Citation
A spatial analysis of the dimensions of economic health in the southeastern United States (1950 and 1960)

Material Information

Title:
A spatial analysis of the dimensions of economic health in the southeastern United States (1950 and 1960)
Creator:
Romsa, Gerald Harry, 1969 ( Dissertant )
Cross, Clark I. ( Thesis advisor )
Brunn, Stanley D. ( Reviewer )
Anderson, James R. ( Reviewer )
Reynolds, John E. ( Reviewer )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1969
Language:
English
Physical Description:
ix, 101 leaves. : illus. ; 28 cm.

Subjects

Subjects / Keywords:
Agricultural population ( jstor )
Censuses ( jstor )
Cities ( jstor )
Counties ( jstor )
Economic regions ( jstor )
Factor analysis ( jstor )
Farms ( jstor )
Median family income ( jstor )
Poverty ( jstor )
Urban populations ( jstor )
Dissertations, Academic -- Geography -- UF ( lcsh )
Economic conditions -- Southern States ( lcsh )
Geography thesis Ph. D ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
The increasing attention given by geographers to the problems of poverty, economic health and underdeveloped areas is a healthy sign that they are progressing with the temporal needs of today's society. This concern with areal differentiation in economic growth and levels of economic well being is not a new geography. It is, instead, a more specialized topical outlook on the net result of economic activity as viewed in a spatial framework. This approach attempts to relate regional inequalities in levels of living with man's economic endeavours as they are tied to his cultural, technological and resource environment. The trend to look for factors other than resource endowment is perhaps the major criterion which separates present geographical research from previous regional studies.. Ullman has pointed out that natural resources and cultural differences are not sufficient to explain the disparity in economic growth. Differences in social conditions and advances in technology often provide additional information in explaining the areal variation in prosperity. The inclusion of additional variables to explain this differential in economic growth has led to the use of multivariate statistical techniques as analytical tools. The usefulness of these techniques lies in their applicability in the analysis of large numbers of indices; in their isolation of significant variables and in the testing of hypotheses. These are important contributions when dealing with interrelated data. The realization by social scientists that the distribution of wealth and poverty is uneven has focused their attention on the need for more regional studies. National and state data often tend to hide small pockets of high deviation in levels of income. This has led to economic regionalization based on counties, and, where possible, smaller units. Spatial interaction is another geographic concept which is increasingly being realized as playing an important role in explaining the difference in levels of progress. This concept was introduced by Ullman and refined by Hagerstrand at Lund University. Hagerstrand used simulation models in an attempt to measure the degree of resistance that cultural and physical barriers imposed upon the acceptance of new ideas among contiguous areas in Sweden. Hartshorne points out that the comparison of economic growth among underdeveloped regions should be relative rather than concrete. A direct comparison between two culturally different regions may not show the disparity in economic achievement accurately unless the indices utilized are standardized to represent with less bias the contrasting cultural and political values. This also is true of temporal studies. Although this problem at present has not been resolved, studies within a similar cultural and political unit based upon certain multivariate techniques are feasible.
Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 97-100.
General Note:
Manuscript copy.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
022059447 ( AlephBibNum )
13489737 ( OCLC )
ACY4880 ( NOTIS )

Downloads

This item has the following downloads:


Full Text







A SPATIAL ANALYSIS OF THE DIMENSIONS OF
ECONOMIC HEALTH IN THE SOUTHEASTERN
UNITED STATES (1950 AND 1960)









By
GERALD HARRY ROMSA


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY






UOF F LIBRARIES




UNIVERSITY OF FLORIDA
1969































To my parents and my wife

Mary













ACKNOWLEDGMENTS


I wish to express my gratitude to my advisor Dr.

Clark I. Cross and the other members of my committee, Dr.

Stanley D. Brunn, Dr. James R. Anderson and Dr. John E.

Reynolds for their guidance, thoughts, constructive criti-

cisms, and patience during the preparation of this disserta-

tion.

In addition, I want to acknowledge the University

of Florida Computer Center's facilities and services which

were utilized for the data analysis in this study.


1ii














TABLE OF CONTENTS


Page


. .iii


. vi


. ix


ACKNOWL',%UEDGMENTS . . . . . . . . .


LIST OF TABLES . . . . . . . . .


LIST OF FIGURES . . . . . . . . .


CHAPTER
I THE CONCEPT OF POVERTY . . . . .


Introduction . . . .
Objective . . .
Need for Such a Study . .
Poverty in the United States
Some Approaches to the Study
Review of the Literature .
The Problem . . . .


STUDY AREA AND METHODOLOGY .


The Setting . . .. .
Historical Background . . . .
Recent Trends in the Economy . .
Outline of the Analytical Model . .
The TModel. . . . . . . .
Variables Used . . . . . .
Summary . . . . . . . .


THE DIMENSIONS OF ECONOMIC HEALTH IN THE
SOUTHEASTERN UNITED STATES (1950-1960)
1950 Factor Analysis . . . .
The 1950 Dimensions . . . . .
1960 Analysis . . . . . .
The 1960 Dimensions . . . . .
Comparison of the Two Time Periods .
Regression Analysis . . . . . .
The Hypotheses . . . . . .
Summary . . . . . . .


. . 27
. 29
. . 30
. . 35
. 36
. . 37
* . 40
. . 42


. . 45
. . 46
. . 48
. . 60
. .62
* . 75
.* 78
* . 79
* . 84


. 1


. . . .
. . . .
. . . .
. . . *
of Poverty
.
. . e .


III


. . .









Page


CHAPTER

IV SUMMARY AND CONCLUSIONS . . . . . . 86
1950 Analysis . . . . . . . 87
1960 Analysis . . . . . . 89
Comparison of Changes in the Two Time
Periods Studied . . . . . . 92
Conclusions. . . . . . . .. 93
Future Research Efforts . . . . . 94

BIBLIOGRAPHY . . . . . . . . 97

BIOGRAPHICAL SKETCH . . . . ...... . ..101












LIST OF TABLES

Table Page

1 Per Cent of Families Earning Less than
$3,000 a Year by Regions (1960) . . . 8

2 Distribution of Poverty by Residence (1960) 9

3 Median Income by Residence (1960) . .. . 9

4 Variables Used . . . . . . . 41

5 Percentage of Total Variance Explained by
Each Dimension in 1950 . . . . . 47

6 Percentage of Variance (Communality) of
Each of the 26 Variables Accounted for
by All Six Components in 1950 . . . 47

7 Dimension I Urban Industrial Counties
in 1950 . . . . . . . . . 49

8 Counties with Extreme Factor Scores on
Dimension I in 1950 . . . . . . 50

9 Dimension II Tenant and Small Farm
Poverty in 1950 . . . . . . .. 52

10 Counties with Extreme Factor Scores on
Dimension II in 1950 . . . . . 52

11 Dimension III Wealthy Urbanized Counties
in 1950 . . . . . . . 55

12 Counties with Extreme Factor Scores on
Dimension III in 1950 . . . . 55

13 Dimension IV Rural VWhite Counties in 1950 57

vi








Table Page

14 Counties with Extreme Factor Scores on
Dimension IV in 1950 . . . . . . 57

15 Dimension V Locational Advantage in
Relation to National Market (New York)
in 1950 . . . . . .. .. 59

16 Counties with Extreme Factor Scores on
Dimension V in 1950 . . . . . . 59

17 Dimension VI Locational Advantage in
Relation to Local, Regional and National
Markets in 1950 . . . . . . . 60

18 Counties with Extreme Scores on Dimension
VI in 1950 . . . . . . . . 60

19 Percentage of Total Variance Explained by
Each Dimension in 1960 . . . . . 61

20 Percentage of Variance (Communality) of
Each of the 26 Variables Accounted for
by All Seven Components in 1960 . . . 61

21 Dimension I Urban Industrial Counties
in 1960 . . . . . . . . . 63

22 Counties with Extreme Factor Scores on
Dimension I in 1960 . . . . . . 64

23 Dimension II Rural Negro Counties . . 66

24 Counties with Extreme Factor Scores on
Dimension II in 1960 . . . . . . 67

25 Dimension III Tenant and Small Farm
Poverty in 1960 . . . . . . 67

26 Counties with Extreme Factor Scores on
Dimension III in 1960 . . . . . 69

27 Dimension IV Locational Advantage in
Relation to National Market (New York)
in 1960 . . .. . . ....... 70
vii








Table Page

28 Counties with Extreme Factor Scores on
Dimension IV in 1960 . . . . . 71

29 Dimension V Wealthy Urbanized Counties
in 1960 . . . . . . . . . 72

30 Counties with Extreme Factor Scores on
Dimension V in 1960 . . .. . .. 72

31 Dimension VI Locational Advantage in
Relation to National, Regional and
Local Markets in 1960 . . . . . 73

32 Counties with Extreme Factor Scores on
Dimension VI in 1960 . .... 74

33 Dimension VII Areas of Declining Prosperity
in 1960 . . . . . . . . 74

34 Counties with Extreme Factor Scores on
Dimension VII in 1960 . . . . 74

35 Expected Correlations Between Median
Family Income and the Derived Dimensions. 79

36 Normality of 1950 Factor Scores . . . 80

37 1950 Correlation Matrix . . . . . 80

38 1950 Summary Table . . . . . 81

39 Normality of 1960 Factor Scores . . . 81

40 1960 Correlation Matrix . . . . 81

41 1960 Summary Table . . . . . .. 82

42 Comparison of the Beta, Correlation
Coefficients and the Coefficient of
Determination Between the Two Time Periods 82


viii













LIST OF FIGURES


Figure Page

1 Study Area; the Southeastern Core . . . 28

2 Urban Industrial Counties, 1950 . . . 51

3 Tenant and Small Farm Poverty, 1950 . . 54

4 Urban Industrial Counties, 1960 . . . 65

5 Tenant and Small Farm Poverty, 1960 . . 68













CHAPTER I

THE CONCEPT OF POVERTY


Introduction

The increasing attention given by geographers to the

problems of poverty, economic health and underdeveloped

areas is a healthy sign that they are progressing with the

temporal needs of today's society. This concern with areal

differentiation in economic growth and levels of economic

wellbeing is not a new geography. It is, instead, a more

specialized topical outlook on the net result of economic

activity as viewed in a spatial framework.1 This approach

attempts to relate regional inequalities in levels of liv-

ing with man's economic endeavours as they are tied to his

cultural, technological and resource environment.

The trend to look for factors other than resource

endowment is perhaps the major criterion which separates

present geographical research from previous regional stud-

ies, Ullman has pointed out that natural resources and cul-

tural differences are not sufficient to explain the dispar-
2
ity in economic growth. Differences in social conditions

and advances in technology often provide additional informa-

tion in explaining the areal variation in prosperity. The

inclusion of additional variables to explain this
1









differential in economic broth has led to the use of multi-

variate statistical techniques as analytical tools. The

usefulness of these techniques lies in their applicability

in the analysis of large numbers of indices; in their isola-

tion of significant variables and in the testing of hypoth-

eses. These are important contributions when dealing with

interrelated data.

The realization by social scientists that the dis-

tribution of wealth and poverty is uneven has focused their

attention on the need for more regional studies. National

and state data often tend to hide small pockets of high

deviation in levels of income. This has led to economic

regionalization based on counties,and, where possible,

smaller units.

Spatial interaction is another geographic concept

which is increasingly being realized as playing an important

role in explaining the difference in levels of progress.

This concept was introduced by Ullman and refined by Hager-

strand at Lund University. Hagerstrand used simulation

models in an attempt to measure the degree of resistance

that cultural and physical barriers imposed upon the accept-

ance of new ideas among contiguous areas in Sweden.

Hartshorne points out that the comparison of eco-

nomic growth among underdeveloped regions should be rela-

tive rather than concrete.4 A direct comparison between two

culturally different regions may not show the disparity in









economic achievement accurately unless the indices utilized

are standardized to represent with less bias the contrast-

ing cultural and political values. This also is true of

temporal studies. Although this problem at present has not

been resolved, studies within a similar cultural and polit-

ical unit based upon certain multivariate techniques are

feasible.


Objective

The purpose of this study is to identify the salient

socio-economic dimensions or characteristics of five south-

eastern states (Mississippi, Alabama, Georgia, South

Carolina and North Carolina) for two time periods, 1950

and 1960. These characteristics will be examined to ascer-

tain if they reflect any areal patterns or changes through

time. If areal variations do exist, they will be analyzed

in regard to their stability, location, effects of urbaniza-

tion and the importance of push and pull factors.

In addition the extracted dimensions will be tested

individually and as a group for their validity in explain-

ing the high deviation in median income throughout the

study area.


Need for Such a Study

The concern of some Americans for a more equitable

distribution of this nation's wealth has shown itself

through civil legislation and action programs by state,









local and federal agencies. Programs such as the Area

Redevelopment Act and the Appalachia Act are examples of the

federal government's increased participation. To implement

these programs rapidly and to the best advantage requires

a thorough knowledge of the problems. Thus research is

essential if the complex facets of poverty are to be under-

stood. An example of such a research program in the South

is The Southern Regional Research Project S-44, "Factors in

the Adjustment of Families and Individuals in Low Income

Areas of the South." This is a cooperative study done by

rural sociologists in nine southern states. These studies

were done in selected counties and dealt primarily with the

attitudes of the rural poor.

There is a need, however, for a comprehensive study

of the Southeast which deals vith the facets associated with

poverty on a larger scale. The Southeast has often been

classified as a retarded area. Recent studies have sug-

gested that this blanket statement can not be applied to

the entire region, as it has made some great economic strides

in the past few decades. At present, there is a lack of

geographic research on this region employing multivariate

statistical techniques, which has emphasized the degree of

variation in economic growth through space and time and

classified economic regions accordingly.

In the remainder of this chapter an overview of the

distribution of poverty in the United States and some






5



approaches to its study are discussed. The pertinent lit-

erature is reviewed as a basis for formulating hypotheses.

The analytical model used in this study is developed in

Chapter II and a brief description of the study area is

given. Chapter III deals with the analysis of the data,

while in ChapterlV the conclusions reached are examined

in view of the hypotheses presented in Chapter I.


Poverty in the United States

Amid the highest standard of living ever achieved

by the United States there exists a forgotten minority of

poverty stricken people. The problems of the poor are often

masked by the middle class way of life in America. There is

no room in this society for the economically less fortunate.

The middle class viewpoint is well expressed in The Affluent

Society by John Galbraith.5 Galbraith's major theme was how,

with increasing automation, the good life could be continued

for all, with less work and more leisure. Poverty was per-

ceived only in two forms: case poverty due to physical

and mental handicaps, and insular poverty such as is found

in West Virginia.

More recent articles have shown that the problems of

poverty are not isolated but nationally widespread. There

are, however, large concentrations of the poor in certain

geographic areas such as the South and in the ghettos of

urban centers. Depending upon the criteria used, there are









estimated to be from 36 to 50 million economically deprived

persons within this wealthy nation.7

This staggering figure is made even more realistic

if one considers the economic, social and psychological

costs of maintaining these individuals in their present way

of life. Not only does the Gross National Product suffer

from the loss of such a large number of human resources,

but there are also heavy social costs. These include wel-

fare payments, costs of mental and physical health, crime

and juvenile delinquency. The national cost of welfare pro-

grams has been estimated at 13 to 15 billion dollars.

Poverty was first recognized as a social problem in

the United States approximately 100 years ago. It existed

before but had been regarded as an individual rather than a

social problem. Prior to this period, individuals attempted

to attain economic freedom by leaving the urban ghettos or

the worn out soils of the East and moving westward. This

movement could only be sustained in a highly agrarian soc-

iety. The slaves on the southern plantations however did

not have this escape valve.

