 8 0 1 3 . 1 . 7 0 1 6 . 5 0 1 4 . 0 1 4 . 5 0 1 4 . 4 0 40 1 9 . 8 0 1 9 , 40 2 1 - 1 0 2 1 . 7 J 2 4 . 4 0 2 3 . 5 U 1 8 . II U 1 8 . 3 1 ) 2 0 . 8 0 1 . 00 " 1 1 . U P 1 5 . 3 U 1 3 . 6 0 1 7 . O u i 3 . 1 O i l . 7 U 1 1 . 2 0 1 6 . 1 0 1 4 . 5 0 1 7 . 1 0 1 5 . 5 0 1 8 . 3 0 1 5 . 2 0 1 7 . 5 0 1 6 . 6 0 1 8 . 7Q 1 9 . / 0 1 9 . 7 J 2 0 . 7 U 2 ? . 2 U 2 1 T 6 U 2 3 . 0 0 ^ 3 ^ 1 0 1 7 . 1 0 1 > . 3 0 l 7 . d 2 3 . 3 0 2 5 . 0 J 2 8 . 7 U 2 7 . 8 U 1 42 4 9 . 9 0 1 6 . 1 0 1 6 . 3 0 1 3 . 4 Q i 6 . 5 0 1 2 . 9 0 l l _ 0 1 5 . 8 0 1 8 . 7 u 2 U . 2 U 1 7 . 8 U l 7 . 2 l i 2 J . 7 b 2 3 41 _ 2 0 2 0 . 3 0 1 7 . 2 0 1 6 . b O lS V i 0 1 4 . 6 b i ' 4 , 9 0 1 5 . 4 0 1 2 . 8 0 1 2 . 9 0 1 1 . 3 0 7 . 9 u ~ 1 3 .2 0 i 2 . 7 u l J . d O 1 7 . 9 0 1 9 . 8 0 1 8 40 3 0 1 2 . 50 13.601 1 0 . 4 0 24 . 1 0 2 1 39 6 0 T 5 *. 5 0 1 3 . 9 0 1 4 , _____ 2 4 . 5 0 2 3 . 3 U 2 4 . v U l P . l U 38 1 o '. l . U U . 4 iiu T '. '5 u i 7 . B 0 1 * . 5 U 2 0 . 5 0 1 6 37 . jj'u l? u . 8 . 5 0 1 9 . 3 0 2 1 . 3 0 l 8 . 2 0 i a 7 . / U1 7 . 6 0 1 9 . 90 . 3 0 1 5 . ' UU1 5 . i u i 6 . 9 ^ 1 0 0 1 6 . 9 0 1 9 . 1 0 2 0 ^ 5 0 * 1 6 . 8 0 2 2 . 5 0 2 1 . 80 1 • 0 0 . _______________________________ 9 0 1 4 . 1 0 1 3 * 7 O i l . 3 0 1 3 . 20 . 6 0 2 3 . 5 J 2 3 . 0 0 2 ^ • 2 U 1 7 . 4 0 1 9 , 5 0 2 3 . 4U 10 .1 0 7 • 5 0 1 0 . * 0 1 0 i 0 . 6 0 i 2 . 2 0 1 6 . 5 0 2 7 > 2 u ? 4 . 3 U 2 4 , 6 0 2 7 . 5 0 1 4 . 9 0 1 7 . 2 0 1015. 0 0 1 6 . 3 0 1 6 . 0 0 1 7 . 2 0 1 9 , 2 0 . 3 0 1 0 • * 0 9 . 8 0 7 . 6 0 1 0 . 0 0 9 . 0 0 lT D O 43 2 9 . 2 0 2 9 . 6 J 2 7 . 6 0 2 5 . O u ? 4 . / U 2 4 . 9 0 2 6 . 9 0 2 3 . 9 0 2 3 , 5 0 1 9 , 3 0 2 1 . 3 0 2 1 . 6 0 2 3 . 1 0 2 4 , 0 0 2 1 . 6 0 2 3 . 6 0 2 5 . 2 0 2 6 . 7 J 2 6 . 3 0 2 1 . 2 ii i 9 7 v G 2 U . 4 'tT 2T T 4 022. 8 0 2 2 . 3 0 2 0 . 1 02 0 . d o S i T . ^ f l i O . 2 0 2 3 . 5 0 2 1 . 5 0 2 1 . 1 0 44 1 6 . 1 0 1 2 . 2 U 1 3 . 7 U 1 3 . 1 U 1 6 . 9 U 1 8 . 6 0 2 1 . 6 0 2 2 . ^ 0 2 0 . 6 0 1 8 . 3 0 1 9 . 0 0 1 7 . 7 0 1 5 . 2 0 l S . 6 d l 7 . 6 O 1 5 . 2 0 2 1 . 3 0 1 7 . 4 (1 1 6 .5 1 4 . 4 0 1 2 .9 01 6 . 6U 3 0 . 6 U 2 5 . 9 0 3 1 . 6 0 3 0 . 2 0 3 4 . 5 0 2 5 . 0 0 l,b2,b3,b4,b5, c, Ht,T READ t+1 total employment for the RA EMPi READ t+0 male population Inventory by occupation category for the RA SOQ, - PROFt,BDSi,CLERi SKOj ,U»SKi _____ READ neighbor occupation soclal-dlstance (matrix) fob PTOTi READ t+0 land uae Inven­ tory for the RA ATOTj .AVACj^ ,AUNUSEj ,AAG. ARE% ,APlq J-l PRINT employment access­ ibility rank for the RA PRINT AEi ORDER the residential areas according to decendlng rank of A^ CiVLCULATE potential social accessibility for all cate­ gories of decision makers 1 ------------- YES CALCULATE areal density constraint for dwelling units In the RA DU601 ZHt - VACi ATOTi-AUMtiSEi CALCULATE t+1 Increase In dwelling units for the RA Hi - V *1C VAC1 y Aic VACt 1-1 READ t+1 tlme-distance of residential area to all other RA's (matrix) »y READ t+0 dwelling unit total for the RA _______ DU60i FLINT total social access­ ibility rank for the RA PRINT ASi SUBJECTIVE ACCESSXBILITT MCLEL Is the amount of vac­ ant land in the RA greater than zero? V A C j O 0_______ y**°** ASki y 4^(“ij+T)bk J-l Ol F1INT social accessibility rink for the RA by category ______ PRINT A S H FLOW CHART GF ONE ITERATION GF Tf£ MXIPTED CALCULATE the decrease in svallable vacant land due to pre-emption by non-resldentlal location seekers In the RA VACi - AVACi-PTOTi "• CALCULATE potential access­ ibility of RA to social opportunities for each (k) citegory of decision makers FIGURE C.l 168 READ t+1 projected nonrealdentlal land uae changes for the RA CALCULATE potential access* ibility of RA to total employment opportunities n V” EMPi 1 “ Z_(Dij+T)to Is the increase In dwelling units greater than the maximum dens­ ity allows? CALCULATE decrease in t+1 dwelling units for the RA due to pre-emption by non. residential location seek­ ers DP60j DHi --VACi ATOTi-AUNUSEi 12 Hi> ZHi CALCULATE potential accessUillty to employment and scd a l opportunltlea for tie RA A1 - AEi+ASi PRINT potential subjective accessibility rank for the RA PRINT A1 PRINT the allocation for the RA PRINT DHit % ALLOCATE surplus dwelling units to next RA Hl+1 - Hi+l+Ok-ZHi) GO TO next residential area 169 COMPUTER PROGRAM OF THE SPATIAL ALLOCATION MODEL (HyrL ► l ( t- U V tH SK'I ? . 