SPATIAL ATTRIBllTES OF SOCIAL AREA DIMENSIONS
IN
~IIA\I1.
FLORIDA S'-fSA: 1970
TboIoas D. Boswell. AEohbi A. Adedibu
and Ki.berly J. Zokoski
Modernization in developed societies results in a siEting and sorting
of urban populations into relatively ~geneous neighborhoods. As Cities
develop, syste.atic relationships between geographical space and social space
e.erle. Three classical .odels have been developed to describe these sortingOUt processes. In 1923 BurieSS suglested that growing cities e~hibit a concentric zone pattern of land use types. Each tone is viewed as e~pandinl its
territory outward by invading the ne~t outer zone.' IlDyt added to the Burgess
IIoOdel by notinl the general presence of B sectoT or wedge pattern. lie stated
that high-class residential areas develop in distinct sectors of the city and
~igrate outward through time fro~ the central business district in segmented
concentric bands. The inner zones are filled in as available housing filters
down to the nellt lower class.' lIarris and UIIllIan added the concept of multiplc
nuclei in which discrete districts occur around several foci in a city rather
than around a single centrnl bUSiness dlstrict. J
~actorial ecology studies conducted during the last twenty years suggest
that these three models are complementary, rather than contradictory. TtpicalIy, a set of ...ocial indicator variables is selected for factor analy... i....
Social area analysis (SAA) studies focus on three composite variables (factors)
th3t usual I)· e..erge from these investilations. The first is regarded;J"~ a .eas_
ure of" social rank (or socio_econoaic status) and is IIoOSt f"requently associated
with variables measuring inco_e, rent, occupational class, and edm:ational lev_
els. The second is a measure of" faAllly status (or f"a.lly cycle stage) and Is
AIOSt often defined by variables representing degree of" f"e.ale participation in
the labor force, fertility, and the prevalence of single or .iltl_fa-Ily dwell_
InJs. The third COAlponent usually suggests ethnlcity (or ethnl~ status) and 15
associated with SCAle indicator of .inority status." E-pirical ~vidence fTOAl
SAA studies sUJgests that the Burless, Hoyt. and Ibrrls and Ull. .n -odels each
describe a separate aspect uf !ocial and reSidential differentiation within
cities. Social rank is said to exhibit a sector or wedge pattetn. fa.ily status a concentric ton~l dis~ributior, and ethniC status a ~Illple nuclei char_
acteristiC. ~
This paper atte.pts to ceteTAline the spatial patterns of social di-enslons f"or the Miaai 901SA. 7 Three hypotheses are tested: (I) Social rank .cst
consistently f"its the sector pattern sUl&ested by Hoyt; (2) fa-ily status is
AIOSt closely associated with the concentric zone pattern as stated by Burgess;
and (3) ethnicity fits best the aultiple nuclei pattern suggested by HarriS and
Ullaan. First, the results of a factor analysis will be discussed. Second,
two-way analyses of variance and T_tests will be used to test the validity of
the three hypotheses. Finally. a distance continuUAI vRriable will be corre_
lated with each of the three social area dieensions as a f"urther test of the
Burless raodel.
Factor Analyaia to Eatabli.h Social Ar8a
~"8ion.
A factor analysis was perfon-ed for 242 census traCts in the Miami SMSA
usinl eilht variables (Table I).' These variables are similar to those most
frequently encountered in SAA studies.' An oblique rotation was used, 50 that
the variable cJusters could be accurately defined." The three factors which
e-erled are not strongly intercorrelated with each other (the strongest
correlation was r . _.IS). Thus, each factor represents an independent social
di-ension of the Mia.i population.
TABLE 1
OBLIQUE FACTOR STRUCTURE MATRIX FOR MIAMI. FLORIDA
Fac~ors
Variables
S~SA41970
(loadings)
I II
11*
Occupa~ion'
.859
.880
.836
Education'
Income l
Fertility"
FLFPR"
OUHS'
Negro'
Cuban'
_.453
-.508
.272
-.710
_.I7l
_ .403
-.032
.108
_. lSI
_ .022
- .133
-.588
.149
-.270
_ .445
.886
.<145
.703
.813
.254
.011
Source: U.S. Bureau of the Census. Cen8UB of Populat£o~ and Hou8£ng: 1970,
Census Tracts, Miami. Final Report PliC(l)-129 (W3shington, D.C.: U.S.