Increasing industrialization and urban growth after

the Civil War brought with it many social evils: poor tene-

ment housing, inadequate schools, low wages, bad working

conditions and no job security. Henry George in his Pro-

gress and Poverty referred to the United States of 1869 as









a land where "amid the greatest accumulation of wealth, men

die of starvation and puny infants suckle dry breasts."10

The plight of the working classes, many of whom were immi-

grants, eventually led to some social reforms. Community-

minded citizens led the battle for public housing regula-

tions, public health, and abolition of child labour. Labour

unions played a major role in obtaining higher wages and

better working conditions.

Major Federal government action on poverty was ini-

tiated by President Franklin D. Roosevelt in the 1930's.

In the State of the Union message in 1933 the former Presi-

dent referred to the "one third of the nation that is ill-

clothed, ill-housed and ill-nourished."1l It was during

this period that the familiar antipoverty tools: social

security, old age assistance, aid to dependent children,

and unemployment compensation were conceived. Furthermore

the Tennessee Valley Authority Act was a recognition that

there were regional inequalities in economic growth, which

resulted in pockets of distressed areas that were resistant

to change and progress.

The next major Federal attack on poverty commenced

in the 1960's. The emphasis was on legislation. The Area

Redevelopment Act of 1962, the Economic Opportunity Act and

the Appalachian Regional Development Act of 1964 are a few

examples. The passage of the acts emphasized rehabilitation

and not monetary relief. They also recognized the areal









distributions of poverty and the need to focus upon the

problems of regional or subregional nature which can not

be solved on the national level. The new progress in

legislation was initiated by a small educated white minor-

ity and a group of dedicated Negro leaders. Civil Rights

marches, sit-ins, and riots have forced the nation's atten-

tion to the problems of her minority poor when other less

peaceful efforts failed.12

The geographic distribution of poverty can readily

be seen if the per cent of families earning less than

$3,000 a year is examined by the major divisions of the

country (Table 1).


Table 1

Per cent of Families Earning Less Than
$3,000 a Year by Regions (1960)

Region Total White Nonwhite
Northeast 9.8 8.4 26.6
North Central 11.5 10.3 29.7
South 24.6 17.9 58.3
West 11.7 11.0 20.7

Source: Louis A. Ferman, Joyce L. Kornbluh and Alan Haber
(eds.), Poverty in America (Ann Arbor: University
of Michigan Press, 1965), p. 72.


This table shows the enormous variation in income across

America. It further points out the plight of the non-

whites. Although the urban centers harbour the majority

of the poor, the rural areas have a higher percentage of

poor families (Tables 2 and 3).








Table 2

Distribution of Poverty by Residence (1960)

Residence % of population % of poor (families
earning less than
$3,000 per year)

Urban 71 54
Rural nonfarm 22 30
Farm 7 16


Source: Louis A. Ferman, Joyce L. Kornbluh and Alan Haber
(eds.), Poverty in America (Ann Arbor: University
of Michigan Press, 1965),p. 104.


Table 3

Median Income by Residence (1960)

Residence Total Population White Nonwhite

Urban $6,580 $6,678 $4,469
Rural Nonfarm $5,486 $5,549 $2,645
Rural Farm $3,779 $3,863 $1,323

Source: Louis A. Ferman, Joyce L. Kornbluh and Alan Haber
(eds.), Poverty in America (Ann Arbor: University
of Michigan Press, 19657, p. 107.

In comparing the income between rural and urban

residents, one should consider the difference in cost of

living. Nevertheless the differences in income are greater

than the variation in the cost of living. From these

figures a picture arises that reveals small concentrated

areas of poverty within cities and a more dispersed form of
rural poor. Within the rural sections of the country, there

are extreme variations in the concentrations of the poor.
Examples of areas with a high rate of unemployment and a

concentrated number of low income groups are: Appalachia,









the cutover regions of the Great Lakes, northern New
13
England, the northern Rockies and the Indian reservations.

Other regions containing large but less concentrated numbers

of poor families are parts of the northern Great Plains,

scattered areas in the West and Midwest and extensive sec-

tions of the Piedmont and the Atlantic Coastal Plain.1

This pattern of dispersion not only shows the

national scale of the problem, but also the variations

from the industrial North to the rural West and the South.

The distribution of natural resources can not account for

all of the variation in poverty. Part of this is due to

demographic factors. Minority nonwhite groups are con-

centrated in many of the distressed economic sections.

Some depressed regions have a high percentage of unskilled

labour, while many of the predominantly white rural coun-

ties have high concentrations of older people. The latter

is the result of the out-migration of the young to more

attractive economic areas. Continuous out-migration over

a period of time, of the highly motivated segment of the

population, will tend to economically depress an area. This

was illustrated and substantiated by Spinelli in his study

of the net out-migration in southeastern Ohio, 1950-1960.


Some Annroaches to the Study of Poverty

One of the most difficult problems in studying pov-

erty is its definition. 4What constitutes poverty? Many









popular definitions adopted are based upon incomes. Sta-

tistics on family and individual income are readily avail-

able and accurate. The present $3,000 criterion used for

families and the $1,500 for single individuals was defined

and used in 1964 by the President's Council of Economic

Advisors. The Council estimated that in 1962 there were

between 33 and 35 million Americans living at or below this

income level.16

Economists such as Leon Keyserling believe that this

criteria are too low. Keyserling raises the income criteria

for families living at the poverty level to $4,000 and for

single persons to $2,000 (1960 dollars). In his analysis

he includes a second grouping called the deprived category.

This group includes all families with incomes between

$4,000 and $5,999 and individuals with earnings between

$2,000 and $3,999. This latter group can subsist on their

incomes, but can not afford many of the luxuries that our

present society has to offer. Keyserling estimates that 73

million Americans fall within these two categories.17

One of the weaknesses of these estimates, based upon

the above definitions, is that they do not account for the

differences in the cost of living throughout the country.

They also do not contain flexibility to take into account

inflationary periods of the economy which increase the cost

of living for individuals on fixed incomes. The social

security administration has made some attempt to account









for the variation in cost of living, but the invariant

$3,000 base is still largely used by them. A more flexible

standard which distinguishes the differences in need between

size of families, rural and urban residence, was published

in 1964 by the Department of Health, Education and Welfare.

This minimum budget ranges from $1,100 for a farm family of

two, aged 65, to over $5,000 for a nonfarm family of seven

or more individuals. This more flexible budget estimated
18
the poor at 34.6 million. More realistic income criteria

are those suggested by the Bureau of Labor Statistics.

Their index was developed for an urban family of four liv-

ing in 20 of the major cities. This index varied from

$5,370 in Houston to $6,567 in Chicago.19

The above estimates show the disagreements among

experts on what is a proper income level for the definition

of poverty. These estimates, furthermore, have not taken

into account transfer payments in the form of financial

assistance from welfare agencies or relatives. They also

have not made any allowance for taxation and therefore,

because of these omissions, do not represent true disposable

income. Indeed most students of poverty feel that the use

of a single index is not sufficient. The trend is to use

several indicators such as color, age, income and education.

Some social scientists have approached the problem

from the sociological and psychological viewpoint. One of









the first to use this definition was Oscar Lewis, who states:

Poverty in modern nations is not only a state
of economic deprivation, of disorganization or
of the absence of something. It is also positive
in the sense that it has a structure, a rationale,
and defence mechanisms without which the poor
could hardly carry on. In short it is a way of
life, stable and persistent passed down from gen-
eration to generation along family lines. The
culture of poverty has its own modalities and
distinctive social and psychological consequences
for its members.20

Economically retarded areas have within them distinct

cultural features. The poor having little money, virtually

no savings and no economic security, have evolved unique

consumer habits, family structures and a way of life. This

subculture, if it may be so called, is composed of the

chronically unemployed or those having irregular employment.

Those who are employed work in the non-unionized low-paying

sectors of the economy. Their low salaries can not provide

for the costs of higher education and health services.

Often their meager earnings are spent on entertainment

luxuries,instead of the basics of food, shelter and cloth-

ing, in an attempt to escape their environment for short

periods of time. The end result is often a low resistance

to disease, ill health and a shorter life span.

The everyday problems of this subculture, due to the

constant stress of maintaining life without hope, has created

in turn psychological manifestations. This has lead to a

lack of motivation and imagination among the destitute. Pro-

tective shells against the outside world are erected. Within










these shells provincialism and deep-rooted prejudices are

maintained. The outside world is viewed with hostility.

This hostility is expressed in their distrust of various

government agencies and programs.

The sociological approach in defining poverty,

although accurately describing the subculture of the poor,

has certain drawbacks. The major restriction is the lack

of sociological data for large sections of the country. To

gather the data requires personal interviews, which not only

can be costly, but also time consuming. In addition, this

definition emphasizes to a great extent the end results

rather than the causes.

A more recent concept, economic health, is gaining

in popularity as it attempts to analyze which indices are

responsible for a lower level of living and how these

indices may vary through space and time. The term economic

health is not universally used, but is represented in a

large number of multivariate studies that use numerous

social, economic and demographic indices. These indicators

are then reduced to a smaller number by elimination or

union with other variables into new ones. This new set of

indices, which accounts for the largest possible variance

within a workable matrix, is then used to explain the dif-

ference in economic growth. This matrix can be applied to

inter-, or intra-regional studies and to a certain extent










be used to explain the changes through time. Economic

health thus tries, by the use of a large number of variables,

to explain why and how different regions of a country,

relative to each other, differ in their overall economic

prosperity. The method of analysis is not new, but it was

not until rapid calculating computers became available that

large regions could be examined on the basis of a multitude

of variables, and at the same time remove personal bias in

regard to the importance of the variables used. This form

of analysis is being used successfully by various social

scientists engaged in regional studies.

The concept of economic health will be used in this

study to explain the differences in prosperity, if any,

within the five southeastern states in the study area.

Twenty-six variables which are indicators of prosperity

will be utilized to ascertain what role they play within

the study area, and if their importance changes through

space and time.



Review of the Literature

The dependence of the Southeast on agriculture has

been documented by geographers (Hart, Vance).21 This

historical reliance upon a few major crops such as cotton

and tobacco had its drawbacks. These crops were export

orientated and consequently fluctuated greatly in price.









Soil mismanagement often led to the abandonment of farms or

a loss in fertility and yields. The plantation form of

agriculture was not conducive to the modernization of agri-

cultural practices, as cheap labour after slavery made it

economically unfeasible to adopt some of the new techniques

until relative prices of the inputs changed. Rural values

predominated and eventually evolved into a way of life.

Political power came under the control of a small, land

based minority to the detriment of education and industrial-

ization.

This theme, that the South's historical development

acted as a barrier to progress, pervades the literature on

the South. Nichols in a paper "Southern Tradition and

Regional Economic Progress" suggests that five social, polit-

ical, psychological and philosophical factors were respon-
22
sible for the lag in southern economic growth.22 These were:

the dominance of agrarian values, rigidity of a social class

structure which prevented the establishment of a strong mid-

dle class, the control of the political parties by a few which

led to an undemocratic political structure. Furthermore

a lack of social responsibility hindered the development

of a first class public school system. This, tied with the

conformity of thought and behavior which was expected of a

good southerner, left little or no sympathy for dissenters,

many of whom migrated North. The loss of this liberal ele-

ment hindered change and helped maintain the status quo.









In spite of these handicaps the South in the last

few decades has made tremendous strides forward. MacDonald

"On South's Recent Economic Development" points out that

personal income in this region, has increased since 1940,

at a faster rate than in the nonsouth.23 Part of this

increase is attributed by MacDonald to a shift in the Amer-

ican economy from coal and steel to plastics and petroleum,

and the increased exploitation of other southern resources

especially forestry and agriculture. The other main factor

responsible for the rapid increase in personal income was

a slow population growth rate, due to the heavy out-migra-

tion of low income groups. This recent progress of the

South was further verified by Lassiter. In his study

"Education for Males by Region, Race and Age "Lassiter

pointed out that at present the investment in education is

providing better financial returns for white southern males

than white northern males.24

This rapid rate of growth however has not been con-

sistent throughout this region. Moon and McCann, who looked

at the levels of adjustment in 30 rural counties in the

states of Kentucky, Tennessee, North Carolina, Texas,

Louisiana, Mississippi and Alabama, found statistically sig-

nificant variation.25 Similar conclusions were arrived at

by Tarver and Beale in "Population Trends of Southern Non-

metropolitan Towns."26 They concluded that the significant

factors explaining the differences in population growth









among 801 southern towns with populations between 2,500 and

9,999 were: the size of the town and its regional location.

Fuguitt on the other hand found that county seats in rural

nonmetropolitan areas grew at the expense of neighboring
27
centers.2

The importance of urbanization as a significant fac-

tor in explaining regional growth in the United States has

received considerable attention. Schlesinger has argued

that since the Civil War, cities dominated the American
28
growth process. Priedman and Alonso depict a spatial

economy in which the processes of economic development are

leading to greater specialization and increased polarization

of growth in a few urban industrial complexes, each of which

is functionally integrated with a surrounding economic area29

These centers not only grow so rapidly as to create
problems of an entirely new order, but they also act
as suction pumps, pulling in the more dynamic ele-
ments from the more static regions. The remainder
of the country is thus relegated to a second class,
peripheral position. It is placed in a quasi-
colonial relationship to the center, experiencing
net outflows of people, capital, and resources,
most of which is redound to the advantage of the
center where economic growth will tend to be rapid,
sustained and cumulative. As a result, income dif-
ferences between the center and the periphery tend
to widen.30


This viewpoint is also expressed by Schultz, who

argues that economic development occurs primarily within

an industrial urban matrix where the factor and product

markets work best. Agricultural activities located most









favourably in regard to these centers can utilize the fac-

tor and product markets more efficiently than those located

on the periphery of these urban systems.3 This was

partially substantiated by Bryant who used multiple regres-

sion to measure the causes of inter-county variation in

farmers' income in the United States.32 He showed that

variation in farm income is largely determined by age, land

and capital inputs, education and color. Location in respect

to regional urban areas was only significant in the eastern

United States, implying that farmers west of the Mississippi

were more dependent upon national and international markets.

0. D. Duncan, et al., in Metropolis and Region,emphasize the

domination of hinterland activities by metropolitan centers3

Both agricultural and non-metropolitan manufacturing activ-

ities were shown to be a function of general accessibility

to the urban system of the nation and distance to the local

metropolis. Berry concluded in his study of rural poverty

in Ontario that

areas lacking the dynamic conditions of industrial
urban development experience high rates of out-
migration and commitantly realize differences in
per capital income, even though labour migrations
generally appear sufficient to overcome differences
attributable to differences in original resource
endowments. The factors linking rural poverty to
industrial urban growth polarized in urban centers
and varying in amplitude according to the rank of
these centers in an urban hierarchy, are thus clear.

This urban influence has also been clearly demon-

strated by Thompson, et al.,35 Berry,36 Bell and Stevenson,37









38
and Hodge. All of these authors used factor analysis on

a large matrix of economic indices in order to derive a

smaller number of dimensions or factor loadings which would

explain more easily the variations in income levels. In all

cases an urban dimension was derived from the analysis that

was correlated with the rapidly growing areas. The study

by Thompson, et al., on the economic health of New York

State, is of particular interest here. It not only demon-

strated that the large metropolitan cities were the nodes

of economic growth, but also that the predominantly agri-

cultural counties near these centers fared better than

smaller cities, which,had populations of 50,000 or less.