0 »•« J G m »;■. f MPU i )O U Ttur,7A F ;l*IN P U T) C illfU , P (JiH (5),C ,li,l , 1rvpE,T0T:M P<45).40H KHf5.45),NElG H ( l ' j . i I . P T ' I I K I b I . A h t * <3 . 4 & ) , T I * ( 4 5 . 4 5 ) , D U 6 0 < 4 b > , A t ( 4 5 ) , A W ( 4 9 , » ) , Z ( 4 5 > 2 , L k * i ji \ ( 4 3 ) Ht»L _ ______ __________________ V5 5 K b A D 1 , ! I Y P F . P O . r t . L . U . T ' It I t o t (1 I ) 9 9 9 9 , 1 0 0 u 1 tUKM AI(1l / 9 f 8.2> 1UCO I t < 1 I Y P E . N f c . 2 ) Rf c AU P . I U l E M P HHJM 3 D 1 .iT y P b .8 0 .K .C ,C ,T 3 0 1 M JM M 4 I i * l L l « s r Ot U P u r l A T A * / * n T T P E n F HUN i s * . l 2 7 * U P A R A M e T b R b * t 9 F 9 1 .2) It l IT Y P E . N E . 2 ) H WI N T 3 0 2 . T O T E P P I M I I Ypb . N e . D H b A U 3 . W O H K R . K E I 3 M I t ( [ I Y P P . Mt . 1 > PHIJT 3 0 3 , HORKR,NEIGh Htf cD 4 , P | O i , * P f c * i *P«0. PH [ M 3 O 4 . PT O T . A H b A _____________________________ _____________________ 2 I U rlM A 1 ( VX . t H > 3 4 (•tJ H rtA il4 5 (1 0 X ,5 t5 /J ,4 l4 0 X ,5 F 5 .3 /)» lC X |5 F 6 .3 ) M i k Ma T < 4 5 < I D A , F S / ) . 4 4 ( i o x , 3 t ' 6 / ) . 1 0 X . 3 F 6 > pHINl 306 bd 10 1 = 1 . 4b p PIwi jnci,] _ __ __________ _______ 10 11 HhAJ; 1 1 . i I | M ( I , J ) , j ' « i , I ) t U Hf-.A I I H x . 1 6 F 4 . 2 ) u0 IP 1=1.45 14 PKJN I J H 7 , iT J H { 1 , J ) , J * 1 A I) Ul> 2 0 1=1.44 _______ 0*1*1 LftJ 2 0 n = j » 4 5 2 0 I I 1< I . 4 ) = TI m ( X , I ) P M m I 3 f ' 7 . ( ( T l M l I . j ) . l * l , 4 5 ) , w« i , 4 5 ) H E A D 2 1 ,1 )1 .1 6 0 k h i n i jnu.Liuon | t | | I Y k F . t J . 2 ) tiU _ 2.0 l U 1' ___________ 3 0 4 ►OHMA T I * 0 P M P L ( I Y P E N I * '. '4 '5 l / ' i , t ' 8 ) )" 3 C J r U * M A I < * i l P 0 9 K E H * . 4 5 < / x , b F 6 ) / * n vF J G H * , 5 ( / X . 5 F 6 . 3 1 1 304 K i H M A l ( » 0 P r o f * , 4 b ( / , f 6 ) / # 0 A P 6 A A , 4 5 < / X , 3 F 8 ) ) 3 0 5 r ORUA I ( » U * / * 0 R E A U U ( . TIKES*) 3 1 ,0 t OWPA I ( « 0 P f c A D h O m * , 1 3 ) 3 0 / t0 h M » IiX .1 6 F 5 .2 ) 3C O t O H M * I ( * 0 * / * 0 D u 6 U * . 4 5 ( / * . F 8 ) )" 21 tOHPA I i 16 x , F 7 > u u JU 1=1.45 t A t <1)=0. UO 30 J = 1 . 4 5 I t I t f c U O A R t | 1 r l ( 1 , J ) ) . Mb . 0 ) G O T C 4 9 9 I t (Lfci.yARt I ) . N E . U U O ^O 899 __ ____ » * T 1h « i , J ) * I ' I t ( L t u v t H I t ) . t t • 0 >GO T u 8 9 9 I t ( t , t I , J . IliO TO 8 9 V J t I L t O . A O I b I1 J . ' N f c . O l G U TC 8 9 9 I- =K 3 0 .41 89V 850 It It II t * 11 11 Ab It (Lfclf/*4(t I.KE.UlUO 10 8 9 9 o t , 0 U , J . ) ( j( J TO. <19 9 I L L G v A H ( 1 'J T t M t ( j ) > . ,vfc . 0 ) GO TO 8 9 9 TU PnP i j > / t ( t_t o o AH ( F j . t- E • u ) j L , I U 899 ( L t o V A 2 < A P ( I ) I , i .fc . 0 ) 0 0 _ T Q 8 9 9 ( 1 ) = AE ( I ) *F i I I Y P t . b o . 3 ) ( jO U> 1 0 0 .41 . OH** A I | . / n A ( * K E O fc H P L O r K f c b T AC Z S S S ( p 11. I T Y * ) CAl L h A M H O f . L H A N . ) GO TO 4 0 0 H « IM 8 9 8 . i , J . F . 8 0 . T . U t * I I , - ) , AE ( I ) . TOTEMP rt(K » |.)*A E J3 M (K ,|> /< T |K < JiL )*T >**B < J> K KIM 11 V H V r M I m A'A I ( • 1 ‘ m j OI A l A C C t - b b l t H Y FT: ? C L A S S E S * ) K -IN I lil) M.MKa I t 5 f 1 P . J ) _ ____________ llC l i l I/O 1 J I I J * l » 4 3 1JU A M J , 1 > * A M J , 1 J + A M J # 1 > I M 1 T YKk . F O . 3 ) u 0 l U 1 5 0 KM I N I 1 1 1 u a l l k a m a i i i a w , l h a .n x > ____________________________________ lib i t 200 I I I MJ MKA I l * l « A M r t K 0 S O C I A L A C C E S S I 3 I T ¥ • > 1 5 U III! 1 6 1 ’ I = 1 » 4 5 ItU A c t 1 ) ! A f I I ) * A W( 1 ( 1 ) K K IM 1A1 l t l M J H K A I l * l R A ‘J K t D C U M H i N f c l . ACC ES 5 I tt 1 L I T T I N D E X * ) CALL K A N M I I A k .L H A N K V ' 2CU K K I M 171 1 7 1 I- I I KKA I i • I h U i i S K h U l D A L L O C A T I r NS * / • AREA*ilOX,«ALLOC.*) 1.0 2 5 0 ||=1.45 I i u NANK4 I i > AHcA ( 2 , I ) = A * b A { 2 i 1 1 f OT < I ) ___________ I M A H t A ( 2 . I ) . G T . U l U C i T. l Z i o U M = AHtr A ( 2 » I > / I A *- 6 A ( 1 1 I ) - A R E A ( 3 . I ) ) » P U 6 0 ( I > K K IM 2 0 1 . 1 . l)H 1 CO 1 U 2 5 u 2 C 1 1 OHKA I ( 1 5 , M 6 . 