Govemment Printing Office. 1972).
*The signs of the loadings for Factor II have been reversed to facilitate
its interpretation.
I Occupation _ percent of all persons 16 years of age and older employed
as (l) professional, technical and kindred workers, piUS (2) managers and
administrators, except farm, plus (3) sales workers.
• Education _ median schOOl years completed for persons 25 years of age
and older.
~
Income _ median family income.
• Fertility _ child/woman ra~io _ n~ber of children less
age per 1,000 wOmen in ~he 15 ~o 44 year age class.
~han
5 years of
" FLFPR • percent of women 16 years of age and older participating in the
labor force.
• OUHS
_ percent of all housing units in one unit structures.
1 Negro. percent of total population that is Negro.
, Cuban _ percent of ~otal population that is of Cuban stock (born in
Cuba or born in ~he U.S. with at least one parent born in Cuba).
The first factor represents 80ciat rank.
Census tracts strongly
by this factor are represented by high occupational status, high
educational levels. high income, and a low percentage of blacks in their
populations. The second factor is representative of tracts characteri~ed by a
larger than average proportion of Cubans in the population and is therefore a
measure of othnicity.
Because blacks are often characteri~ed by lower incomes
and a generally disadvantaged position. their Characteristics are normally
inversely aSSOCiated with the social rank dimension,
rather than through a
separate factor (as was the case with the Cubans).
characteri~ed
The third factor is broadly interpre~ed as an aspect of {amity 8ta~U8.
It is probably most representative of aspiring middle-class suburban households where wives normally work outside the home. This dimension is characterized by single famity dwelling units, high female labor force participation
rates. and moderate fertility. Most studies have found that female labor
2
force participation is negatively correlated with the family status factor,
with middle_class mothers staying home to care for their young children. II
This was not the case in Miami, however. (SM studies of several other cities
have likewise found the family status factor troublesome to interpret.")
Spatial Analysis of the Social. Al"9a Dimensions
Factor scores for each of the 242 census tract units used in this study
serve as inputs for analyses of variance and T-tests. Parting from tradition,
we subjectively delineated a set of five sectors and five concentric zones
(Fig. I). These will Serve as frames of reference for testing the validity of
the Burgess, Hoyt, and ttarris and Ulllllan models with respect to social rank,
family status, and ethnicity. In most similar studies sectorS and zones have
been constructed geometrically based on a pricPi knowledge. The only exception
of which we are aware is a study of Brisbane, Australia by Timms, in which nine
sectors were defined along the city's major arterial roads. ' l Five sectors
radiating from the city's eBD were defined for this investigation using five of
Miami's main transportation arteries as edges. "
Although 1I0yt envisioned his
sectors as being centered along major trasport routes, familiarity with the
Miami area suggests that these selected routes usually separate, rather than
unify, neighborhoods. IS The selection of these boundaries was based on a visual
interpretation of the maps of factor SCOreS displayed in Figures 2, 3, and 4.
The five concentric zones have been delineated through use of ten minute
isochrones centered about the CBD."lsochrones were used under the assumption
that driving time was more important than linear distance in affecting levels
of interaction with the CBD.
It was hypothesized that Social rank will exhibit a sectoral Or wedge
pattern in ~liami. The distribution of standardized factor scores for this
social area dimension was mapped (Fig. 2) and a two-way analysis of variance
was conducted to test for the appearance of wedge, zone, and multiple nuclei
patterns {Table 2)."
Because an analysis of variance test indicates whether
or not a significant difference exists between sectors and zones in the
aggregate, but does not indicate specifically within which wedges or zones
these differences exist, T-tests for each possible pair of sectors and each
pair of zones was also performed (Table 3)."
Social rank in Miami varies
significantly by both zones and sectors (Table 2). In addition, interaction
TABl.E 2
TWO-WAY ANAl.YSIS Of VARIANCE FOR SOCIAl. RANK IN MIAMl, FLORIDA SMSA-1970
Source of
Variation
Sum of
Squares
Degrees
Mo,.
of
Square
F
Freedom
Zones
Wedges
Interaction
Residual
Total
35.04
44.93
28.62
122.06
242.28
••
"
8.76
11. 23
2.38
221
'"
."