Berry arrived at the same conclusions in his analysis of

rural poverty in Ontario. Location of the rural townships,

in relation to urban areas, played an important role in

determining the level of economic achievement as measured

by 47 variables. In only a few exceptions, most of which

were due to the presence of mineral resources, did a few

townships deviate from the pattern of hinterlands dominated

by large urban industrial clusters. Using similar procedures

Hodge analyzed 473 trade centers in Saskatchewan. He found

that trade center viability was largely a function of rank

on the urbanism scale; the lower the rank the greater the

probability of decline.

The apparent importance of the urban or industrial

dimension is clearly seen in these studies. That these








complexes do grow at the expense of their hinterlands is

readily verified, if one looks at the variables which are

correlated with the urban industrial dimensions. These are

generally high income groups, youthfulness of the popula-

tion, a high level of education and a high level of capital

expenditures. All are indices of a healthy economy. The

converse is true; the dimensions representing poverty have

weaker associations with the above indices.

Are those indices stable through time? The decline

of certain regions relative to others would seem to indicate
39
the contrary. King found that they were not stable. In

his examination of Canadian cities he found that the per-

centage of variance explained by the dimensions arising from

his analysis varied through time, although the factor load-

ings associated with the dimensions remained fairly stable.


The Problem

The problem involved in this study relates to the

variations in the economic environment among 454 counties

in Alabama, Georgia, Mississippi, North and South Carolina.

Specifically does variation exist? What dimensions account

for this variation? Do they vary over time? What percent-

age of poverty do they explain? Are some counties progres-

sing at the expense of others? The literature reviewed pro-

vides a basis from which to examine several useful spatial
hypotheses of economic health in the selected southeastern
states. The hypotheses to be tested in the study are:









(1) There is variation in economic health among the

selected counties and these can be explained in terms of a

few dimensions arising out of factor analysis.

(2) These dimensions are not stable but change

through time as do the conditions required for economic growth.

(3) Within the study area there will be nodes of

rapidly growing areas which thrive at the expense of their

hinterlands.

(4) These nodal areas are associated with urbaniza-

tion and it is postulated that the degree of growth within

these sectors, relative to each other, is a function of popu-

lation size and location, location in relation to local,

regional and national metropolitan centers.

This study will be successful if it can demonstrate

the underlying characteristics of areal variation in economic

health in the five selected southeastern states for two points

in time, and ascertain if these characteristics have changed

between the two time periods (1950 and 1960).












REFERENCES


Richard Hartshorne, "Geography and Economic Growth,"
in Essays on Geography and Economic Development, Norton
Ginsburg (ed.) (Chicago: The University of Chicago Press,
1960), pp. 24-25.

Edward L. Ullman, "Geographic Theory and Under-
developed Areas," in Essays on Geography and Economic Devel-
opment, Norton Ginsburg ed.) (Chicago: The University of
Chicago Press, 1960), pp. 26-27.

Ibid., p. 29.

4Hartshorne, op. cit., pp. 18-23.

5John Galbraith, The Affluent Society (Boston:
Houghton-Mifflin Co., 1958).

Ibid., pp. 323-333.

7Michael Harrington, The Other America, Poverty in
the United States (Baltimore: Penguin Books, 1966), pp.
176-179.

Lar A. Levitan, "Programs in Aid of the Poor,"
Poverty and Human Resources Abstracts, Vol. I (1966), pp.
11-25.

9Herman P. Miller, Poverty American Style (Belmont:
Wadsworth Publishing Co., Inc., 196b), pp. 2-3.

0Louis A. Ferman, Joyce L. Kornbluh and Alan Haber
(eds.), Poverty in America (Ann Arbor: University of
Michigan Press, 1965), p. 1.

11Ibid.

12 Miller, op. cit., p. 5.









13
Gordon E. Record, "The Geography of Poverty in
the United States," in Problems and Trends in American
Geography, Saul B. Cohen (ed.) (New York: Basic Books,
Inc. 1967), p. 105.

14Ibid.

5Joseph Spinelli, "A Study of Net Out-Iigration in
Southeastern Ohio, 1950-1960" (Unpublished Master's Thesis,
Department of Geography, Ohio State University, 1966).

16Miller, op. cit., p. 2.

17Ibid.

18
H18arrington, loc. cit.

19Ibid.

20
2Elizabeth Herzog, "Some Assumptions About the Poor,"
Social Science Review, XXXVII (December, 1963), 390.

21
21John Fraser Hart, The Southeastern United States
(Princeton: D. Van Nostrand Co., Inc., 1967); and Rupert B.
Vance, Human Geography of the South (Chapel Hill: Univer-
sity of North Carolina Press, 1935).

22
22William H. Nichols, "Southern Tradition and
Regional Economic Progress," Southern Economic Journal, XXVI
(January, 1960), 187-198.

23Stephen L. MacDonald, "On South's Recent Economic
Development," Southern Economic Journal, XXVIII (July, 1961),
30-40.

24Roy L. Lassiter, Jr., "Education for Males by
Region, Race and Age," Southern Economic Journal, XXXII
(July, 1965), 15-22.

25
Seung Gyu Moon and Glenn C. McCann, Subregional
Variability of Adjustment Factors of Rural Families in the
South (Raleigh: North Carolina Agricultural Experiment
Station, Southern Cooperative Series, Bulletin No. 111,
January, 1966).









26James D. Tarver and Calvin L. Beale, "Population
Trends of Southern Nonmetropolitan Towns," Rural Sociology,
XXXIII (March, 1968), 19-29.

27
2Glenn V. Fuguitt, "County Seat Status as a Factor
in Small Town Growth and Decline," Social Forces, XLIII
(December, 1965), 245-251.

2Arthur I'. Schlesinger, A History of American Life,
Vol. XI, The Rise of the City 1878-1l89 (New York: The Mac-
millan Co., 1949).

29John Friedman and William Alonso (eds.), Regional
Development and Planning (Cambridge: Massachusetts Institute
of Technology Press, 19b4), pp. 1-11.

0Ibid., p. 3.

31T. W. Schultz, Economic Organization of Agri-
culture (New York: McGraw-Hill Co., 1953).

32
3Keith W. Bryant, "Inter-County Variation in
Farmers' Income," Journal of Farm Economics, XLVIII
(August, 1966), 557-577.

33Otis Dudly Duncan, et al., Metropolis and Region
(Baltimore: The John Hopkins Press, 19b0).

4Brian J. Berry, Strategies, MIodels and Economic
Theories of Development in Rural Regions. Agricultural
Economic Report No. 127 (Washington: Economic Research
Service, U.S. Department of Agriculture, 1967).

35
5John H. Thompson, et al., "Toward a Geography of
Economic Health: The Case of New York State," in Regional
Development and Planning, John Friedman and William Alonso
(eds.) (Cambridge: Massachusetts Institute of Technology
Press, 1964), p. 187.

3Brian J. Berry, "Identification of Declining
Regions: An Empirical Study of the Dimensions of Rural
Poverty," in Thoman, R. S. (ed.), Areas cf Economic Stress
in Canada (Kingston: Queens University, 19b5).

37W. H. Bell and D. W. Stevenson, "An Index of Eco-
nomic Health for Ontario Counties and Districts," Ontario
Economic Review, II, No. 5 (1964), 1-7.





26



38
3Gerald Hodge, "Do Villages Grow? Some Perspectives
and Predictions," Rural Sociology, XXX, No. 2 (June, 1966),
183-196.

3Leslie J. King, "Cross Sectional Analysis of
Canadian Urban Dimensions 1951 and 1961," Canadian Geographer,
X, No. 4 (1966), 205-224.













CHAPTER II

STUDY AREA AND METHODOLOGY


The study area is composed of five states, Missis-

sippi, Alabama, Georgia, North and South Carolina. This

region covers an area of 241,968 square miles, and in 1968

had a population of 16,326,746. These states are similar

in their socio-economic-political characteristics (Hart,l

Vance,2 Odum ). All five withdrew with the Confederacy and

after Reconstruction the states voted solidly Democratic

until 1964, with the exception of North Carolina which went

Republican once.

Perhaps due to the apparent homogenity of the South,

few authors have attempted to subclassify it. What region-

alization of the South has taken place is largely based

upon physical divisions such as the Delta, Piedmont,

Appalachia and the Coastal Plain (White and Foscue,4

Griffin, Young, and Chatham5). Bogue and Beale however,

using 64 variables, have broken the South down into four

distinct regions.6 One of which, the Southeast Coastal Plain

and Piedmont, coincides very closely with the area under

study (Figure 1). Five distinct characteristics separate

this region from the others. It is the least urbanized, has

the highest percentage of Negro population, the highest
27
















7 4
~22202

2*2224
422 2,


^ 1.'1 ..* ^ / /- -^

;.. / i i i '.j ..' \ i ^ -

V.^^Ip ^ yd 5 l -''
.9P 2 '2I -; I

<* 2 ^ ? i/ ^ 4'-
Y 2, i-vE */ *1 i,



M~~ l->-<^ ^- i i
2s






9 20 'Z 2 2
^~i 11 in \rr^^
i ^ -- ~i i ,
< II i- ,' i -^ ,- -^ -v i'i




9 4- I i -
pT-^isn^

S~~ ~~~ b ^ '^ ^^-^1
C/ i, 5 j r/ i^' i 11^ ^
r i I: ; ; '** B I j --P-- '
0) 0


20o 22 i
.2 '2 2,
h E ^ k--^^'-- *) 11 i4-
^ h u s^ ~- ^- ^1^^ <^









percentage of the labour force in agriculture, and the

industrial base is weak and highly dependent upon textiles.

These four characteristics are the cause of the fifth, low

median income, as the lowest median family income of any

region in the United States is found here. Furthermore

Bogue and Beale found enough variation within this distinct

region to classify it into 19 subregions, implying that

there may be differential growth patterns.

The choice of this study area was dependent to a

great extent on the previous work done by Bogue and Beale.

Although the two areas are not exactly the same, they may

be considered as one in their characteristics. The reason

for this discrepancy is that state boundaries were maintained

so as to reduce the effect of state governments in the anal-

ysis.



The Setting

Three distinct physiographic features fall within

the study area, the Coastal Plain, Piedmont and Appalachia.

The boundary between the Piedmont and the Coastal Plain is

demarcated by a Fall Line and a change in soils. The Fall

Line is characterized by a zone of waterfalls and rapids in

the streams which imposed a limit on navigation. Cities

such as Raleigh, Columbia, I.'acon, Columbus and Montgomery

are found along this break. In the Carolinas and Georgia

the Fall Line is marked by a long, low narrow belt of sand










hills, the remnants of ancient beaches. The Piedmont soils

are residual, the product of weathering of the underlying

rock over eons of time. The Coastal Plain soils on the

other hand were formed of sedimentary materials when the

plain was submerged.

Both physiographic divisions have a variety of local

topography ranging from level to hilly, with the older Pied-

mont having the greater extremes of relief. A small portion

of the study area falls within Appalachia. This section

extends from central Alabama to the Blue Ridge of North-

western Georgia and North Carolina. In the Blue Ridge the

mountains rise 3,000 to 6,000 feet above sea level. In the

Appalachian Valley of Alabama the relief is much less, varying

from 500 to 1,000 feet.

The economy was and to a large extent still is

centered on agriculture. Three major crops: tobacco, cot-

ton and peanuts provide the majority of the farmers' cash

receipts. Industrial growth has been based upon service

functions, textiles, food processing, furniture and lumber-

ing. In many respects the economy is similar to that of

underdeveloped countries, in that it is dependent on a few

staple agricultural products and has a weak industrial base.


Historical Background

The retardation of this area,as in the rest of the

South, has largely been due to the historical development of








the region. This historical retardation has been well

documented by historians and economists (Eaton, Maddox ).

It was the development of an agrarian culture that led to

the exclusion of the rise of a strong middle class, and

industry. These could have become the impetus for economic

growth, when agriculture was diminishing in importance.

Once behind, the South was in a poor position to compete

with the industrial North, and it became subservient to a

centralized economy based upon market potentials.

The Southeast was settled by two streams of settle-

ment. The first group occupied the tidewater between

Charleston and Savannah. Plantations were established and

at first indentured servants, later slaves, were brought in

as a source of labour. Rice became the dominant crop along

the coast, while indigo was grown on the plantations further

inland. As the tidewater was being settled a different

group of immigrants were coming into the western Carolinas

and Georgia. These were the Scotch-Irish and Germans who

had come south from Pennsylvania. Many of these settled

permanently in the backcountry and became successful farmers,

raising all their requirements. The excess livestock, hogs,

corn were sold or converted to other more valuable by-

products such as liquor or hides. The heavy migration into

the western area led to a more rapid rate of growth than in

the tidewater. Eventually this may have led to the political

control of the southern colonies by the small farmers, had









not the importance of cotton began to assert itself. Cotton

planters began to move into the Piedmont and it became the

major crop. The search for more and better cotton soils led

to the opening up and settlement of Alabama and Mississippi.

The relatively rapid rise in the dominance of cotton resulted

from two major events, the industrial revolution in Europe,

requiring cotton for its textile mills, and the invention of

the cotton gin by Whitney in 1793.

The emergence of King Cotton as the dominant crop,

controlled by a small group of planters, utilizing slave

labor, pre-empted the use of land for other crops. This

small aristocratic group gained and maintained political

control throughout the area. Furthermore this group man-

aged to impose their values upon the entire southern soci-

ety. The use of slave labor was a barrier to the establish-

ment of more productive agricultural practices which would

utilize a higher share of capital as a factor of production.

The social stigma associated with trade and manufacturing

hindered the development of a middle class and a more diver-

sified economy. Perhaps the greatest damage done was the

concept that education was not for the masses, but for the

select few. The education provided the select few, although

of high calibre, was heavily oriented toward the liberal

arts and was lacking in the field of applied technology.








Other important sources of income were lumber and

naval stores, which initially were of prime importance in

North Carolina but rapidly spread throughout the South.

Tobacco became the second major crop when, in 1850, it was

discovered that bright tobacco did well on the poor sandy

soils of northwestern North Carolina. The third major crop

was the growing of peanuts along the Georgia and Alabama

border.

Thus in the period preceding the Civil War the

southern economy was oriented around the export of a few

major staples as sources of revenue with which to pay for

the importation of manufactured goods from abroad and the

North. This raw material orientation meant that the econ-

omy was affected by any severe fluctuations in prices or

tariffs. The tariff of abominations established by the

North in 1828 was a blow to southern commerce.0

The Civil War brought ruin upon the South. The

abolition of slavery removed a source of cheap labour from

the plantations and left many of these large land holdings

idle. A system of sharecropping evolved to fill the labour

shortage. The system was not satisfactory for it did not

encourage any long term agricultural practices by either

the landowner or the sharecropper. In addition, develop-

ment of either industry or a more efficient system of

agriculture was strongly hampered by the shortage of invest-

ment capital. Inflation during the War had left the








Confederate states with a worthless currency. Deflation

after the war left the South with little buying power.

Some of the capital problems were attributable to national

monetary policies which hurt the South. The National Bank

Reforms in 1865 levied a severe tax on state banks, which

predominantly were located in the South. Furthermore in

these bank reforms, the South did not receive its fair
11
allotment of Federal currency.

The establishment of certain industries such as

iron and steel was also hampered by the monopolistic action

of northern firms who established "Pittsburgh plus basing

point" and by this means removed the locational advantage

of Birmingham. The industrial growth that did take place

occurred mainly in the field of food processing, textiles

and cigarette manufacturing. The use of bright tobacco as

a major ingredient in cigarettes gave rise to cities like

Durham and Winston Salem. The textile industry moved here

from the North during the later part of the nineteenth

century. Many of the mills were established in towns along

the Fall Line where there was a source of power. Despite

these attempts at industrialization, the South did not keep

pace with the North. As most poor or stagnating areas it

suffered from selective out-migration, lack of capital and

the establishment of new firms. All of these factors

encouraged the location of industry in the heart of the

more lucrative northern market.