3 ) 2 1 0 2 l I > * A „ b A < 2 . I ) / t A H t A t l . l ) - A « E A f ? « | ) ) * D U 6 0 < t ) ______ 5>U i A s 0 LO 2 2 ( i J = 1 # 4 3 A *A t(JJ I M I 1 V K K . F U . 2 >A«AM I J > 2«U b U M A = S u M A » A * * C * A K E A ( 2 > J ) A *A t( 1) I M I 1 * W F . F J . 2 >A«AM( I ) __ A « { i * A « * l > A K F A ( 2 # l ) / SUM A * AP I M A . L b . 1 1 1 > > G0 TO AP = A - / I I ) H k I N1 2 0 1 , 1 , H I ) bt rc 2 5 U 24 U Ak = 0 K K IM 2 01 ,1, A 2 5 U CON r 1 IVOC l i l l 1C V I I I/O 1 0 1 = 1 , 4 5 1U b(I>=A(I) K(1)=I tt I =1,44 •J-I * 1 t t 2t L S J . 4 5 i M b l l . ) . LK . « ( I I ) U 0 IK,1K=M<|) | r f c MH = n ( l ) TO 20 ______ ____ b ( L >= n ( I I M LIsM I) h ( 1 > = I F Mw AC I >=1 I F ’ rr U LO l l 1 IvuF KKJI.T JO _____ _ 2U M;KKa I ( . AHI a ACCESS.*) Kk I M 4f, , IK I 1 ) . «< I ) , I = 1 , 45) 4 U M' k KA I I I s . I A.?) Ktr f L)k N tr,a a ______ ______ ___ SELECTED BIBLIOGRAPHY SELECTED BIBLIOGRAPHY I. BOOKS Alonso, William. Location and Land U se: Toward a General Theory of Land R e n t . Cambridge: Harvard University Press, 1965. Barlowe, Raleigh. Land Resource Economics. Prentice-Hall, Inc., 1958. Englewood Cliffs, N. J.: Bartholomew, Harland. Land Uses in American Cities. Harvard University Press, 1955. Chapin, F. Stuart, Jr. Urban Land Use Planning. University of Illinois Press, 1966. Firey, Walter. Land Uses in Central Boston. University Press, 1947. Galbraith, John K. Co., 1958. The Affluent Society. Cambridge: 2nd e d . Urbana: Cambridge: Boston: Harvard Houghton Mifflin Gist, Noel P. and Fava, Sylvia Fleis. Urban Society. New York: Thomas Y. Crowell Company, 1964. 5th edition. Grigsby, W. G. Housing Markets and Public Policy.Philadelphia: University of Pennsylvania Press, 1963. Hawley, Amos. Human Ecology. New York: Ronald Press,1950. Hauser^ Philip M. and Schnore, Leo F., ed s . The Study of Urbanization. New York: John Wiley & Sons, Inc., 1966. Hollingshead, August and Redlich, Fredrick, eds. Social Class and Mental Illness: A Community Study. New York: John Wiley & Sons, I n c ., 1958. Hoyt, Homer. The Structure and Growth of Residential Neighborhoods in American Cities. Washington, D. C.: Federal Housing Adminis­ tration, 1939. Hurd, Richard M. Principles of City Land Values. Record and Guide, 1924. Isard, Walter. Location and Space Economy. Press, 1958. 171 New York: Cambridge: The The M. I. T. 172 Laumann, Edward C. Prestige and Association in an American Community: An Analysis of the Urban Stratification System. Indianapolis: The Bobbs-Merrill Company, Inc., 1966. Lieberman, Stanley. Ethnic Patterns in American Cities. The Free Press of Glencoe, 1963. Loomis, C. P. and Beegle, J. A. Prentice-Hall, 1950. Rural Social Systems. Marshall, Alfred. Principles of Economics. MacMillan Company, 1916. 7th ed. New York: New York: London: Mitchell, Robert B. and Rapkin, Chester. Urban Traffic; A Function of Land U s e . New York: Columbia University Press, 1954. Park, Robert E. 1952. Quinn, James A. Human Communities. Human Ecology. Glencoe, 111.: New York: Ratcliff, Richard U. Urban Land Economics. Book Company, 1949. Free Press, Prentice-Hall, 1950. New York: McGraw-Hill Shevky, Eshref and Bell, Wendell. Social Area Analysis. Calif : Stanford University Press, 1955. Stanford, Shevky, Eshref and Williams, Marilyn. The Social Areas of Los Angeles, Analysis and Typology■ Berkeley and Los Angeles: University of California Press, 1949. Veblen, Thorstein. The Theory of the Leisure Class, An Economic Study of Institutions. New York: The Macmillan Company, cl912. Warner, Lloyd W. , ej:. al. Social Classes in Amer i c a . Harper Torchbooks, 1960. Wingo, Lowdon, Jr. Transportation and Urban L a n d . Resources for the Future, 1961. II. New York: Washington, D. C . : ARTICLES IN BOOKS Bell, Wendell. "Social Areas: Typology of Urban Neighborhoods." Community Structure and Analysis. Edited by Marvin B. Sussman. New York: Thomas Y. Crowell Company, 1959. Burgess, Ernest W. "The Growth of the City." The City. Edited by Robert E. Park, Ernest W. Burgess, and Roderick D. McKensie. Chicago: The University of Chicago Press, 1925. Burgess, Ernest W. "Urban Areas." Chicago: An Experiment in Social Science Research. Edited by T. V. Smith and L. D. White. Chicago: University of Chicago Press, 1929. 