1.00
•• significantly different at a a .05
3
15.86'
20.34·
4.32'
Critical
value of F
at a • . OS
2.37
2.37
1. 15
F
Critical F
6.69
8.58
2.47
TA8LE 3
T_TESTS FOR DIFFERENCES 8ETWEEN ZONES AND WEDGES
FDR SOCIAL RANI!: IN MIAMI, FLORIDA SMSA_1970
~
SOCIAL RANK 8Y ZONES
Zones
Zones
,
2
S
-1.23
0
2
3
•
0
•
Wedges
,
•
S
0
•
N
.'2
."
0
•
N
N
SOCIAL
2
3
Zone
Means
•
3
"""
"
Wedges
2
0
.'6
N
0
.>S
•
S
Wedge
Means
WEDGES
3
0
•
.36
_.58
0
-.21
0
N
0
N
N
N
.
diagonal elements
significant difference 0<,
.os
no' significantly different at II. •
.S<
0
.69
.os
exists between zones and sectors, which also suggests a multiple nuclei
pattern. I' Since the zones and sectors used in this study contain the same
number of degrees of freedom, their F_values can be compared directly. However,
the degrees of freedom for interaction are different. Therefore, the three
calculated F-values have been divided by the critical values of F needed to
achieve a .05 level of significance to obtain a better estimate for purposes of
co~arison.
When this is done it is clear that the level of interaction is
lower than the main effects of the zone and wedge patterns. Furthermore, the
wedge effect is somewhat stronger than the 10nal pattern.
The highest ranking social status wedges are those numbered one, four,
and five, with wedge number five being especially distinctive (Table 3 and Fig.
2). The fifth wedge, extending south of the C80, includes such high land-value
districts as 8rickell. Key Biscayne, Coconut Grove, and Gables Estates. The
wedge with the lowest social rank is number two. which inCludes such depressed
areas as Opa Locka, West Little River, Model City. and River North. 10
•
!!g:~~
~ s S 2 1
ggg:88
N
..
,
'";'
IOOIIJlI
The ~one with the highest social rank is number three. the middle :tone.
Johnston similarly has noted that in Toronto. Ontario. and cities in Australia
and New Zealand there is considerable evidence that the most prestigious
suburbs are not on the urban periphery but rather in intermediate :tones between
the CBO and the fringe.:tl This is also the CaSe in Miami.
~~st studies dealing with the spatial distribution of social status
have found that status exhibits a wedge pattern. Anderson and Egeland found
this to be true in Akron. Dayton. Indianapolis, and Syracuse;:tz as did
Schwirian and ~~tre with eleven Canadian cities zl and Schwirian and Smith with
San Juan, Puerto Rico. 2~ Several investigators of other cities have found that
social rank varies by both zones and wedges. TIlis was found in investigations
dealing with Chicago.'s Rome.'· and Brisbane.'1 Our research shows Miami
followS this pattern.
Ethnioity (Cubans)
Above we hypotheshed that Miami's Cuban population would exhibit a
mUltiple nuclei pattern. Timms in Brisbane found this gattern associated with
ethnicity.'· as did Murdie in Toronto.'· Eichelberger' and Woodbury'l characterized the population of ~liami 's Cubans as being a combination of a nucleated
and wedge patterns.
Both the wedges and the zones in Miami display significant differences
with respect to the Cuban population (Table 4). Cubans arc concentrat<Xi in the
first zone and third wedge (Table 5 and Fig. 3). an area west of the CBO known
as Little Havana. A second node appears in sector two, zones two and three
(the suburban community of Hialeah).