Thus the South entered the twentieth century in a

disadvantaged position. It was an area short of skilled

labour, investment capital, and entrepreneurs, technolog-

ically and productively lagging behind the other sections

of the country.


Recent Trends in the Economy

Despite these handicaps, great strides have been

made in the last 30 years throughout the study area and

the entire South. This has come about due to the mechaniza-

tion of agriculture, and better utilization of other natural

resources. The development of better educational facilities,

bi-racial labour force, more internal savings and capital

investments has encouraged the location of industry, espe-

cially those that benefit by decentralization.12

The trend has been to a diversification of both the

agricultural and industrial segments of the economy. Exam-

ples are the growth of the dairy industry in Mississippi,

the establishment of beef cattle on the Black Prairie of

Alabama and the Piedmont of Georgia, the location of the

broiler industry since the war in northern Alabama and

Georgia. Abandoned cropland, formerly in cotton, has been

reverting or restocked to pine woodland. Forest management

has been initiated and encouraged by the rapid expansion of

the pulp and paper industry. Better conservation practices,

increased use of fertilizers and mechanization of farms










have resulted in higher yields. Prices of important cash

crops such as cotton, tobacco and peanuts have been stabi-

lized by a system of acreage allotments.

Industrial growth has been spurred by the establish-

ment of large important industries. Some major examples

are: the atomic energy plant near Aiken, South Carolina;

the Redstone guided missile project and other space facilities

at Huntsville, Alabama, the large Lockheed aircraft plant in

Marietta, Georgia. Less dramatic has been the location of

electronic firms in the Golden Triangle of North Carolina.

The location by northern companies of branch firms in south-

ern cities has aided in the establishment and diversifica-

tion of industrial growth. Atlanta owes a great deal of

its growth to the establishment of these regional branches.

It has become an assembly center for Ford and General Motors.

Many insurance and airline firms have their regional offices

here. In addition the overall economic growth in the South

has been aided by the establishment of large military bases

and the rapid increase in tourism, both of which have been

attracted by the mild climate.


Outline of the Analytical Model

Briefly the steps undertaken in the study are as

follows: (1) transfer raw data into an I x N matrix (M

variables and N observations), (2) perform principal com-

ponents analysis of the M x N intercorrelation matrix of









the M variables and rotation of the resulting eigenvectors

to a normal varimax position to satisfy the criterion of a

simple structure, (3) compute the factor scores of N obser-

vations on the R rotated factors, (4) regress a dependent

variable D on the factor scores.



The Model

The mathematical model used to determine the under-

lying dimensions of economic health is known as a principal

axis factor analysis solution.13 The choice of the model

is based upon several considerations. The study of the

covariance of economic, social and demographic character-

istics of regions is a problem of a multivariate nature.

Data of this nature are often not-normally distributed and

are highly interrelated. Regression models using a large

number of variables lacking the assumptions of normality

and independence will give a coefficient of determination

which is biased upwards. In addition, the coefficients and

standard errors of the coefficients may be affected. Fac-

tor analysis, however,can be used on data lacking these a

priori assumptions. This technique can simultaneously

manage a large number of variables, compensate for random

error, and disentangle complex interrelations into their

major and distinct regularities.14 The essence of this

technique is to reduce a large number of variables into a

smaller group of distinct orthogonol factors, components,









or dimensions with a minimum loss of information. On the

basis of these factors the unique variations of a domain can

be discerned. A further step is the obtaining of factor

scores which is the location in space of each observation

on the particular dimensions. These factor scores can be

used as independent variables regressed on any desired
15
dependent variable. This method is suggested by Scott.5

Scott contends that when the assumptions of independence

and normality are not known, this regression technique is

more accurate than the classical least squares technique.16

The essential features of factor analysis are as

follows:17 from a rectangular X matrix of raw data it is

possible to compute a R matrix of simple correlation coef-

ficients.

xll x12 ... xln
x21 x22 ... x2n
xml xm2 ... xmn

XX' = R

rll r12 ... rln
R r21 r22 ... r2n

rn1 rn2 ... rnn

The problem in factor analysis is to find a matrix A such

that R = AA'. This is done by solving for, using an iterative

process, the eigenvalue or characteristic roots L, where L

is the sum of the squared factor loadings of the N variables

on a dimension. The factor loadings are the correlations









between the variables and the dimensions. One can solve

for as many characteristic roots as there are variables;

however, one generally only solves for the most important

roots, that is for the ones that account for the majority

of the variance. The equation for solving L is (R LI) =0

where I is the inverse of R. After finding the eigenvalue

L one can solve for the eigenvector V which is the column

vector of factor loadings, by the following equation (R =

LI) V = 0. Repetition of this procedure yields the factor

matrix A.


all a12 ... aml
a21 a22 ... am2 r is the number of character-
A .. ... istic roots solved for.
aml ... ... amr

Dividing the characteristic roots found by the total com-

munality (communality is the sum of independent common

variances of the components on each variable) yields the

variance accounted for by each dimension.

It is often of interest to find the factor scores,

that is the ranking of each observation on each of the

derived components. The matrix of factor scores is given

by the following equation F- XAL'. In addition the factors

are rotated so as to comply with Thurston's concept of sim-

ple structure, such that each of the original variables

relates highly with one and only one of the new dimensions.l8

This assures that the first factor arising from the analysis









accounts for the maximum variance and each successive factor

accounts for decreased proportions of the explained variance.

The factor scores have further application in a

regression model to explain the per cent of variation of

a dependent variable which is accounted for by the orthogonal

dimensions. This procedure produces an index which has the

largest possible variance of any linear combination. A

lower coefficient of determination is obtained in compari-

son with the least squares method. Factor regression

accounts for errors in the variable high intercorrelations

and gives regression coefficients which are more accurately

representative of economic theory.19



Variables Used

The variables used for the two time periods were

chosen to represent social, economic, demographic and

locational characteristics of the study area. In all, 26

indices were used. They were obtained from the censuses

of population, agriculture, and the city and county data

book for the respective time periods. Three locational

indices were calculated for the study. The following list

is provided to show the nature of the variables.








Table 4

Variables Used


1 Total population
2 Per cent of the population urban
3 Per cent of the population nonwhite
4 Per cent of the population 65 years old or older
5 Per cent of the population 25 years old or over who
completed less than five years of school
6 Per cent of the population 25 years old or over who
completed high school
7 Per cent of housing units sound
8 Median value of owner-occupied dwelling units
9 Median gross rent
10 Value added by manufacturing
11 Total capital expenditures by manufacturing firms
12 Total retail sales
13 Total wholesale sales
14 Per cent of the population under five years of age
15 Per cent of the labour force unemployed
16 Per cent of the farms operated by tenants
17 Average value of land and buildings per farm
18 Total value of farm products produced (sold)
19 Number of commercial farms earning less than $2,500
annually
20 Number of farmers whose off-farm income is greater
than farm income
21 Per cent of families earning between $0 $999 annually
22 Per cent of families earning between $1,000 $1,900
annually
23 Per cent of families earning between $2,000 $2,999
annually
24 Distance in miles from the center of a county to the
nearest city with a population of at least 50,000
25 Distance in miles from the center of a county to the
nearest city with a population of at least 500,000
26 Distance in miles from the center of a county to New
York City


Through these variables an attempt was made to strike

a balance between the rural and urban segments of the study

area. This choice also took into account that the majority

of these indices have been found to be significant by other

researchers, in explaining the variation in economic health.









The major problem with the data is that the areal units

(counties) vary immensely in size. Large counties having

extensive variation within them may not accurately reflect

this in the data. In spite of this inherent difficulty,

county data are at present the best available for such an

analysis.



Summary

This chapter is divided into two sections. The

first describes the study area and the historical basis

for its poverty. This region was selected for analysis

as it is representative of the southeastern United States.

Furthermore, it appears to have had a substantial number

of economic changes within it in the last few decades, and

thus would be an ideal testing area for the analysis of

spatial and temporal patterns of economic health.

Factor analysis was chosen as an appropriate model

because it can handle nonnormal and highly intercorrelated

data. In addition it opens up other avenues of research

beyond the location and identification of the underlying

dimensions.












REFERENCES


John Fraser Hart, The Southeastern United States
(Princeton: D. Van Nostrand Co., Inc., 1967).

Rupert B. Vance, Human Geography of the South
(Chapel Hill: The University of North Carolina Press, 1935).

3Howard Wd. Odum, Southern Regions of the United States
(Chapel Hill: The University of North Carolina Press, 1936).

4G. Langdon White and Edwin J. Foscue, Regional Geog-
raphy of Anglo-America, 2nd edition (Englewood Cliffs:
Prentice-Hall, 1956).

5Paul F. Griffin, Robert N. Young and Ronald L. Chatham,
Anglo-America A Regional Geography of the United States and
Canada (San Francisco: Fearon Publishers Inc., 1962).

Donald J. Bogue and Calvin L. Beale, Economic Areas
of the United States (New York: The Free Press of Glencoe,
Inc., 1961), p. ALII.

7Ibid., pp. 269-270.

Clement Eatoh, A History of the Old South (New York:
The I.acmillan Co., 1965).

James G. MIaddox, et al., The Advancing South: Man-
power Proseccts and Problems (New York: The Twentieth-
Century iFund, 1967).

10
lWilliam E. Laird and James R. Rinehart, "Exogeneous
Check on Southern Economic Development," South Atlantic
.Quarterl, Vol. LXV, No. 4 (1966), pp. 491-5-08

Ibid.

12B. U. Ratchford, "Economic Development in the South,"
South Atlantic Quarterly, Vol. LXIV, No. 4 (1965), pp. 496-
505.
43 "





44



13The type of factor analysis used in this study is
the version known as principal component analysis.

14R. J. Rummel, "Understanding Factor Analysis,"
Conflict Resolution, Vol. XI, No. 4 (1967), p. 444.

15John T. Scott, Jr., "Factor Analysis and Regres-
sion," Econometrica, Vol. 34, No. 3 (July, 1967), pp. 552-
562.

16Ibid.

17For a more detailed explanation of the model see:
H. H. Harmon, Modern Factor Analysis (Chicago: University
of Chicago Press, -TT)- T .; i. G Kndall, A Course in Multi-
variate Analysic (London: Charles Griffin, 957), pp. 10-
36; R. J. Rummel, "Understanding Factor Analysis, Conflict
Resolution, Vol. XI, No. 4 (1967), pp. 444-480; Mary Megee,
"On Economic Growth and the Factor Analysis Method," South-
ern Economic Journal, XXXI (July, 1964), 215-228.

18Qazi Ahmad, Indian Cities: Characteristics and
Correlates (Chicago: University of Chicago Press, 1965),
p. 25.


19Scott, op. cit., p. 552,













CHAPTER III

THE DIMENSIONS OF ECONOMIC HEALTH IN THE
SOUTHEASTERN UNITED STATES (1950-1960)


The analysis as reported in this chapter as well

as in the subsequent chapter, was undertaken at the Univer-

sity of Florida's computing center. The original data, for

the two time periods, were transformed into symmetric 26 by

26 correlation matrices. These matrices were then subjected

to a principal axis solution which yielded six eigenvectors

for the 1950 data and seven eigenvectors for the 1960 data,

based upon the criteria oflAI. The components were then

rotated to a normal varimax position. The underlying dimen-

sions obtained in this manner were used to identify and to

help interpret the areal variation in economic health.

The basic dimensions of economic health are identi-

fied by the analysis of the derived factor loadings and fac-

tor scores. Factor loadings are the correlations of the

original variables with the newly derived dimensions. If

each of the original variables were plotted in n-dimensional

space (n representing the number of derived factors), the

coordinates of the variables would represent the correlations

or loadings on each principal axis (dimension) by the indi-

vidual variables. These factor loadings range from 1.0,

indicating a high positive correlation with a given dimension
45









(which must be interpreted), to a -1.0, indicating a high

negative correlation. Variables with high positive loadings

on a factor are expected to be most characteristic of the

quality represented in the factor, while high negative load-

ings on the same factor should indicate a contrasting qual-

ity. Factor scores on the other hand measure how closely

each observation (county) ranks on each of the individual

dimensions. For example, a high positive factor score for

a county on a given component can be interpreted to mean

that the county is highly representative of that component.

With this in mind, it becomes possible to identify, and

interpret the areal variation of economic health. In order

to facilitate interpretations, of the dimensions only high

loadings on each of the components are shown, as are only

highly negative and positive factor scores in Tables 7-18,

21-34. The factor scores will also be employed later in

this chapter in a stepwise regression as a means of testing

how well the derived dimensions explain the variation in

poverty.


1950 Factor Analysis

Table 5 summarizes the proportion of the total inter-

county variance in 1950 accounted for by each of the six

derived factors. The six rotated factors together explain

70.5 per cent of the total variance of the 26 variables.

Thus these six components are considered representative of

the variables that they replaced.









Table 5

Percentage of Total Variance Explained by Each
Dimension in 1950


Dimension

I
II
III
IV
V
VI


Eigenvalue

8.71
3.09
2.14
1.91
1.45
1.02


Total


Per cent of total variance

33.5
11.9
8.2
7.4
5.6
3.9

70.5


Another indication of the representation of the

original variables by the ncw dimensions is given by the

communalities. These values are the total sum of squares

of the correlations of a particular variable with all six

components. As would be expected, all variables are not

equally represented (Table 6).


Table 6

Percentage of Variance (Communality) of each of the 26
Variables Accounted for by All Six Components in 1950

Variables Communality in Percentage

1 96.6
2 74.5
3 79.1
4 19.3
5 83.0
6 82.2
7 87.1
8 53.4
9 75.5
10 84.2
11 59.6
12 77.2
13 69.4











Variables

14
15
16
17
18
19
20
21
22
23
24
25
26


Table 6--Continued

Communality in Percentage

78.1
10.9
80.2
67.9
84.7
87.9
69.6
93.2
95.5
46.3
39.4
63.6
75.1


Two variables, per cent of the population 65 years of age

and older and per cent unemployed, are poorly accounted for

as only 19.3 and 10.9 per cent of their total sums of squares

are explained. This implies that these variables were not

significant in the analysis. On the other hand the remain-

ing variables are summarized quite adequately by the new

dimensions. Thus the variations of the selected indices

are effectively captured by this statistical technique, a

desirable trait in such an analysis.


The 1950 Dimensions

Factor I (Urban Industrial Counties), the most

significant dimension, contributing 33.5 per cent of the

total variance, isolates the urban ghettos of the Southeast.

This component is associated with poverty found in some of

the urban industrial complexes within this region. Table 7

shows the correlations between this dimension and urban









functions such as retail and wholesale sales. High loadings

are also associated with total population, value added by

manufacturing, capital expenditures, poverty and youthful-

ness of the population. Counties which rank high on this

dimension are those which contain large rapicdy growing

regional centers such as Atlanta, Birmingham, LMobile, and

Charlotte (Table 8 and Figure 2). These centers, due to

their prosperity and rapid rate of economic growth, attract

a large number of immigrants from the hin'terlan i, many of

whom are unskilled and consequently are employed in low

paying industries or seasonal work.