173 Duncan, Otis Dudley. "Characteristics of the Socioeconomic Index." Occupations and Social Status. Edited by A. J. Reiss, Jr. New York: Free Press, 1961. Duncan, Otis Dudley. "A Socioeconomic Index for All Occupations." Occupations and Social Status. Edited by A. J. Reiss, Jr. New York: Free Press, 1961. Foley, Donald L. "An Approach to Metropolitan Spatial Structure." Explorations into Urban Structure. Edited by M. M. Webber, J. W. Dyckman, D. L. Foley, A. Z. Guttenberg, W. L. C. Wheaton, and C. B. Wurster. Philadelphia: University of Pennsylvania Press, 1964. Garrison, William L. "Toward a Simulation Model of Urban Growth and Development." Proceedings of the Symposium in Urban Geography, Lund, 1960. Lund Studies in Geography. Lund, Sweden: C. W. K. Gleerup, 1962. Marble, Duane F. "A Theoretical Exploration of Individual Travel Behavior." Quantitative Geography. Edited by William L. Garrison and Duane F. Marble. Part I: Economic and Cultural Topics. Evanston: Northwestern University, Department of Geography, 1967. Nystuen, John D. "A Theory and Simulation of Intraurban Travel." Quantitative Geography. Edited by William L. Garrison and Duane F. Marble. Part I: Economic and Cultural Topics. Evanston: Northwestern University, Department of Geography, 1967. Park, Robert E. "The Urban Community as a Spatial Pattern and a Moral Order." The Urban Community. Edited by Ernest W. Burgess. Chicago: University of Chicago Press, 1926. Riecken, Henry W. and Homans, George C. "Psychological Aspects of Social Structure." Handbook of Social Psychology. Edited by G. Lindzey. Vol. II. Reading, Massachusetts: Addison-Wesley, 1954. Shepard, R. N. "On Subjectively Optimum Selections Among MultiAttribute Alternatives." Decision Mak i n g . Edited by Ward Edwards and Amos Tversky. Harmondsworth, England: Penguin Books Ltd., 1967. Wheaton, William L. C. "Public and Private Agents of Change in Urban Expansion." Explorations into Urban Structure. Edited by M. M. Webber, J. W. Dyckman, D. L. Foley, A. Z. Guttenberg, W. L. C. Wheaton, and C. B. Wurster. Philadelphia: University of Pennsylvania Press, 1964. 174 III. ARTICLES IN PERIODICALS Adams, John S. and Sanders, Ralph. "Urban Residential Structure and the Location of Stress in Ghettos." Earth and Mineral Sciences. (January, 1969), 20-33. Alonso, William. "A Theory of the Urban Land Market." Papers and Proceedings of the Regional Science Association, VI (1960), 149-158. Anderson, Theodore and Egeland, Janice. "Spatial Aspects of Social Area Analysis." American Sociological Review, XXVI (June, 1961), 392-398. Barber, Bernard. "Family Status, Local Community Status and Social Stratification: Three Types of Social Ranking." Pacific Sociological Review, IV (Spring, 1961), 3-10. Berry, Brian J. L. and Rees, Philip H. "The Factorial Ecology of Calcutta." American Journal of Sociology, LXXIV (March, 1969), 445-491. Berry, Brian J. L., Simmons, James W . , and Tennant, Robert J. "Urban Population Densities: Structure and Change." Geographical Review, LII:3 (July, 1963), 389-405. Blumenfeld, Hans. "Are Land Use Patterns Predictable." Journal of the American Institute of Planners, XXV (May, 1959), 61-66. Brown, Robert C. "The Use and Misuse of Distance Variables in Land Use Analysis." The Professional Geographer. XX (September, 1968), 337-340. Burgess, Ernest W. "Residential Succession in American Cities." Annals. CXL (November, 1928), 112. Carroll, J. Douglas and Bevis, Edward W. "Predicting Local Travel in Urban Regions." Papers and Proceedings of the Regional Science Association. Ill (1967), 183-197. Carrothers, Gerald A. P. "An Historical Review of the Gravity and Potential Concepts of Human Interaction." Journal of the American Institute of Planners, XXII (1956), 94-102. Cassetti, Emilio. "Urban Population Density Patterns: An Alternative Explanation." Canadian Geographer, XI:2 (1967), 96-100. Centers, Richard. "Marital Selection and Occupational Strata." American Journal of Sociology, LIV (May, 1949), 530-535. Cressy, Paul F. "Population Succession in Chicago: 1898-1930." American Journal of Sociology, XLIV (July, 1938), 59-68. 