TABLE 4
l1'IO-WAY ANALYSIS OF VARIANCE FOR ETIINICITY
(CUBANS) IN MIAMI, FLORIDA SMSA_1970
Source of
Variation
Zones
Wedges
Interaction
Residual
Total
Sum of
Squares
2B.SO
30.22
41.74
129.97
242.34
Degrees
of
Freedom
Mean
Square
,,
m"
'"
7.12
7.56
3.48
.588
F
12.11*
12.85*
5.91*
Critical
value of F
at a • . 05
2.37
2.37
1.75
F
Critical
F
6.69
5.42
3.38
1.01
*. significantly different. at co • . 05
Family Stat\.ls
Family st.at.us. we hypot.hesized, will exhibit. predominately a concentric
zone pattern. Most studies have found this to be true. TilM>s confirms this in
Brisbane' l2 McElrath in Rome; Il Murdie in Toronto;" and Berry and Horton in
ChiCago. ~s Both zones and wedges display significant differences in Miami (Table
6); however, contrary to the studies just ~entioned, t.he differences bet.ween
wedges is much stronger.
,
TABLE 5
T-TESTS FOR DIFFER£NCI:S BETIl'EEN ZONES AND lIiEDGES
(-'OR ETIINICITY (CUBA.'IlS) IN MIA-'ll, FLORiDA SMSA_1910
Zone
Means
5
Zones
,,,
."
.'6
o
N
.09
•5
.85
."
o
Wedges
,,,
o
.w
1. 22
•5
N
- .02
0
N
N
."
o
.56
dialonal ele.ent$
$ienificant dif"rerence iI't a • . OS
not significantly different lit a • . 05
TABLE 6
TWO-WAY ANALYSIS OF VARIMCE FOR FANllY
STATUS IN MIA.'1I, FlORIDA S/l1SA·1910
Source of"
Varia'tion
"'" .,
Square$
Decrees
.f
....
,
Square
Freeda-
Zones
Wedces
Interaction
Residual
Total
..
19.10
98.54
6.81
119.96
242.92
signiClcantly different
••
4.18
24.64
"
'"
'"
..
.S7
.S<
1.01
••
.os
•
8.80·
45.38'
1.06
.
Critical
value of
••
2.37
2.37
l. 15
,
.os
,
critical
3.11
19.15
.00
,
There Is a gener~l increase in the aean fa-ily st~tus scores for zones
as distance incre~ses away f~ the CBO to zone nu.ber four (Table 7 and Fi~.
4), then there is a decline in the fifth zone. As for the wed~e" there is a
pro~ressive decrease ~vin, counterclockwise f~ the second to first sectoTS.
Very low value, are encountered In the first sector, which is I.r,ely ac_
counted for by the presence of an elderly population livin, in .ulti-faaily
dwellln,s on Ni . .i Beach. The hi,hest values are found in the Middle-class
suburbs of wed,es two and four.
TABLE 7
T-TESTS FOR DIFFERENCES BE1'lI'EEN ZOSES AND WEDGES
FOR FANILY STATIJS IN )1IA.\lI, FLORIDA SNSA-1970
FJI.'II LY STATIJS SY ZOKES
Zones
Zones
2
,
1
2
3
0
,
,
3
,
S
.ss
0
S
0
,
,
.13
.01
0
."
.38
S
Wedge
Means
0
FAMI LY STATIJS BY WEDGES
Wed,es
WedKll5
2
,
S
o
,
1.1'
0
,
N
3
0
1
2
3
Zone
Means
,,
."
.10
0
0
.33
0
.01
dia,onal element
sicnificant difference at 0 • . OS
not sianificantly different at 0 • . OS
Althou,h Burgess illustrated his ecological urban ,rowth .adel through
a set of concentric zones, some of his statements appear to suggest the
presence
a distance continu~ effect.'· In fact, several authors of SAA
studies have operationally defined their distance variable as being scaled
alon, a continuum, rather than discrete zones.'7 To test for the effects of a
distance continuum, straight line distances" between the ceo and each of the
census tracts were corTelated with the factor scores for social rank, ethnicity,
and faMily status. All three coefficients were statistically significant at
the .OS level;" but this seeMS to have been due ~re to the lar,e number of
census tract units used rathoT than truly Meaningful relationships. The only
notable exception of the three was the correlation coefficient between distance'
and ethniclty (r ••. 449). Thus, approxiMately 20.1\ of the vari.bility oT the
Cuban distribution is inversely related to distance frOM the CSO, which reconfirMS the strenfth of the concentration in the area surround in, and includin, Little Havana . •
or
,
Sttttrnary arid COl1CtuawnB
The factor analysis of eigh~ variables for 242 census trac~s in MisaU
produced three factors. These are tenerally consis~ent ~lth the three social
area analysis diaensions of SOCial rank, ethnicity (In this case Cubans), and
.iddle-Class fnaily status (with hith fe.ale labor force participation rates).