Dime vision I


Table 7

Urban Industrial Counties in 1950


Prima y V riables F

Total population
Per cent of population urban
Value added by manufacturing
Total Capital expenditures
Retail sales
Wholesale sales
Number of persons under five year:, of age
Per cent of families earning between
0--999
Per cent of families earning between
$1,000- i,999
Per cent of families earning between
$2,000-$2,999


actor Loadings

.92
.44
,88
.73
.85
.82
.83

.76

.90


.65









Table 8

Counties with Extreme Factor Scores on Dimension I
in 1950


High Positive

Jefferson, Ala.
Fulton, Ga.
Mobile, Ala.
Mecklenburg, N. C.
Forsythe, N. C.
Guilford, N. C.
Chatham, Ga.
Greenville, S. C.
Charlestovwn, S. C.
Spartanburg, S. C.
Gaston, N. C.
Fayette, Ga..
Bibb, Ga.
Richland, S. C.
Montgomery, Ala.
Buncombe, N. C.
Hinds, Miss.
Wake, N. C.
Durham, N. C.
Richmond, Ga.


12.32
11.84
3.94
3.66
3.51
3.10
2.78
2.52
2.33
2.29
1.97
1.59
1.59
1.50
1.48
1.48
1.40
1.38
1.35
1.30


High Negative

Chattahooche, Ga.
Fannin, Ga.
Houston, Ga.
Carteret, N. C.
Dare, N. C.
Onslow, N. C.
Lafayette, Miss.
Ohowan, N. C.
Stone, Miss.
Orange, N. C.


The second dimension which explains 13.3 per cent

of the total variance is a rural poverty indicator. This

component is related mainly to types of farming and sources

of farm income. High positive loadings are found on total

value of farm products produced and number of commercial

farms earning less than $2,500 annually, the tenancy rate,

and number of farmers whose off-farm income is greater than

farm income (Table 9). This dimension isolates the poor

tenant and the small individual farmer who, because of lack

of finances, education, or the small scale of his holding,

has not been able to make a financial success of his


-2.18
-1.34
-1.03
- .82
- .81
- .78
- .76
- .71
- .71
- .70













+ +


+
0


cn) CM

ED


'
':'''
'''''
''''
'''


C


C

o

L.
0


0) 0


I iv









enterprise. Factor scores fail to reveal any easily defined

grouping of counties. Instead they suggest that this type

of poverty is quite widespread, especially in areas outside

the dominance of any large metropolitan center (Table 10).

There does appear, however, to be a concentration of poor

counties in the old cotton belt of the Southeast, suggesting

that these rural counties have not fully recovered from the

negative effects of the sharecropping system of agriculture

(Figure 3).


Dimension II


Table 9

Tenant and Small Farm Poverty in 1950


Primary Variables Factor Loadings

Per cent of farmers that are tenants .49
Total value of farm products produced (sold) .83
Number of commercial farms earning less than
$2,500 annually .93
Number of farmers \h'ose off-farm income is
greater than their farm income .39
Per cent of families earning between
$0-$999 .55


Table 10

Counties with Extreme Factor Scores on Dimension II
in 1950


High Positive

Bolivar, Miss.
Sunflower, MIiss.
Robeson, N. C.
Leflore, Miss.
Washington, Miss.
Coahoma, Miss.
Cullman, Ala.
De Kalb, Ala.
Johnston, N. C.


5.38
5.30
3.61
3.38
3.28
3.15
3.09
3.08
3.00


High Negative

Evans, Ga.
Glynn, Ga.
Camden, Ga.
I:uscogee, Ga.
McIntosh, Ga,
Chatham, Ga.
Liberty, Ga.
Fulton, Ga.
Charlton, Gao


-2.43
-1.93
-1.89
-1.64
-1.60
-1.59
-1.52
-1.51
-1.43








Table l0--Continued

High Positive High Negative

Sampson, N. C. 2.76 Long, Ga. -1.41
Horry, S. C. 2.74 Bryan, Ga. -1.39
Orangeburg, S. C. 2.40 Camden, N. C. -1.37
Tallahatchie, Miss. 2.36 Currituck, N. C. -1.37
Columbus, N. C. 2.36 Doughert,, Ga. -1.33
Florence, S. C. 2.34 Clink, Ga -1.32
Williamsburg, S. C. 2.22 Turner, Ga. -1.32
Marshall, Ala. 2.15 Peach, Ga. -1.31
Pitt, N. C. 2.15 Dare, N. C. -1.28
Quitman, Miss. 2.09 Hancock, Miss. -1.28
Cleveland, N. C. 2.05 Jones, Ga. -1.26
Tunica, Mirs. 2.02 Wayne, Ga. -1.26


The third most important factor (Wealthy Urbanized

Counties) identifies a pattern of urban wealth. It is

associated with variables, such as per cent of the popula-

tion urban, median value of owner-occupied homes, and per

cent of the population that completed high school (Table 11).

The economic influence of Atlanta is clearly seen in De Kalb,

Clayton and Cobb which rank high on this dimension. These

counties are to an extent dormitory suburbs of Atlanta and

as such contain a high proportion of residents with above

average incomes. Many of the other counties listed in Table

12 owe their prosperity to the location of state institutions,

an example is Orangeburg, South Carolina. Other counties

that fall into this category are Hinds, Missi'sippi; Mont-

gomery, Alabama and Richland, South Carolira. The influence

of state spcdii g is felf-evident in these counties as the

respective state capitals are located there. Federal pay-

rolls are another important source of prosperity, for example










CM m +
+ + A



- I,
0
o 0
*- O


L~jiwI I
0n


0


44 z


^


O
d Ir,
6 cs\
L4









Houston, Chattahooche and Muscogee counties in Georgia, the

sites of large military bases. Prosperity in these urban-

ized counties can be traced to concentrations of highly edu-

cated people working in professions that pay above average

salaries.

Table 11

Dimension III Wealthy Urbanized Counties in 1950


Primary Variables


Factor Loadings


Per cent of the population urban
Per cent of the population that completed
less than five years of school
Median value of owner-occupied homes
Average value of farms (land and buildings)
Per cent of the population that completed
high school


.73

-.43
.70
.55

.82


Table 12
Counties with Extreme Factor Scores on Dimension III
in 1950


High Positive

Chattahooche, Ga.
De Kalb, Ga.
Harrison, Miss.
Glynn, Ga.
Houston, Ga.
New Hanover, N. C.
Dougherty, Ga.
Muscogee, Ga.
Forrest, Miss.
Clarke, Ga.
Adams, Miss.
Hinds, Miss.
Montgomery, Ala.
Richland, S. C.
Jackson, Miss.
Orange, N. C.
Pasquotank, N. C.
Chatham, Ga.


5.54
4.39
3.72
3.71
3.33
3.08
3.06
2.98
2.73
2.71
2.70
2.62
2.44
2.35
2,31
2.24
2.16
2.11


High Negative

Fayette, Ga.
Jefferson, Ala.
Dawson, Ga.
Jasper, S. C.
Heard, Ga.
Union, Ga.
Stokes, N. C.
Forsythe, Ga.
Yancey, N. C.
Berkeley, S. C.
Calhoun, S. C.
Banks, Ga.
Echols, Ga.
Lowndes, Ala.
Fulton, Ga.
Gilmer, Ga.
Greene, Ala.
Clebourne, Ala.


-2.25
-2.16
-1.85
-1.53
-1.42
-1.40
-1.35
-1.35
-1.31
-1.31
-1.31
-1.30
-1.30
-1.29
-1.28
-1.27
-1.26
-1.26










Dimension IV (Rural White Counties) shows the con-

trast in levels of living between primarily white and black

counties. Variables that correlate highly negative with

this component are: per cent of the population nonwhite,

per cent completed less than five years of school and per

cent of farmers that are tenants (Table 13). A positive

correlation is found between this dimension and number of

farmers whose income from off-farm employment is greater

than faii income. High positive factor scores are asso-

ciated with the Appalachian counties of Alabama, Georgia,

North and South Carolina, with the heaviest concentration

in western North Carolina (Table 14). The predominantly

Negro counties rank low on this factor (Table 14). These

Negro counties are located in a belt which runs from east-

ern North Carolina to the Mississippi Delta. The largest

concentrations are found in the Delta, southwestern Georgia

and the tobacco area of eastern North Carolina. The major-

ity of the Negroes in this belt are the descendants of the

former slaves who worked on the cotton and tobacco planta-

tions. Thus this area of colored poverty has its roots in

the plantation system of land tenure, whose remnants still

remain. Although both groups of counties are poor, the

Negro counties are more so, indicating the traditional

disparity between the black and the white segments of our

society.









Table 13


Dimension IV


Rural White Counties in 1950


Primary Variables


Factor Loadings


Per cent of population nonwhite
Per cent of farmers that are tenants
Per cent of population that completed less
than five years of school
Number of farmers whose income from off-farm
employment is greater than farm income


Table 14


Counties with Extreme Factor
in 1950


High Positive


Wilkes, N. C.
Cullman, Ala.
Walker, Ala.
Ashe, N. C.
Fannin, Ga.
Madison, K. C.
Caldwell, N. C.
Randolph, N. C.
Greenvill S. C.
De Kalb, Ala.
Watauga, N. ,.
Buncombe, N. C.
Catawaba, N. C.
Jackson, N. C.
Cherokee, N. C.
Spartanburg, S. C.
Davidson, N. C.
Burke, N. C.
Mitchell, N. C.


2.30
2.25
2.13
2.06
2.01
1.99
1.89
1.89
1.82
1.80
1.77
1.66
1.66
1.65
1.63
1.59
1.58
1.55
1.55


The last two components


Scores on Dimension IV


High Negative


Lee, Ga.
Union, Miss.
Calhoun, Ga.
Bur: e, Ga.
Turner, Ga.
Coahoma, Miss.
Terrel, Ga.
Peach, Ga.
Lowndes, Ala.
Stewart, Miss.
Issaquena, Miss.
Greene, Ala.
Baker, Ga.
Washington, Miss.
Bullock, Ala.
Edgecombc, N. C.
Quitman, Ga.
Webster, Ga.
Adams, Miss.


-2.56
-2.54
-2.20
-2.02
-2.01
-2.01
-1.92
-1.84
-1.81
-1.79
-1.78
-1.77
-1.73
-1.713
-1.71
-1.70
-1.66
-1.63
-1.62


together contribute 10.1 per


cent to the total explained variation. These two dimensions

are difficult to identify. They appear to be indicators of

urban influence upon their hinterlands (Tables 15 through

18). Factor V indicates that a close proximity to a large


-.85
-.72

-.78

.60









city may not always be beneficial. For a small urban center

generally is at a disadvantage competing with a large

regional center for a regional market, especially if the

regional center contains a large proportion of the regional

market. A large city in this instance may cause services

and industries to gravitate to it. This may be true of the

southwestern Mississippi counties that rank highly negative

on Dimension V (Table 16), These counties are depressed

economically not only because of their poor agricultural

base but also due to the lack of thriving urban centers.

In this case the presence of a large urban center is not

beneficial, rather it appears that New Orleans is growing

at the expense of this area. This in turn suggests that New

Orleans is losing its national significance to other more

rapidly growing centers such as Atlanta. This is in con-

trast to the more prosperous counties of North Carolina,

whose produce is related more to national markets. Thus

this dimension appears to reflect the importance of national

markets to economic growth.

The last dimension is another locational index as it

has negative loadings with all the distance variables (Table

17). The positive loading with median value of owner-occupied

dwellings is an indicator of wealth. This factor reflects

a locational advantage in regard to regional and local mar-

kets, A glance at the factor scores will verify this loca-

tional advantage. For example. counties with high positive








scores arelfound to be centered near large metropolitan cen-

ters (Table 18). The economic effect of Atlanta is clearly

seen on the Georgia counties. The counties with high nega-

tive factor scores tend to be more isolated and economically

depressed.

Table 15

Dimension V Locational Advantage in Relation to
National Market (New York) in 1950


Primary Variables


Factor Loadings


Per cent of population 65 years of
age or older -.31
Median gross rent .31
Avel-ge value of farm (land and buildingr) .48
Total value of farm products produced .27
Number of farmers hose incon from off-
farm employment is greater than farm income -.22
Distance in miles fron- the center of a county
to a city of at lear b 500,000 .66
Distance in miles from the center of a county
to New York -.81

Table 16

Counties with Extreme Factor Scores on Dimension V
in 1950


High Positive

Horry, S. C.
Yadkin, N. C.
Glynn, Ga.
Alleghany, N. C.
Robeson, N. C.
Surry, N. C.
Columbus, N. C.
Stokes, N. C.
Davie, N. C.
Harnett, N.C.
Forsythe, N. C.
Pitt, N. C.
Florence, S. C.
Cabarrus, N. C.
Dillon, S. C.
Camden, Ga.
Nash, N. C.


2.22
2.18
2.05
2.05
2.00
1.97
1.95
1.94
1.81
1.79
1.79
1.78
1.74
1.72
1.68
1.68
1.63


High Negative

Evans, Ga.
Pike, Miss.
Hinds, Miss.
Harrison, Miss.
Lincoln, Miss.
Copiah, Miss.
Amite, Miss.
Wilkinson, Miss.
Walthall, Miss.
Marion, Miss.
Forrest, Miss.
Jefferson, Miss.
Lafayette, Miss.
Marshall, Miss.
Tuscaloosa, Ala.
Jefferson Davis,
Dallas, Ala.


-2.55
--2.08
-2.07
-1.99
-1.97
-1.89
-1.80
-1.74
-1.73
-1.73
-1.71
-1.68
-1.68
-1.61
-1.61
Misse-l.56
-1.56










Table 17

Dimension VI Locational Advantage in Relation to
Local Regional and National Markets in 1950


Primary Variables


Factor Loadings


Median value of owner-occupied
dwelling units
Median gross rent
Distance in miles from the eel'ter of a
county to a city of at least 50,000
Distance in miles from the center of a
county to a city of at least 500,000
Distance in miles from the center of a
county to New York


Table 18


Counties with Extreme Factor
in 1950


High Positive

Fannin, Ga.
Fayette, Ga.
De Kalb, Ga.
Edgecombe, N. C.
Clayton, Ga.
Wilson, N. C.
Turner, Ga.
Hertford, N. C.
Evans, Ga.
Tunica, Miss.
Nash, N. C.
Wake, N. C.
Greene, N. C.
De Soto, Miss.
Henry, Ga.


12.76
2.90
2.15
2.02
1.95
1.86
1.86
1.81
1.76
1.75
1.69
1.66
1.63
1.55
1.52


Scores on Dimension VI


High Negative


Lowndes, Ga.
Georgetown, S. C.
Issaquena, Miss.
Wayne, Ga.
Choctow, Ala.
Washington, Ala.
Clarke, MIiss.
New Hanover, N. C
Columbus, N. C.
Marion, S. C.
Louderdale, Miss.
McIntosh, Ga.
Choctow, Miss.
Jasper, Miss.
Geneva, Ala.


-2.51
-1.84
-1.79
-1.76
-1.76
-1.64
-1.62
-1.55
-1.52
-1.46
-1.44
-1.44
-1.43
-1.42
-1.41


1960 Ana ysis

In 31950 seven 'imen3ions (,nrrg d from the factor

analysis solution. T1 so seven components explain 77.9 per

cent of the total variance (Table 19). The higher percentage


.69
.22

-.47

-.43

-.26









of explanation over the 1950 period is due to the additional

dimension and a higher percentage accounted for by the other

dimensions. A check of the communalities reveals that the

higher percentage of explained variation is due to a better

factoring of the variables, including old age and unemploy-

ment, which were poorly accounted for in 1950 (Table 20).