175 Curry, Leslie. ’’Central Places in the Random Spatial Economy." Journal of Regional Science, VII:2 (1967), 236. Czamanski, Stanislav. "Effects of Public Investments on Urban Land Values." Journal of the American Institute of Planners, XXXII (July, 1966), 204-217. Duncan, Otis Dudley and Duncan, Beverly. "The Measurement of IntraCity Locational and Residential Patterns." Journal of Regional Science, 11:2 (1960), 37-54. Duncan, Otis Dudley and Duncan, Beverly. "Residential Segregation and Occupational Stratification." American Journal of Sociology, LX (March, 1955), 493-503. Duncan, Otis Dudley and Lieberman, Stanley. "Ethnic Segregation and Assimilation.11 American Journal of Sociology, LXIV (January, 1959), 368-369. Dziewonski, Kasimierz. "A New Approach to Theory and Empirical Analysis of Location." The Regional Science Association Papers, XVI (1965), 17-25. Feldman, Arnold S. and Tilly, Charles. "The Interaction of Social and Physical Space." American Sociological Review, XXV (December, 1960), 877-884. Firey, Walter. "Residential Sectors Re-Examined." XVIII (1950), 451-453. Appraisal Journal, Firey, Walter. "Sentiment and Symbolism as Ecological Variables." American Sociological Review, X (April, 1945), 140-148. Ford, Richard G. "Population Succession in Chicago." American Journal of Sociology. LVI (September, 1950), 156-160. Frieden, Bernard J. "Locational Preferences in the Urban Housing Market." Journal of the American Institute of Planners, XXVII (November, 1961), 316-324. Getis, Arthur. "The Determination of the Location of Retail Activities with the Use of a Map Transformation." Economic Geography, XXXIX (January, 1963), 14-22. Guttenberg, Albert Z. "Urban Structure and Urban Growth." Journal of the American Institute of Planners, XXVI (May, 1960), 104-110. Haig, Robert M. "Towards an Understanding of the Metropolis." Quarterly Journal of Economics, XL (May, 1926), 402-434. Hamburg, J. R. and Lathrop, G. T. "An Opportunity-Accessibility Model for Allocating Regional Growth." Journal of the American Institute of Planners. XXXI (May, 1965), 95-103. 176 Hansen, Walter G. "How Accessibility Shapes Land Use." Journal of the American Institute of Planners, XXV (May, 1959), 73-81. Harris, Britton. "A Note on the Probability of Interaction at a Distance." Journal of Regional Science, V:2 (1964), 31-35. Harris, Chauncy and Ullman, Edward L. "The Nature of Cities." The Annals of the American A cademy of Political and Social Science, CCXLII (November, 1945), 7-17. Herbert, John D. and Stevens, Benjamin H. "A Model for the Distribu­ tion of Residential Activity in Urban Areas." Journal of Regional Science, 11:2 (1960), 21-36. Hoyt, Homer. "Residential Sectors Revisited." XVIII (1950), 445-450. Appraisal Journal, Hunt, T. C. "Occupational Strata and Marriage Selection." Sociological Review, V (August, 1940), 495-504. American Kain, John F. "The Development of Urban Transportation Models." The Regional Science Association Papers, XIV (1964), 147-173. King, Morton B. "Socio-economic Status and Sociometric Choice." Social Forces, XXXIX (March, 1961), 199-206. Lakshmanan, T. R. "An Approach to the Analysis of Intraurban Location Applied to the Baltimore Region." Economic Geography, XL:4 (1964), 348-370. Lenski, Gerhard. "Status Crystallization: A Non-vertical Dimension of Social Status." American Sociological Review, XIX (August, 1954), 405-464. Lieberman, Stanley. "Suburbs and Ethnic Residential Patterns." American Journal of Sociology, LXVII (May, 1962), 673-682. Lukermann, Fred and Porter, Philip W. "Gravity and Potential Models in Economic Geography." Annals of the Association of American Geographers, L (1960), 493-504. Marble, Duane F. "Transport Inputs at Urban Residential Sites." Papers and Proceedings of the Regional Science Association, V (1959), 254. Morrill, Richard L. "Expansion of the Urban Fringe: A Simulation Experiment." The Regional Science Association Papers, XV (1965), 185-199. Morrill, Richard L. "The Negro Ghetto: Problems and Alternatives." Geographical Review. LV:3 (July, 1965). 