The spatial attributes of the three SAA di.ensions are less clear.
Althouth social rank exhibits significant differences between both ~ones and
wedtes. as well as interaction effects, it is .est clearly represented by a
wedge pattern, as suggested by l~yt. This supports the hypothesi~ed pattern
and is si.ilar to the findints of .est SAA studies in other cities. Fa.ily
status. on the other hand, does not reflect a hypothesi~ed ~onal tendency.
Instead, it exhibits a strong wedge pattern. unlike the results of ~st other
social area invcstigations. This exceptional situation is related, at least in
part. to the historical developacnt of Miami Beach as a co-bination resort and
retire.ent co.plex. The distribution of Cubans (ethnlclty) was hypothesi~ed to
exhibit a .ultiple nuclei pattern. While the results do not totally refute
this hypothesis. they do indicate a so.ewhat more cQBIpllcated cOlll>ination of
wedge, ~one, and nuclei co~ponents. This reflects the concentration of Cubans
in two nodes (Little flavana and Ili:lleah) and :I tendency for this population to
diffuse outward toward the western periphery within sectors.
The notion of a diStance continuum effect on the throe SAA dimensions
ul50 WIIS teste<l. Although there is a statistically significant correlation in
each case, only ethnicity is interpreted as being meaningful. There Is a
11IOder:lte inverlle relationship between the degree of concentr:>tion of the Cut-an
l>opuilltion and disttmce frOll Mlallli's central business district. Thus, not only
IS there a difference in the distribution of Cubans llmong zones (as indicated
by th~ analysis of variance test). but this difference Is a result of a decline
in concentration along a continuu. liS distances fro~ the CRO are increased.
fven though there arc also differences a.ong zones for both social rank and
fa.lly status. a distance continuu. effect waS not found to exist.
Although .any SAA studies have preceeded this one. this Investigation
.akes three contributions. First, the zones and sectors have been delineated
less arbitrarily than in other studies. which have generally utilized siaple
geo.etr;c shapes. Major transportation arteries and isachrones appear to be
.est tonsistent with the distributions displayed in the three factor score .aps.
These have been selected as boundaries. This ..y be an i.portant consideration.
since the way ~Ones and sectorS are drawn can effect vhether significant zonal
or sectoral differences exist. Perhaps this Is why both significant wedge and
zone patterns have been found for all three SAA di.enslons, and a .ultiple
nuclei distribution for two out of the three. This suggests that a strict.
~tually eXClusive, interpretation of the zone. ~edge. and ~Itiple nuclei
hypotheses Is InappTOpriate. Rather, ele.ents of all three can exist at the
sa-e ti.e for a particular di.ension. Ho~ever. a5 in the case of ~ia.i. one
pattern ..y be .ere prevalent than the other two with respect to a given
di..ensioa.
Second, the reliability of social area analysis and the tone. sector.
and lIIultiple nuclei lIIOdels h:>ve been tested in the context of a city ~here they
have not been comprehellslvely applied before. In fact, there has been only one
other SAA study that has de:>lt with any cities in the Southeastern United
States, and that inves!,.iiption used I\)SO census da!"a, The study establiShed
social lIrea dimensions through usc of f;)ctor analysis, but It did not mal' or
analyze the spatial attributes of these dimensions. ~1
A third contribution of this study is in testing the effects of both a
distance continuum and discrete concentric zones. Oiscr~te ~ones are more
appropriate for social rank and family Status, we found, whereas a distance
contlnuu. effect is .ere operative in the residential patterns of Cubans.
"
Burgess, "The Growth of the City," Proceedings of thB Ame,..ican
Society 18 (1923): 85_89, and E. W. Burgess, "The Growth of the
City," in ThB City, cd. R. E. Park, E. W. Burgess, and R. O. McKen~ie (Chicago:
University of Chicago Press, 1925), Pl'. 47-62.