Dimen s

I
I]
II
IV
V
VI
VI

Total


Table 19

Percentage of Total Variance Explained by Each
Dimension in 1960

pion Eigenvalue Per cent of tot,

8.80 33.8
[ 3.44 13.3
I 2.40 9.2
J 1.86 7.2
1.47 5.7
1.13 4.4
I 1.12 4.3

77.9


al variance


Table 20

Percentage of Variance (Communality) of Each of the 26
Variables Accounted for by All Seven Components in 1960

Variables Communality in percentage

1 97.1
2 65.9
3 77.6
4 92.4
5 85.4
6 78.0
7 91.3
8 89.5
9 83.5
10 87.1
11 58.9
12 81.5
13 67.2
14 67.2









Table 20--Continued

Variables Communality in percentage

15 73.1
16 74.8
17 68.6
18 65.1
19 83.9
20 75.5
21 91.6
22 95.4
23 94.2
24 57.2
25 54.2
26 69.2


The 1960 Dimensions

The largest dimension, as in 1950, is labelled

urban industrial counties. It contributes 33.8 per cent of

the total variance. High positive correlations are found

between this dimension and population, value added, capital

expenditures, retail sales, wholesale sales, and poverty

(Table 21). The concentration of the poor in the industrial

area is not unexpected. As employment diminishes in the

rural areas due to increased mechanization and larger land

holding the only promise of jobs lies in the growing indus--

trial and service sectors of the economy. Consequently

there is a large migration from the country to the cities.

At times the flow is too great to be absorbed at once into

the economy, resulting in temporary unemployment. On the

other hand, due to their lack of skills or color, many of

those who are absorbed by the labour market are engaged in

low paying occupations, especially in some branches of the

service sector of the economy. Consequently the faster a









city grows and the larger it becomes the larger the concen-

tration of urban poverty. For as a city achieves a higher

rank in the hierarchy of urban systems, it assumes more

service functions, many of which are low paying. In addi-

tion the expansion of urban functions often makes the city

more attractive to future immigrants; this then compounds

the problem of absorbing these migrants into the existing

labour force rapidly. Furthermore, if the supply of labour

greatly exceeds the demand, wages may remain low unless the

workers are unionized or covered by wagc laws. This notion

is suggested by the county factor scores in Table 22. It

is observed that the cities ranking highest on the urban

industrial poverty dimension are also the largest in popula-

tion (Figure 4).

Table 21

Dimension I Urban Industrial Counties in 1960

Primary Variables Factor Loadings

Total population .93
Per cent of the population urban .43
Value added by manufacturing .84
Total capital expenditures .73
Retail sales .80
Wholesale sales .80
Per cent of families earning between
$0 $999 .80
Per cent of families earning between
$1,000 $1,999 .90
Per cent of families earning between
$2,000 $2,999 .91









Table 2

Counties with Extreme Factor
in 1960


High Positive

Fultan, Ga.
Jefferson, Ala.
hecklenburg, N. C.
Mobile, Ala.
Guilford, N. C.
Forsythe, N. C.
Greenville, S. C.
Chatham, Ga.
Charleston, S. C.
Spartanburg, S. C.
Gaston, N. C.
Richland, S. C
Hinds, Iliss.
De Kalb, Ga.
Etowah, Ala.
Jackson, liss.
Buncombe, N. C.
Wake, N. C.


11.99
11.33
4.89
4.13
3.63
3.12
2.90
2.58
2.53
2.48
2.08
1.90
1.82
1.62
1.61
1.57
1.51
1.46


2

Scores on Dimension I


High Negative

Houston, Ga.
Chattahoochee, Ga.
Clayton, Ga.
Catoosa, Ga.
Lafayette, Miss.
Hancock, Miss.
Lumpkin, Ga..
Stone, 1iss.
Columbia, Ca.
Walthall, ILiss.
Candler, Ga.
Liberty, Ga.
Watauga, N. C.
Lamar, Ga.
George, Miss.
Oconee, Ga.
Wayne, Ga.
Polk, N. C.


The second most important dimension contributes 13.3

per cent to the total variance. This component is called

Rural Negro Counties as it relates to the poverty found in

black rural counties. It has high factor loadings on per

cent of the population nonwhite, per cent of the population

who have completed less than five years of school, per cent

of the population under five years of age and per cent of

farmers who are tenants (Table 23). The other factor load-

ings also emphasize the rural nature of this dimension. The

ranking of factor scores clearly indicates the differential

in levels of living between the white and nonwhite (Table

24). High positive scores are associated with predominantly


-1.17
- .99
- .97
- .92
- .86
- .75
- .71
- .70
- .70
- .67
- .67
- .67
- .67
- .66
- .62
- .61
- .60
- .60












*: o

0 4.

00













<0







cZ -

< Z

L,
QO ^^
Z c^_____ W









Negro counties. These counties are located in an arc that

runs from eastern North Carolina through the old cotton belt

to the Mississippi Delta where the heaviest concentration

is found. White counties with low factor scores are gen-

erally found in the Appalachian region of the study area.

The distribution of the counties is similar to that found

in 1950.

The third dimension is also an indicator of rural

poverty but it is not related to color. This component

reflects the residual poverty left over from the sharecrop-

ping era. This group of rural poor is made up of tenant

and small commercial farmers who have not been able to cope

with the changing trends in agricultural production (Table

25). In many instances their deprivation is attributed to

the lack of education, capital and land resources. Although

this form of poverty is widespread (Table 26), Figure 5

indicates that it is most prevalent in the areas where share-

cropping prevailed.


Table 23

Dimension II Rural Negro Counties in 1960

Primary Variables Factor Loadings

Per cent of the population nonwhite .82
Per cent of population that completed
less than five years of school .72
Per cent of population under five years
of age .77
Per cent of farmers who are tenants .74
Average value of farm (land and buildings) .47









Table 23--Continued


Primary Variables


Factor Loadings


Total value of farm products produced
Number of farmers whose off-farm income
is greater than farm income
Per cent of dwelling units sound


.35

-.48
-.45


Table 24

Counties with Extreme Factor Scores on Dimension II
in 1960


Hi gh Positive

Tunica, Miss.
Coahoma, Miss.
Bolivar, Miss.
Washington, Miss.
Sunflower, Miss.
Leflore, Miss.
Sharkey, Miss.
Quitman, Miss.
Humphreys, Miss.
Tallahatchie, Miss.
Lee, Ga.
De Soto, Miss.
Lowndeo, Ala.
Burke, Ga.
Pitt, N. C.


High Negative


3.55
3.15
3.10
3.09
3.02
2.91
2.66
2.44
2.23
2.20
2.04
2.03
2.01
1.94
1.91


Ashe, N. C.
Wautauga, N. C.
Jackson, N. C.
Dare, N. C.
Chattahoochee, Ga.
Avery, N. C.
Macon, N. C.
Tishomingo, Miss.
Buncombe, N. C.
Towns, Ga.
Fannin, Ga.
Marion, Ala.
Alleghany, N. C.
iitchell, N. C.
Darlington, S. C.


Table 25


Dimension III


Tenant and Small Farm Poverty in 1960


Primary Variables


Factor Loadings


Per cent tenants
Average value of farms
Value of farm products produced
Number of commercial farms earning less
than $2,500 annually
Per cent of families earning between '0-$999
Per cent of families earning between $1,000-
$1,999
Number of farmers whose off-farm income is
greater than farm income


.38
-.38
.65

.89
.45

.26


.65


-1.92
-1.92
-1.87
-1.83
-1.81
-1.79
-1.79
-1.78
-1.77
-1.74
-1.73
-1.71
-1.66
-1.59
-1.58






68







rn+
+ + A















I I_
C









I
















Jo
O '























ZI


1 - :-.! -.. .-v..:..'-..









Table 26

Counties with Extreme Factor Scorer. on Dimension III
in 1960


Higp Positive

Robeson, N. C.
Cullman, Ala.
De Kalb, Ala.
Hinds, Miss.
Panola, Miss.
Dallas, Ala.
Williamsburg, S. C.
Sunflower, Miss.
Orangeburg, S. C.
Horry, S. C.
Columbus, N. C.
Sampson, N. C.
Marshall, Miss.
Madison, Miss.
Marshall, Ala.
Anderson, S. C.
Florence, S. C.


3.37
3.32
3.23
2.84
2.62
2.58
2.56
2.51
2.50
2.49
2,48
2.47
2.25
2.19
2.14
2.12
2.08


High Negative

Fulton, Ga.
Camden, Ga.
McIntosh, Ga.
Clinch, Ga.
De Kalb, Ga.
Jackson, Miss.
Charlton, Ga.
Peach, Ga.
Mecklenburg, N. C.
Muscogee, Ga.
Bryan, Ga.
Chatham, Ga.
Currituck, N. C.
Liberty, Ga.
Dougherty, Ga.
Lee, Ga.
Jones, Ga.


The fourth dimension (Locational Advantage in Rela-

tion to National Market) has high correlations with distance

to New York and distance to a city of at least 500,000 per-

sons (Table 27). Counties with extreme negative scores on

this dimension are generally concentrated in southwestern

Mississippi (Table 28). Other low factor scores are found

in central Alabama and northern Mississippi, These factor

scores, although they do to a certain extent reflect poverty,

reflect largely the proximity of these counties to large

regional centers such as Memphis, Birmingham and New Orleans.

It is of interest to note that no large regional grouping

appears around Atlanta. The absence of such a grouping

around Atlanta may be attributed to the closer proximity

of New York. On the other side of the scale are found the


-2.71
-2.53
-2.40
-2.05
-1.86
-1.80
-1.76
-1.74
-1.63
-1.59
-1.58
-1.55
-1.46
-1.42
-1.39
-1.39
-1.37










tobacco-producing and the furniture manufacturing counties

of North and South Carolina. Although these counties are

not in close proximity to regional markets they are close

to the national market. This favourable location is impor-

tant in regard to their specialization in the growing of

tobacco and in the manufacturing of cigarettes and furniture.

Furthermore these commodities have been less susceptible to

mechanization than other enterprises such as cotton growing.

This is in contrast to southwestern Mississippi which suf-

fers from being unable to compete effectively in the agri-

culture market. As a result there has been a large rever-

sion to forestland, especially with the coming in of large

pulp and paper companies. The surplus of labour resulting

from the dwindling agricultural base has resulted in a high

unemployment rate and out-migration.


Table 27

Dimension IV Locational Advantage in Relation to
National Market (New York) in 1960

Primary Variables Factor Loadings
Distance in miles to a city of at
least 500,000 persons .72
Distance in miles to New York -.78
Per cent of population 65 years of age
or over -.19
Median gross rent .29
Total value of farm products produced .21
Per cent of population nonwhite -.14
Per cent of population unemployed -.17
Distance in miles from center of county
to nearest city of at least 50,000 .22








Table 28

Counties with Extreme Factor Scores on Dimension IV
in 1960

High Positive Loadings Hijh Negative Loadings

Gilmer, Ga. 4.61 Pearl River, Miss.
Horry, S. C. 2.52 Harrison, Miss.
Florence, S. C. 2.43 Mobile, Ala.
Columbus, S. C. 2.20 Pike, Miss.
Dillon, S. C. 2.19 Hancock, Miss.
Robeson, N. C. 2.13 Copiah, Miss.
Darlington, S. C. 2.10 Tuscaloosa, Ala.
Alleghany, N. C. 2.10 Lincoln, Miss.
Marlsboro, N. C. 2.02 Hinds, Miss.
Marion, S. C. 1.87 Amite, Miss.
Yadkin, N. C. 1.85 Marion, Miss.
Surry, N. C. 1.83 Wilkinson, Miss.
Scotland, N. C. 1.81 Lawrence, Miss.
Alexander, N. C. 1.79 Jefferson Davis, Miss.
Carteret, N. C. 1.71 Walker, Ala.
New Hanover, N. C. 1.71 Adams, Miss.
Wilkes, N. C. 1.71 George, Miss.


-2.42
-2.08
-2.08
-2.04
-2.03
-1.93
-1.92
-1.87
-1.79
-1.78
-1.76
-1.73
-1.71
-1.70
-1.65
-1.62
-1.61


Dimension V labelled wealthy urbanized counties,

reflects urban prosperity. It has high factor loadings on

per cent of the population urban, per cent of population

completed high school, per cent of population with sound

housing, median value of owner-occupied homes and median

gross rent (Table 29). Counties which rank high on this

component are prosperous small cities such as Macon, Albany,

and Columbus in Georgia, Huntsville, Alabama and Wake Forest

in North Carolina (Table 30). Other counties that have high

positive scores are Cobb, Clayton and De Kalb in Georgia.

These counties contain some of the more prosperous suburbs

of Atlanta. Negative factor scores are associated with

counties both rural and urban, that have a high percentage

of low income groups (Table 30).








Table 29

Dimension V Wealthy Urbanized Counties in 1960


Primary Variables


Factor Loadings


Per cent of population urban
Per cent of population completed less
than five years of school
Per cent of population completed high
school
Per cent of population with sound housing
Median value of owner-occupied homes
Median gross rent
Retail sales
Average value of farm (Land and buildings)


.67

-.52

.80
.80
.70
.75
.39
.48


Table 30

Counties with Extreme Factor Scores on Dimension V
in 1960


Hig-h Positive Loadings


High Negative Loadings


De Kalb, Ga.
Houston, Ga.
Clayton, Ga.
Dougherty, Ga.
Harrison, Miss.
Cobb, Ga.
Muscogee, Ga.
Nadison, Ala.
Onslow, N. C.
Choctow, Ala.
Clarke, Ga.
Orange, N. C.
Montgomery, Alao
Wake, N. C.
Cumberland, N. C.
Camden, Ga.


4.12
3.69
3.52
3.31
3.10
3.08
3.05
2.91
2.84
2.69
2.60
2.35
2.31
2.23
2.15
2.01


Webster, Ga.
Jefferson, Ala.
Fulton, Ga.
Glascock, Ga.
Quitman, Ga.
Hancock, Ga.
Talbot, Ga.
Issaquena, Miss.
Stewart, Ga.
Banks, Ga.
Wheeler, Ga.


The last two dimensions like those in 1950 are dif-


ficult to identify.


Dimension VI appears to be another


locational index. This component is related to national, re-

gional and loca] market potentials. Table 32 suggests that


-2.45
-2.43
-2.43
-1.82
-1.74
-1.67
-1.58
-1.58
-1.54
-1.50
-1.50









counties which rank high on this factor are located close

to urban centers. They tend to have lower unemployment

rates and a younger population than counties which have a

less favourable location.

The last component is difficult to name. Table 33

shows that the salient variables correlated with this factor

are old age and median value of owner-occupied dwelling units,

and value added by manufacturing. This dimension may isolate

areas of declining prosperity, suggesting the out-migration

of the younger and better educated labour force. Although

Table 34 is not conclusive, it implies that out-migration

may be the key to the explanation of this dimension.



Table 31

Dimension VI Locational Advantage in Relation to
National, Regional and Local Markets in
1960

Primary Variables Factor Loadings

Distance in miles from the center of the
county to a city of at least 50,000 -.60
Per cent of the labour force unemployed -.83
Distance in miles from the center of the
county to a city of at least 500,000 -.10
Dif.tance in miles from the center of the
county to New York -.19
Per cent of the population 65 years of
age or older -.10








Table 32

Counties with Extreme Factor Scores on Dimension VI
in 1960


High Positive Loadings


Crawford, Ga.
Edgefield, S. C.
Guilford, N. C.
Baker, Ga.
Hertford, N. C.
De Kalb, Ga.
Wake, N. C.
Fayette, Ga.
Glascock, Ga.
Webster, Ga.
Candler, Ga.
Forsythe, N. C.