177 Muth, Richard F. "The Spatial Structure of the Housing Market." Papers and Proceedings of the Regional Science Association, VII (1961), 207-220. Muth, Richard F. "The Variation of Population Density and its Components in South Chicago." The Regional Science Association Papers, XV (1965), 173-183. National Opinion Research Center. "Jobs and Occupations: A Popular Evaluation." Opinion News, IX (September, 1947), 3-13. Olsson, Gunnar, "Central Place Systems, Spatial Interaction and Stochastic Processes." Regional Science Association Papers, XVIII (1966), 13-45. Olsson, Gunnar and Gole, Stephen. "Spatial Theory and Human Behavior." The Regional Science Association, XXI (1968), 229-242. Rodwin, Lloyd. "The Theory of Residential Growth and Structure." Appraisal Journal, XVIII (1950), 295-315. Schneider, Morton. "Gravity Models and Trip Distribution Theory." Papers and Proceedings of the Regional Science Association, V (1959), 51-58. Simmons, James W. "Changing Residence in the City: A Review of Intra Urban Mobility." Geographical Review, LVII (October, 1968), 622-651. Smith, Joel, Form, William H . , and Stone, Gregory P. "Local Intimacy in a Middle-Sized City." American Journal of Sociology, LX (November, 1954), 276-284. Stegman, Michael A. "Accessibility Models and Residential Location." Journal of the American InstH-uLe of Planners, XXXV (January, 1969), 22-29. Stewart, John Q. "Empirical Mathematical Rules Concerning the Distribution and Equilibrium of Population." Geographical Review, XXXVII (1947), 461-483. Stewart, John Q. and Warntz, William. "Macrogeography and Social Science." Geographical Review, XLVIII (April, 1958), 167-184. Stouffer, Samuel A. "Intervening Opportunities: A Theory Relating Mobility and Distance." American Sociological Review, V (December, 1940), 845-867. Ullman, Edward L. "The Nature of Cities Reconsidered." The Regional Science Association Papers and Proceedings, IX (1962), 7-23. Warntz, William. "The Topology of Socio-Economic Terrain and Spatial Flows." Regional Science Association Papers, XVII (1966), 48. 178 Wheeler, James O. "Occupational Status and Work Trips: A Minimum Distance Approach." Social Forces, XLV (1967), 508-515. Wheeler, James 0. "Some Effects of Occupational Status on Work Trips." Journal of Regional Science. IX:1 (April, 1969), 69-77. Wheeler, James 0. "Work-Trip Length and the Ghetto." XLIV (1968), 107-112. Land Economics, Wingo, Lowdon, Jr. "An Economic Model of the Utilization of Urban Land for Residential Purposes." Papers and Proceedings of the Regional Science Association, VII (1961), 191-205. Winsborough, Hal H. "City Growth and City Structure." Regional Science, IV:2 (1962), 35-49. Journal of Wolpert, Julian. "Behavioral Aspects of the Decision to Migrate." Regional Science Association Papers. XV (1965), 159-172. Wolpert, Julian. "The Decision Process in Spatial Context." Annals of the Association of American Geographers, LIV (December, 1964), 537-558. Wolpert, Julian. "Distance and Directional Bias in Inter-Urban Migratory Streams." Annals of the Association of American Geographers. LVII (September, 1967), 605-616. IV. RESEARCH AND TECHNICAL REPORTS Bouchard, Richard J. and Pyers, Clyde E. "Use of Gravity Model for Describing Urban Travel." Highway Research Record No. 88 : Travel Patterns. Washington, D. C . : Highway Research Board, 1965. Brigham, Eugene F. "Some Pitfalls in the Analysis of Residential Locational Preferences." Santa Monica: The RAND Corporation, October, 1963. Butler, Edgar A., jet. _al. "Moving Behavior and Residential Choice: A National Survey." Center for Urban and Regional Studies. Institute for Social Science Research. Chapel Hill: University of North Carolina Press, March, 1968. Chapin, F. Stuart, Jr. and Hightower, Henry. "Household Activity Systems— A Pilot Investigation." Institute for Social Science Research. Chapel Hill: University of North Carolina, 1966. Chapin, F. Stuart, Jr. and Weiss, Shirley F. "Factors Influencing Land Development." Institute for Research in Social Science. Chapel Hill: University of North Carolina, August, 1962. 