I.
E. W.
Sociok>!lica~
2.
11. Boyt:, The StpuctUJ"6 and Gra",th of Residential Neighborhocds in AmeriCIan
Cities (Washington, O. C.: Federal Housing Administration, 1939).
3.
C. lJ. Harris and E. L. Ullman, "The Nature of Cities," Annaltl of
American Academy of Political SCienctl 242 (1945): 7_17.
4.
trn.
For the best review of this body ot' literature sec O. Timms, The Urban
Mosaic: Tololal'ds a Theory of Residential DiffsNJntiation (Cambridge: Cambridge
University Press,
1971).
5.
For an elaboration of the details of social area analysis see: E. Shevky
and M. Williams, The Social Araaa of Los Angeles: Anal!lsis and Typology
(I.os Angeles: University of Californi.a Press, 1949); E. She"ky and W. Bell,
Social Araa Analyais' TI16o>'y, Illuatrativ8 Application and Computational
ProceduNJs (Stanford California: Stanford University Press, 1955; and TillVllS,
Urban Mosaic, Pl'. J23-210.
6.
R. A. Murdie, "Th" Social Geography of the Cit:y: Theoretical and Empir;cal
Background," in .I.. ternal Stl~<ctw'8 of the City: ReadiO'lgs 00'1 Space and
EmJiJ"Q>111l6O'1t, cd. L. S. Bourne (New York; Oxford University Press, 1971),
279-90.
7. This paper represents a partial sequel to an artiCle which the two senior
auth01's published earlier. T. lJ. Iloswell and A. A. Adedibu, "The Social
Structure ot' Nine Florida Sl-lSA's for 1970," The FZori<kl Geogl'aphe .., 12 (1978):
1_8.
8.
In Actu:.1 t'act there were 115 census tracts, but: many were split into 2 or
more parts.
Thu'S. i t was possible to derive variables t'or a total of 244 areal
units.
We let't out two census tracts (numbers 114 and I IS) which were located
in predominantly rural areas.
9.
One of the main advantages of the SAA technique is that compared to more
general t'actorial ecologies it provides a standard set of variables whi.ch
facilitates comparisons. Timms, Urban Mosaic, p. 206.
10.
N. II. Nie et al., Statistical Package fo", the Social Sciences (New York:
McGraw-IIill Book Col, 1975), Pl'. 468-514.
11.
See Timms, U..ban
MQ8ai~,
PI'. 148-176.
12.
For instance see K. P. Schwirian and M. ~latre. "The Ecolo~ical Structure
of Canadian Cities," in Comparative Ul"ban Stpucture: Studi"a ;... the Eaology of
Citiea cd. K. P. Schwirian (Lexington, Mass.: D. C. lIeat:h and Co., 1974). Pl'.
309-323; and T. R. Anderson and L. L. Bean, "The Shevky-Bell Social Areas,
Conn rmat ion ot' Resul ts and a Reinterl'retat ion," Social Forces. 40 (1961):
119_24.
13.
Timms, Urban Mosaic, PI'. 240-41.
14. These nve arteries arc: (I) the Miami River and its southeastern
extension into Biscayne Bay just: north of Virginia Key, (2) U. S. ltighway
South (South Dixie Ilighway), (3) U. S. Highway 41 (Tamiami Trail and S.W. 8th
Street), ('I) U. S. Highway 27 (Okeechobee Road) and the ~ljami Canal and (5)
Interstate Highway 95.
IS.
1I0yt, StPUCtlll"6 and Growth, Pl'.
117-119.
"
16. The isochrones were taken from Metropolitan Oade County Planning
Oepartment, Profile of Metropolitan Dade County: Conditions and Needs
(~iami, Florida: Community Improvement Progr~, Office of the County ~anager,
October, 1972), p. 61.
17. A .OS level of significance has been arbitrarily selected for the
statistical tests used in this paper. The type of two-way analysis of variance
used is the classical experimental model ANOVA developed for unequal cell
frequencies by Nie, et al., Statistical Package, pp. 398-433.
t8. Two-tailed T_tests were utilized. An F-test was conducted in each case
as a test for unequal sample variances.