2.08
2.06
2.04
1.84
1.81
1.76
1.74
1.73
1.71
1.71
1.63
1.63


High Negative Loadings

Bolivar, Miss.
Washington, Miss.
Pamlico, N. C.
Carteret, N. C.
Clay, N. C.
Hyde, N. C.
Avery, N. C.
Brunswick, N. C.
Rabun, Ga.
Pearl River, Miss.
Jackson, Miss.
Adams, Miss.
Jasper, Miss.
New Hanover, N. C.
Pitt, N. C.


-3.85
-3.28
-3.17
-3.10
-2.88
--2.88
-2.72
-2.56
-2.36
-2.31
-2.12
-2.29
-2.02
-2.00
-2.00


Table 33

Dimension VII Areas of Declining Prosperity
in 1960


Primary Variables

Per cent of the population 65 years
of age or older
Median value of owner-occupied
dwelling units
Median gross rent
Value added by manufacturing


Factor Loadings


-.89

-.54
.30
-.31


Table 34


Counties with Extreme Factor Scores on Dimension VII
in 1960


High Positive Loadings

Berkeley, S. C.
Liberty, Ga.
Hinds, Hiss.
Beaufort, S. C.
Cumberland, N. C.
George, Miss.
Camden, Ga,
Chattahoochee, Ga.
Onslow, N. C.
Jackson, Miss.
Charlton, Ga.


High Negative Loadings


1.77
1.73
1.71
1.70
1.58
1.34
1.27
1.25
1.24
1.16
1.16


Choctaw, Ala.
Forsythe, N. C.
Baldwin, Ga.
Fulton, Ga.
Hyde, N. C.
Alleghany, N. C.


-18.77
-1.38
-1.28
-1.21
-1.04
-1.00








Table 34--Continued

High Positive Loadings High negative Loadings

Houston, Ga. 1.16
Mobile, Ala. 1.09
Charleston, S. C. 1.04
Marengo, Miiss. 1.02
Muscogee, Ga. 1.00
Pearl River, Miss. 1.00


Comparison of the Two Time Periods

The purpose of examining two time periods is to

interpret changes through time, in the overall economic

growth of the southeastern United States, to analyze whe-

ther there has been a change in the importance of the dimen-

sions of economic health, and whether new factors have arisen

over the decade 1950-1960. Such a comparative study, however,

is limited to a descriptive analysis as tests of significance

between dimensions over two time periods have not yet been

derived.

The 1950 and 1960 factor analyses revealed essentially

the same patterns of economic health, although in 1960 an

additional dimension emerged. This dimension, however, con-

tributed little to the explanation of the total variation.

In spite of this similarity important changes have occurred

in the rank order of the components, indicating that the

dimensions are not stable through time, and that changes have

occurred in the overall factor pattern. Similar results were

noted by King in his study of Canadian cities and by Hoffman

and Romsa2 in their analysis of urban dimensions in the south-

eastern United States.








Although dimension IV in 1950 (Rural White Counties)

and dimension II in 1960 (Rural Negro Counties) are named

differently, they are considered to be similar as both are

indicators of the color problem. They have similar factor

loadings with the exception that the correlation signs are

reversed. Thus one dimension can be considered a reciprocal

of the other.

The changes in rank order are reflected in dimensions

II through V. Negro poverty has moved from fourth in order

of importance in 1950 to second in 1960 and the per cent of

variation contributed has increased from 7.4 to 13.3 per

cent, an increase of 5.9 per cent. This change in the impor-

tance of the Negro dimension implies that the Negroes have

not shared equitably in the economic progress of the South-

east, during this decade. The decrease of 2.7 per cent in

explained variation of component II in 1950, tenant and small

farm poverty, has reduced it to third in order of importance

in 1960. The decrease in this form of rural poverty is in

accord with the mechanization and the consolidation of farm

units that has occurred in the Southeast. In many instances

this process of consolidation and mechanization has resulted

in a surplus labour supply. Agricultural unemployment is

responsible for the increased concentration of Negro poverty

in the Mississippi Delta. The mechanization of cotton farm-

ing in the Delta has greatly reduced the labour requirements

and has either left the Negro unemployed, or has forced the









excess labour supply to migrate elsewhere, particularly to

the North. This similar process has been at work in south-

ern Georgia, southern Alabama and southwestern Mississippi,

where the pulp and paper companies have become established

and have reverted large areas of crop land to pine forest.

The shift of dimension III in 1950 (wealthy urban-

ized counties) to dimension V in 1960 is an indication that

there has been a narrowing of the range in prosperity

between cities. This is interpreted as demonstrating that

urban centers have shared in economic growth more equitably

than rural areas. This is shown by the increase in rural

poverty in certain sections of the study area as expressed

by components II and IV in 1960. In 1950 these components

accounted for 13 per cent of the total variation while in

1960 they accounted for 20.5 per cent.

The locational factor relating to the national

market has moved in this time period from fifth place in

1950 to fourth place in 1960. On the other hand the regional

and local market locational index has remained stable not

only in rank but it also only registered a slight increase

in total explained variation. This may be the result of an

improved transportation network that has made these markets

more acccssibae. The new dimension that emerged in 1960 is

not only difficult to identify but it appears to have con-

tributed little to the overall analysis.









From these dimensions it is clear that the variance

in economic health throughout the study area in 1950 and

1960 is a function of urbanism, color, location, industrial-

ization, and rural poverty. It is observed that the rate

and form of urban growth is most important, for not only do

the urban areas harbour the majority of the region's poverty,

but also influence to some extent the growth rates of their

hinterlands. It is suggested that the increased rural pov-

erty that is arising in certain sections, due to increasing

technology, will become of greater concern unless an outlet

is found for the surplus supply of labour.



Regression Analysis

The major underlying patterns of economic health

factored from the original 26 variables can be tested for

hypotheses by the use of a regression model. As regression

analysis requires normal and independent data, the testing

of the normality of the factor scores and the independence

of the dimensions is of theoretical importance. In addition

it is of interest to ascertain how well these dimensions

explain the variation in poverty throughout the study area.

To test these and additional hypotheses a dependent variable,

median family income, vias chosen as an index of poverty.

This variable was then regressed on the dimensions in a

stepwise regression.









The Hypotheses

Since the dimensions are theoretically representative

of the original data, it is hypothesized that a high coef-

ficient of determination between the dimensions and median

family income will be obtained. It is hypothesized that the

1960 dimensions will result in a higher coefficient of deter-

mination, as they have higher commonalities than the 1950

components. In addition, hypotheses in regard to the expected

correlations between the dimensions and the dependent variable

can be tested. The expected relationships are shown in Table

35.

Table 35

Expected Correlations Between Median Family
Income and the Derived Dimensions
Median Family Income

Urban Industrial Counties
Tenant and Small Farm Poverty
Wealthy Urbanized Counties +*
Rural Negro Counties
Locational Advantage in Regard to
National Market 1
Locational Advantage in Regard to
Regional and Local Markets +
Areas of Declining Prosperity
Rural White Counties .


The analysis verified the normality of the factor

scores and the independence of the dimensions (Tables 36,

37, 39 and 40). The hypothesis regarding the coefficient of

determination was found to be correct (Tables 38 and 41).

However, the correlations between the dependent and the






80



independent variables were found as expected with one

exception, Urban Industrial dimension had a positive cor-

relation instead of a negative correlation as expected

(Tables 37 and 40).

Table 36

Normality of 1950 Factor Scores


Variable


Mean


Median Family Income 1445.38525
Urban Industrial Counties -0.00002
Tenant and Small Farm Poverty 0.00003
Wealthy Urbanized Counties 0.00022
Rural White Counties -0.00013
Locational Advantage in Regard
to.National Market .-0.00000
Locational Advantage in Regard
to National, Local and
Regional Markets -0.0004


Standard Deviation


572.95850
.99995
i99997
.99997
.99997

.99997


.99995


Table 37

1950 Correlation Matrix


Variable


3 4


1 Median Family
Income 1.000 0.365 -0.124 0.623
2 Urban Industrial
Counties 1.000 -0.000 0.000
3 Tenant and Small
Farm Poverty 1.000 0.000
4 Wealthy Urbanized
Counties 1.000
5 Rural White Counties
6 Locational Advantage in Relation
to National Market
7 Locational Advantage in Relation
to National, Regional and Local Markets


0.334 0.301 0.212

-0.000 0.000 0.000

-0.000 -0.000 -0.000


-0.000
1.000


-0.000 -0.000
0,000 0.000

1.000 0.000


1.000







Table 38

1950 Summary Table


Step Variable
Number Entered


Multiple


0.6232
0.7223
0.7959
0.8510
0.8770
0.8857


2 2
R Increase in R


0.3884
0.52].8
0.6334
0.7243
0.7692
0.7845


0.3884
0.1334
0.1116
0.0909
0.0449
0.0153


Table 39

Normality of 1960 Factor Scores


Variable


Mean


Standard Deviaticn


IMedian Family Income 3
Urban Industrial Counties
Tenant and Small Farm Poverty
Wealthy Urbanized Counties
Rural Negro Poverty
Locational Advantage in Relation
to National Market
Locational Advantage in Relation
to National, Regional and
Local Markets
Areas of Declining Prosperity


072,64746
-0.00003
0.00007
0.00018
0.00013

0.00002


-0.00006
-0.00011


986.50854
.99992
.99998
.99995
.99999


.99995


.99996
.99995


Table 40
1960 Correlation Matrix


Variable


4 5 6 7


1 Median Family
Income 1.000 0.349 -0.363 -0.187 0.046 0,695 0.186 -0.042
2 Urban Industrial
Counties 1.000 -0.000 0.000 0.000 -0.000-0.000-0.000
3 Rural Negro
Counties 1.000 0.000 -0.000 -0.000-0.000 -0.000
4 Tenant and Small
Farm Poverty 1.000 -0.000 0.000-0.00 --0.000
5 Locational Advantage in
Relation to National M:arket 1.000 -0.000 0.000 -0.000
6 WealthyUrbanized Counties 1.000 -0,U00 -0.000
7 Locational Advantage in Relation to
National, Regional and Local iMarkets 1.000 0.000
8 Areas of Declining Prosperity 1.000


__ __








Table 41

1960 Summary Table


Step Variable
Number Entered

1 6
2 3
3 2


Multiple


0.6954
0.7844
0.8584
0.8785
0.8979
0.8991
0.9001


0.483
0.615
0.736
0.771
0.806
0.8o8
0.810


2 Increase in R

i6 0.4836
3 0.1316
8 0.1.215
7 0.0349
1 0.0344
3 0.0022
)1 0.0018


Table 42


Comparison of the Beta, Correlation Coefficients
and the Coefficient of Determination Between the
Two Time Periods


Beta


Urban Industrial


1950

R R2


209.26 .365*


Counties
Tenant and Small *
Farm Poverty -70.92 -.124
Wealthy Urbanized .
Counties 357.09 .623
Rural Negro Counties -- --
Locational Advantage
in Relation to the .
National Market 172.72 .301
Locational Advantage
in Relation to National,
Regional and Local *
Markets 121.44 .212
Areas of Declining


Prosperity
Rural White
Counties


191.42 .334


Beta


. 33


343.94


1960

R R2

.349 .121


.015 -184.26 -.187. .034


.388 686.11l
-- -357.91


.090


45.80


.044 183.09

-41.85*


.695 .
-.363


.483
.131


.046 .002


.186

-.042


.034

.001


.111


Significant at .05 level.

The Beta and Correlation Coefficients (Table 42)

indicate that the variance in median family income in both

periods was a function of location, color, industrialization,


_ I~YI~









urbanization and rural poverty. Dimension VII contributed

very little to the analysis. Although the urban industrial

dimension did not correlate negatively as expected, this

was probably due to the large percentage of high incomes

found in these urban industrial counties which tend to inflate

median family incomes, and thus to a degree hide the large

number of poor families in these counties.

There have, however, been changes in the significance

of these dimensions in explaining the differences in median

family income (Table 41). The influence of the locational

dimensions has dropped rapidly from 13.58 per cent of the

explained variation in 1950 to 3.62 per cent in 1960. Almost

all of this loss is accounted for by the dimension represent-

ing location in regard to the national market, as this com-

ponent v.as not significant in the 1960 regression analysis.

This may reflect the overall economic growth that was achieved

in the South during the two time periods, especially in the

cities. In contrast to the decrease in the importance of

location, urban growth played a larger role in 1960. The

dimension representing urban wealth accounted for an addi-

tional 10 per cent in explained variation. On the other hand

rural poverty had a slightly higher negative correlation in

1960 than in 1950, indicating that in general rural areas are

today at a larger disadvantage. This also applies to the

Negro, who has lost ground during the time period studied.

This appears also to be borne out by the beta coefficients, with

the exception of the beta value, associated with location in









relation to the national market, as the beta coefficients

in 1960 are substantially larger than those in 1950. This

is an indication that the income gap between the poor and

the rich has actually widened in this time period,



Summary

The factor and regression analysis shows that the

urban-rural and the Negro-white dichotomy is increasing.

The analysis indicates that economic growth and opportunity

are being centered around urban complexes to the disadvantage

of rural areas. Furthermore, the economic gap between the

Negro and white segments of the population in this region

has not narrowed. These conclusions are in accord with the

theories regarding poverty and regional growth which were

examined in Chapter I.













REFERENCES


Leslie J. King, "Cross Sectional Analysis on
Canadian Urban Dimensions 1951 and 1961." Canadian Geojraapher.
X, No. 4 (1966), 205-224.

2
WV/ayne Hoffman and Gerald Romsa, "Some Temporal Con-
siderations of Basic Urban Dimensions in the Southeast,"
Paper read before the meeting of the Southeastern Division
of the Association of American Geographers, Greenville, North
Carolina, November 26, 1968.














CHAPTER IV

SUMMARY AND CONCLUSIONS


This chapter summarizes the findings of the factor

and regression analysis for two time periods of five south-

eastern states, Alabama, Mississippi, Georgia, South Carolina

and North Carolina. These states were selected as they

constitute a socio-cconomic region and are representative of

the southeastern United States.

The prime objective of the study was to identify the

major dimensions of economic health of the study area for

two time pe-iods 1950 and 1960, and to ascertain if thes-,

dimensions had c' anged throu ;h ti, e. In the analysis, the

area variation in economic )'e~tth was measured by using the

multivariate procedure of factor analysis, as the 26 variables

selected and used as indices of economic health did not meet

the statistical premises of normality and independence.

Furthermore, factor scores derived from factor analysis were

used in interpreting the areal pattern of the dimensions of

economic health., These same factor scores were also used in

a regression equation to explain the variance in median

family income.










1950 Analysis

The 1950 data were grouped into six dimensions which

represented 70.5 per cent of the total variance of the

original 26 variables. Dimension I was interpreted as being

an indicator of industrialization and urbanization. Counties

ranking high on this component contained the largest urban

centers in the study area. The highest factor scores were

associated with the larger and more rapidly growing centers

such as Atlanta, Birmingham and Charlotte.

Factor II was a low income agricultural component.

Variables characterizing this dimension revealed an agricul-

tural structure which was remnant of the sharecropping era.

Many of the farmers, in the counties that were identified by

this component, are not able to compete successfully with

agricultural enterprises that use large inputs of land,

capital and managerial skill as substitutes for labour and

thus are able to operate on a larger scale. Counties fitting

this pattern were generally located in the old cotton belt.

The third dimension revealed a pattern of urban

counties. These counties contained smaller but wealthier

cities tha l the urban nodes represented by dimension I.