179 Chapin, F. Stuart, Jr. and Weiss, Shirley. "Some Input Refinements for a Residential Model." Institute for Social Science Research. Chapel Hill: University of North Carolina, 1965. Crecine, John P. "A Dynamic Model of Urban Structure." Monica: The RAND Corporation, 1968. Santa Crecine, John P. "A Time-Oriented Model for Spatial Location." Technical Bulletin No. 6. Pittsburgh: Department of City Planning, January, 1964. Curran, Frank B. and Stegmeier, Joseph T. "Traffic Patterns in 50 Cities." Public Roads— A Journal of Highway Research. Vol. XXX, No. 5. Washington: U. S. Department of Commerce, Bureau of Public Roads, December, 1958. Donnelly, Thomas G . , Chapin, F. Stuart, Jr., and Weiss, Shirley F. "A Probabilistic Model for Residential Growth." Institute for Social Science Research. Chapel Hill: University of North Carolina, 1964. Garrison, William L. "Difficult Decisions in Land Use Model Con­ struction." Highway Research Record N o . 126: Land Use Fore­ casting Concepts. Washington, D. C. : Highway Research Board, 1966. Hamburg, J. R . , Lathrop, G. T . , and Young, G. F. "An OpportunityAccessibility Model for Allocating Regional Growth." Highway Research Record, No. 102. Washington, D. C.: Highway Research Board, 1965. Harris, Britton. "Computers and Urban Planning." Institute for Environmental Studies. Philadelphia: University of Pennsylvania, 1967. Harris, Britton. "Quantitative Models for Uroan Development: Their Role in Metropolitan Policy-Making." Institute fpr Environmental Studies. Philadelphia: University of Pennsylvania, May, 1967. Harris, Britton. "Report of Dartmouth Conference on Urban Development Models." Institute for Environmental Studies. Philadelphia: University of Pennsylvania, 1968. Hemmens, George C. "The Structure of Urban Activity Linkages." Institute for Social Science Research. Chapel Hill: University of North Carolina, 1966. Horton, Frank E. and Reynolds, David R. "Urban Environmental Per­ ception and Individual Travel Behavior." Department of Geography, Special Publication No. 2. Iowa City: University of Iowa, 1968. Irwin, N. A. "Review of Existing Land-Use Forecasting Techniques." Highway Research Record N o . 8 8 : Travel Patterns. Washington, D. C . : Highway Research Board, 1965. 180 Kaln, John F. "Commuting and the Residential Decisions of Chicago and Detroit Central Business District Workers." Santa Monica: The RAND Corporation, April, 1963. Kain, John F. "A Contribution to the Urban Transportation Debate: An Econometric Model of Urban Residential and Travel Behavior." Santa Monica: The RAND Corporation, November, 1962. Kain, John F, "The Journey*to-Work as a Determinant of Residential Location." Santa Monica: The RAND Corporation, December, 1961. Kain, John F. "A Multiple Equation Model of Household Locational and Tripmaking Behavior." Santa Monica: The RAND Corporation, April, 1962. Lamb, Donald D. "Research of Existing Land Use Models." Southwestern Pennsylvania Regional Planning Commission. Pittsburgh: CONSAD Research Corporation, March, 1967. Lowry, Ira S. "A Model of Metropolis." Corporation, August, 1964. Santa Monica: The RAND Lowry, Ira S. "Seven Models of Urban Development: A Structural Comparison." Santa Monica: The RAND Corporation, September, Martin, Brian V. , Memmott, Frederick W . , III, and Bone, Alexander J. "Principles and Techniques of Predicting Future Demand for Urban Area Transportation." M. I. T. Report No. 3. Cambridge: M. I. T. Press, June, 1961. Niedercorn, John H. "An Econometric Model of Metropolitan Employment and Population Growth." Santa Monica; The RAND Corporation, October, 1963. Niedercorn, John H. and Kain, John F. "Suburbanization of Employment and Population 1948-1975." Santa Monica: The RAND Corporation, January, 1963. Reilly, William J. "Methods of Studying Retail Relationships." University of Texas Bulletin, No. 2944, 1929. Rushton, Gerard. "Analysis of Spatial Behavior from Revealed Space Preference." Research Report. Computer Institute for Social Science Research. East Lansing, Mich.: Michigan State University, October, 1967. Rushton, Gerard. "The Scaling of Locational Preferences." Research Report. Computer Institute for Social Science Research. East Lansing, Mich.: Michigan State University, April, 1969. Rushton, Gerard. "Spatial Pattern of Grocery Purchases by the Iowa Rural Population." Studies in Business and Economics. Iowa City: University of Iowa, 1966. 