If the sample variances were judged
not to be significatntly different, a pooled variance estimate procedure was
used.
If significant differences in the sample variances did exist, a
separate variance estimate technique was used. Ibid, pp. 267_75.
19. Murdie states that a significant level of interaction can be used as a
test of the multiple nuclei concept. R. A. Murdie, Factorial Ecology of
MstropoUtan Toronto. 19S1-l96l:
An Essay on thB Soaial Geography of ths City.
University of Chicago, Department of Geography, Research Paper, No. 116
(Chicago, 1969), p. ISS.
20.
Metropolitan Dade County Planning Department, Profils, p. 102.
21. R. J. Johnston. "On Spatial Patterns in the Residential Structure of
Cities," Canadian Geog7'aPher 24 (1970): 361_67.
22. T. R. Anderson and J. A. Egeland. "Spatial Aspects of Social Area
Analysis," AmBl""ican Soaio4Jgical RevionJ 26 (1961); 392_99.
23.
Schwirian and Matre, "Ecological Structure."
24.
K. P. Schwirian and R. Smith, "Primacy, Modernization. and Urban
Structure; The Ecology of Puerto Rican Cities," in Compa:rative Urban StructuN:
Studisa in the Eco4Jgy of Cities, ed K. P. Shwirian (Lexington, ~~ss.; D. C.
Ileath and Co., 1974), pp. 324-38.
25.
B. J. L. Berry and f. E. Horton, Geographic Perspectives on Urban Systems:
With Intergrat6d Readings (Englewood Cliffs, New Jersey: Prentice_Hall, Inc.,
1970), pp. 369_375.
26.
D. C.
~lcElrath,
''Thc Social Areas of Romc; A Comparatoive Analysis,"
Amel"'ican Sociological Review 27 (1962):
376~91.
27.
Timms, Urban HOsaic,
2B.
Ibid., p. 241_44.
29.
R. A. Hurdie, Factorial Eco4Jgy, pp. 15B-66.
pp. 229-44.
30.
F. P. Eichelberger, "The Cubans in Miami: Residential Movements and
Ethnic Group Differentiations" (unpublished ~I.A. Thesis. Department of
Geography, University of Cincinnati. 1974), pp. 57-62.
31. K. O. Woodbury, "The Spatial Diffusion of the Cuban Colftlllunity in Oade
County. Florida" (M. A. thesis, Department of Geography, University of Florida.
197B).
32.
Tilllllls. Urban HOaaic. p. 241.
33.
McElrath, "Social Areas of Rome:"
12
Factopial Ecology. p. 161.
34.
~\urdie,
35.
Berry and 110rton, Geogr'Gphic Perspectives.
36.
Timms, Upban MOsaic. pp. 218_19.
37. For instance see: Anderson and Egeland, "Spatial Aspects," Schwirian and
Matre, "Ecological Structure," p. 320; and Schwirian and Smith, "Primacy," p.
330.
38. Straight line distances, rather than isochrones, were used here because
of the difficulty that would be involved in estimating driving times to each
of the 242 CensuS tracts used in this investigation.
39. The coefficients for the correlations between distance and social rank,
ethnicity, :md falOily status were .208, -.449, and .236, respectively.
40. Johnston ("Spatial I'atterns") has suggested logic for testing for a
distance effect within zones. A careful study of the lOapS in Figures 2, 3,
and 4 did not indicate that this task would be worth the effort in our
investigation. Clearly, there are some differences ..ithin tones of" the
sectors for the social rank and ethnicity SAA dimensions. We kno.. this through
the two analyses of variance which indicated significant levels of" interaction
in both cases. Still. with the e~ception of" ethnicity, there does not appear
to be a consistent distance continuum ef"feet.
41. This study dealt with ten cities. The two Southern cities involved were
Atlanta, Georgia and BiT1llingham, Alabama. Maurice D. Van Arsdol, Jr., Santo
F. Camilleri. and Calvin F. Schmid, "The Generality of" Urban Social Area
Inde~es," Amel"'icmn Sociological Howie", 23 ll958): 277-84.
Don 8randes