181 Rushton, Gerard. "Temporal Changes in Space Preference Structures." Research Report. Computer Institute for Social Science Research. East Lansing, Mich.: Michigan State University, April, 1969. San Francisco Community Renewal Program. Housing Market." Cambridge, Mass.: January, 1966. "Model of San Francisco Arthur D. Little, Inc., Swerdloff, Carl N. and Stowers, Joseph R. "A Test of Some First Generation Land Use Models." Highway Research Record, N o . 126. Washington, D. C . : Highway Research Board, 1966. Tomazinis, Anthony and Gabbour, Iskandar. "Trip Length Variations Within Urban Areas." Institute of Environmental Studies. Philadelphia: University of Pennsylvania, 1966. Tri-County Regional Planning Commission. "M. E. T. R. 0. — A Gaming Simulation." M. E. T. R. 0. Project Technical Report # 5, January, 1966. Voorhees, Alan M. "A General Theory of Traffic Movement." Pro­ ceedings of the Institute of Traffic Engineers, 1955, 46-56. Weiss, Shirley F., Smith, John E „ , Kaiser, Edward J., and Kenney, Kenneth B. "A Focused View of the Urban Growth Process." Institute for Social Science Research. Chapel Hill: University of North Carolina, 1966. V. UNPUBLISHED MATERIALS Brown, Lawrence A. and Longbrake, David B. "Migration Flows in Intra­ urban Space: Place Utility Considerations." Paper presented at the Annual Meeting of the Association of American Geographers. Washington, D. C . , August, 1968. Brown, Lawrence A. and Moore, Eric G. "Intra-Urban Migration: An Actor Oriented Framework." Unpublished paper, University of Iowa, 1968. Cox, Kevin R. "Acquaintanceship in a Spatial Context: Conceptual Model and Empirical Relationships." Paper presented at the Annual Meeting of the Association of American Geographers, Washington, D. C., August, 1968. Gabbour, Iskandar. "Travel Cost Variations and the Size of Urban Areas." Unpublished Ph. D. dissertation, University of Pennsylvania, 1967. Hansen, Walter G. "Accessibility and Residential Growth." Unpublished MCP thesis, Massachusetts Institute of Technology, 1959. 182 Harris, Britton. "Basic Assumptions for a Simulation of the Urban Residential Housing and Land Market." Unpublished paper, July, 1966. Harris, Britton. "Note on Residential Location in a Subnucleated Region." Unpublished paper. Institute for Environmental Studies. Philadelphia: University of Pennsylvania, March, 1966. Harris, Britton. June, 1966. "Notes on Accessibility." Unpublished paper, Harris, Britton. "Organizing the Use of Models in Metropolitan Planning." Paper presented to a Seminar on Metropolitan Land Use Models, Berkeley, California, March 19-20, 1965. Harris, Britton. "Preliminary— Note on Aspects of Equilibrium in Urban Growth Models." Unpublished paper, August, 1966. Harris, Britton. "Urban Transportation Planning— Philosophy of A p ­ proach." Paper presented at the National Transportation Symposium and Railroad Conference, San Francisco, May 3, 1966. Lee, Douglass B., Jr. "Household Disaggregation in Urban Models." Paper presented at the Annual Meeting of the Regional Science Association, Cambridge, Massachusetts, November, 1968. Moriarty, Barry M. "Human Interaction and Evolving Urban Spatial Structure." Unpublished paper presented at the Annual Meeting of the Association of American Geographers, Washington, D. C. , August, 1968. Abstract in Annals of the Association of American Geographers. LIX (March, 1969), 195. Newling, Bruce E. "The Spatial Variation of Urban Population Densities." Paper presented at the 21st International Geographi­ cal Union Congress, New Delhi, December 1-8, 1968. Tomazinis, Anthony R. "Spatial Parameters Affecting Urban Traffic." Paper presented to the Origin and Destination Survey Committee Meeting of the Highway Research Board, January 9, 1961. Tomazinis, Anthony. "Transportation Inputs of Urban Activities: An Investigation of Urban Location Theories Concerning the Trans­ portation Inputs of Urban Activities." Unpublished Ph. D. dissertation, University of Pennsylvania, 1963. VI. NEWSPAPERS AND PUBLIC DOCUMENTS State Journal (Lansing, Michigan), February 10, 1969. U. S. Department of Commerce. Bureau of the Census. United States Census of Population and Housing: 1960, Lansing, Michigan.