These cities contain a high proportion of residents who are

employed in certain high paying segments of the service sec-

tor, such as education and the military. Other counties

which are represented by this dimension contain wealthy

dormitory suburbs.









Factor IV was a component defining the largely

white population of Appalachia and the predominantly Negro

Mississippi Delta. This dimension although reflecting

poverty in both areas, identified the disparity in level of

education attained and land ownership between white and the

nonwhite segments of the population. The last two dimen-

sions were a reflection of the effect of location, in regard

to local, regional and national markets. No definitive pat-

tern arose, although there is some empirical evidence that

poverty in southwestern Mississippi is due to its disadvan-

tageous location.

On the basis of the derived dimensions, areal pat-

terns of economic health were derived. In general a picture

arose of scattered peaks representing counties with a high

proportion of urban residents and or industrial development

that are surrounded by a sea of poverty. In the northern

portion of the study area are the poor whites of Appalachia.

In. the Mississippi delta poverty is largely comprised of

Negro agricultural workers, many of whom are unemployed.

Poverty in southwestern Mississippi appears to be a reflec-

tion of isolation. Throughout the entire area are found the

poor tenant and small farmer.

Factor scores obtained from the analysis were util-

ized as independent variables in a regression model to test

the significance of the dimensions in explaining the varia-

tion in median family income. The regression an:-Jysis did









support the area pattern of economic health as identified

by the six dimensions. The six factor scores obtained from

the factor analysis explained 78.45 per cent of the varia-

tion in median family income. Median income was found to be

positively correlated with urbanization, industrialization,

and the white portion of the population, The three dimensions

representing these indices accounted for 63.34 per cent of the

explained variation. Location and agricultural poverty con-

tributed smaller amounts to the total coefficient of deter-

mination.

The beta coefficients also substantiated the correla-

tion coefficients. Dimension III had a beta value of 357.09

implying that a one unit increase in this dimension would

result in an increase of $357.09 in median family income.

Since this dimension is largely a reflection of education,

it bears out the economic value of public investment in the

schools of this region. A one unit increase in the urban

industrial dimension would result in an increased median

family income of $209.26. Being white or located close to

urban centers resulted also in higher incomes. Agricultural

employment, however, resulted in lower incomes as did a

decrease in the white dimension.


1960 Analysis

The 1960 data were reduced by factor analysis to

seven dimensions, which explained 77.9 per cent of the varia-

tion in economic health. The increase in explained variation









over the 1950 period was due to the derivation of an addi-

tional component and a more complete factoring of the

original variables. Except for dimension VII (areas of

declining prosperity) the components obtained for this time

period were similar to those derived in 1950, although their

order of importance had changed.

Factor I as in 1950 was associated with urban

industrial counties. It had similar factor loadings and

was identified with counties containing the larger urban

centers. The second component pointed out the economic

discrimination against the Negro. This factor differentiates

between Negro poverty, of which a large proportion is con-

centrated in the Mississippi Delta, and the poverty of white

Appalachia. Although the poverty of Appalachia has been

recognized by various levels of government through different

aid programs, little has been done for the Delta area.

Dimension III represents rural poverty found

throughout the Southeast. Rural poverty is found among

both the white and the Negro operators of small land hold-

ings or tenants.

Dimension IV, a locational index, reflects the

effect of proximity to urban markets. The areal pattern

of factor scores although difficult to interpret suggests

that Mississippi and Alabama suffer by not having closer

access to the national market.




Full Text

PAGE 1

A SPATIAL ANALYSIS OF THE DIMENSIONS OF ECONOMIC HEALTH IN THE SOUTHEASTERN UNITED STATES (1950 AND I960) By GERALD HARRY ROMSA A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UOFF LIBRARIES UNIVERSITY OF FLORIDA 1969

PAGE 2

UNIVERSITY OF FLORIDA 3 1262 08552 2190

PAGE 3

To my parents and my wife Mary

PAGE 4

ACKNOWLEDGMENTS I wish to express my gratitude to my advisor Dr. Clark I. Cross and the other members of my committee, Dr. Stanley D. Brunn, Dr. James R. Anderson and Dr. John E. Reynolds for their guidance, thoughts, constructive criticisms, and patience during the preparation of this dissertation. In addition, I want to acknowledge the University of Florida Computer Center's facilities and services which were utilized for the data analysis in this study. 11 :

PAGE 5

ACKNOWLEDGMENTS TABLE OF CONTENTS Page iii LIST OP TABLES vl LIST OF FIGURES ix CHAPTER I THE CONCEPT OF POVERTY 1 Introduction . * Objective ^ Need for Such a Study J Poverty in the United States 5 Some Approaches to the Study of Poverty . . 10 Review of the Literature 15 The Problem *1 II STUDY AREA AND METHODOLOGY 27 The Setting * ^9 Historical Background • . . ju Recent Trends in the Economy .J? Outline of the Analytical Model 3b The Model j< Variables Used J^ Summary . • • 4* III THE DIMENSIONS OF ECONOMIC HEAL™ IN THE SOUTHEASTERN UNITED STATES (1950-1960) .... 4. 1950 Factor Analysis 46 The 1950 Dimensions *° I960 Analysis °Y The I960 Dimensions . . . . . • °£ Comparison of the Two Time Penoas (j Regression Analysis '° The Hypotheses < J Summary IV

PAGE 6

Page CHAPTER IV SUMMARY AND CONCLUSIONS 86 1950 Analysis 87 I960 Analysis 89 Comparison of Changes in the Two Time Periods Studied 92 Conclusions 93 Future Research Efforts 94 BIBLIOGRAPHY 97 BIOGRAPHICAL SKETCH 101

PAGE 7

LIST OP TABLES Table Page 1 Per Cent of Families Earning Less than S3, 000 a Year by Regions (I960) 8 2 Distribution of Poverty by Residence (I960) . 9 3 Median Income by Residence (i960) 9 4 Variables Used 41 5 Percentage of Total Variance Explained by Each Dimension in 1950 47 6 Percentage of Variance ( Communal ity ) of Each of the 26 Variables Accounted for by All Six Components in 1950 47 7 Dimension I Urban Industrial Counties in 1950 49 8 Counties with Extreme Factor Scores on Dimension I in 1950 50 9 Dimension II Tenant and Small Farm Poverty in 1950 52 10 Counties with Extreme Factor Scores on Dimension II in 1950 52 11 Dimension III Wealthy Urbanized Counties in 1950 . 55 12 Counties with Extreme Factor Scores on Dimension III in 1950 55 13 Dimension IV Rural White Counties in 1950 . 57 vi

PAGE 8

Table Page 14 Counties with Extreme Factor Scores on Dimension IV in 1950 57 15 Dimension V Locational Advantage in Relation to National Market (New York) in 1950 59 16 Counties with Extreme Factor Scores on Dimension V in 1950 59 17 Dimension VI Locational Advantage in Relation to Local, Regional and National Markets in 1950 60 18 Counties with Extreme Scores on Dimension VI in 1950 60 19 Percentage of Total Variance Explained byEach Dimension in I960 61 20 Percentage of Variance (Comrnunality ) of Each of the 26 Variables Accounted for by All Seven Components in I960 61 21 Dimension I Urban Industrial Counties in I960 63 22 Counties with Extreme Factor Scores on Dimension I in I960 64 23 Dimension II Rural Negro Counties 66 24 Counties with Extreme Factor Scores on Dimension II in I960 67 25 Dimension III Tenant and Small Farm Poverty in I960 67 26 Counties with Extreme Factor Scores on Dimension III in I960 69 27 Dimension IV Locational Advantage in Relation to National Market (New York) in I960 70 vii

PAGE 9

Table Page 28 Counties with Extreme Factor Scores on Dimension IV in I960 71 29 Dimension V Wealthy Urbanized Counties in I960 72 30 Counties with Extreme Factor Scores on Dimension V in I960 72 31 Dimension VI Locational Advantage in Relation to National, Regional and Local Markets in I960 73 32 Counties with Extreme Factor Scores on Dimension VI in I960 74 33 Dimension VII Areas of Declining Prosperity in I960 74 34 Counties with Extreme Factor Scores on Dimension VII in I960 74 35 Expected Correlations Between Median Family Income and the Derived Dimensions. . 79 36 Normality of 1950 Factor Scores 80 37 1950 Correlation Matrix 80 38 1950 Summary Table 81 39 Normality of I960 Factor Scores 8l 40 I960 Correlation Matrix 81 41 I960 Summary Table 82 42 Comparison of the Beta, Correlation Coefficients and the Coefficient of Determination Between the Two Time Periods 82 vi 11

PAGE 10

LIST OP FIGURES Figure Page 1 Study Area; the Southeastern Core 28 2 Urban Industrial Counties, 1950 51 3 Tenant and Small Farm Poverty, 1950 54 4 Urban Industrial Counties, I960 65 5 Tenant and Small Farm Poverty, I960 68 IX

PAGE 11

CHAPTER I THE CONCEPT OF POVERTY Introduction The increasing attention given by geographers to the problems of poverty, economic health and underdeveloped areas is a healthy sign that they are progressing with the temporal needs of today's society. This concern with areal differentiation in economic growth and levels of economic wellbeing is not a new geography. It is, instead, a more specialized topical outlook on the net result of economic activity as viewed in a spatial framework. This approach attempts to relate regional inequalities in levels of living with man's economic endeavours as they are tied to his cultural, technological and resource environment. The trend to look for factors other than resource endowment is perhaps the major criterion which separates present geographical research from previous regional studies.. Ullman has pointed out that natural resources and cultural differences are not sufficient to explain the dispar2 ity in economic growth. Differences in social conditions and advances in technology often provide additional information in explaining the areal variation in prosperity. The inclusion of additional variables to explain this 1

PAGE 12

differential in economic browth has led to the use of multivariate statistical techniques as analytical tools. The usefulness of these techniques lies in their applicability in the analysis of large numbers of indices; in their isolation of significant variables and in the testing of hypotheses. These are important contributions when dealing with interrelated data. The realization by social scientists that the distribution of wealth and poverty is uneven has focused their attention on the need for more regional studies. National and state data often tend to hide small pockets of high deviation in levels of income. This has led to economic regionalization based on counties, and, where possible, smaller units. Spatial interaction is another geographic concept which is increasingly being realized as playing an important role in explaining the difference in levels of progress. This concept was introduced by Ullman and refined by Hagerstrand at Lund University. Hagerstrand used simulation models in an attempt to measure the degree of resistance that cultural and physical barriers imposed upon the acceptance of new ideas among contiguous areas in Sweden. Hartshorne points out that the comparison of economic growth among underdeveloped regions should be relaA tive rather than concrete. A direct comparison between two culturally different regions may not show the disparity in

PAGE 13

economic achievement accurately unless the indices utilized are standardized to represent with less bias the contrasting cultural and political values. This also is true of temporal studies. Although this problem at present has not been resolved, studies within a similar cultural and political unit based upon certain multivariate techniques are feasible. Objective The purpose of this study is to identify the salient socio-economic dimensions or characteristics of five southeastern states (Mississippi, Alabama, Georgia, South Carolina and North Carolina) for two time periods, 1950 and I960. These characteristics will be examined to ascertain if they reflect any areal patterns or changes through time. If areal variations do exist, they will be analyzed in regard to their stability, location, effects of urbanization and the importance of push and pull factors. In addition the extracted dimensions will be tested individually and as a group for their validity in explaining the high deviation in median income throughout the study area. Need for Such a Study The concern of some Americans for a more equitable distribution of this nation's wealth has shown itself through civil legislation and action programs by state,

PAGE 14

local and federal agencies. Programs such as the Area Redevelopment Act and the Appalachia Act are examples of the federal government's increased participation. To implement these programs rapidly and to the best advantage requires a thorough knowledge of the problems. Thus research is essential if the complex facets of poverty are to be understood. An example of such a research program in the South is The Southern Regional Research Project S-44, "Factors in the Adjustment of Families and Individuals in Low Income Areas of the South." This is a cooperative study done by rural sociologists in nine southern states. These studies were done in selected counties and dealt primarily with the attitudes of the rural poor. There is a need, however, for a comprehensive study of the Southeast which deals with the facets associated with poverty on a larger scale. The Southeast has often been classified as a retarded area. Recent studies have suggested that this blanket statement can not be applied to the entire region, as it has made some great economic strides in the past few decades. At present, there is a lack of geographic research on this region employing multivariate statistical techniques, which has emphasized the degree of variation in economic growth through space and time and classified economic regions accordingly. In the remainder of this chapter an overview of the distribution of poverty in the United States and some

PAGE 15

approaches to its study are discussed. The pertinent literature is reviev/ed as a basis for formulating hypotheses. The analytical model used in this study is developed in Chapter II and a brief description of the study area is given. Chapter III deals with the analysis of the data, v/hile in Chapter IV the conclusions reached are examined in view of the hypotheses presented in Chapter I. Poverty in the United States Amid the highest standard of living ever achieved by the United States there exists a forgotten minority of poverty stricken people. The problems of the poor are often masked by the middle class way of life in America. There is no room in this society for the economically less fortunate. The middle class viewpoint is well expressed in The Affluent 5 Society by John Galbraith. Galbraith's major theme was how, with increasing automation, the good life could be continued for all, with less work and more leisure. Poverty was perceived only in two forms: case poverty due to physical and mental handicaps, and insular poverty such as is found in West Virginia. More recent articles have shown that the problems of poverty are not isolated but nationally widespread. There are, however, large concentrations of the poor in certain geographic areas such as the Soiitb and in the ghettos of urban centers. Depending upon the criteria used, there are

PAGE 16

estimated to be from 36 to 50 million economically deprived 7 persons within this wealthy nation. This staggering figure is made even more realistic if one considers the economic, social and psychological costs of maintaining these individuals in their present way of life. Not only does the Gross National Product suffer from the loss of such a large number of human resources, but there are also heavy social costs. These include welfare payments, costs of mental and physical health, crime and juvenile delinquency. The national cost of welfare proo grams has been estimated at 13 to 15 billion dollars. Poverty was first recognized as a social problem in g the United States approximately 100 years ago. It existed before but had been regarded as an individual rather than a social problem. Prior to this period, individuals attempted to attain economic freedom by leaving the urban ghettos or the worn out soils of the East and moving westward. This movement could only be sustained in a highly agrarian society. The slaves on the southern plantations however did not have this escape valve. Increasing industrialization and urban growth after the Civil War brought with it many social evils: poor tenement housing, inadequate schools, low wages, bad working conditions and no job security. Henry George in his Pro gress and__Poverty referred to the United States of 1869 as

PAGE 17

a land where "amid the greatest accumulation of wealth, men die of starvation and puny infants suckle dry breasts." The plight of the working classes, many of whom were immigrants, eventually led to some social reforms. Communityminded citizens led the battle for public housing regulations, public health, and abolition of child labour. Labour unions played a major role in obtaining higher wages and better working conditions. Major Federal government action on poverty was initiated by President Franklin D. Roosevelt in the 1930's. In the State of the Union message in 1933 the former President referred to the "one third of the nation that is illclothed, ill-housed and ill-nourished.""' It was during this period that the familiar antipoverty tools: social security, old age assistance, aid to dependent children, and unemployment compensation were conceived. Furthermore the Tennessee Valley Authority Act was a recognition that there were regional inequalities in economic growth, which resulted in pockets of distressed areas that were resistant to change and progress. The not major Federal attack on poverty commenced in the 1960's. The emphasis was on legislation. The Area Redevelopment Act of 1962, the Economic Opportunity Act and the Appalachian Regional Development Act of 1964 are a few examples. The passage of the acts emphasized rehabilitation and not monetary relief. They also recognized the areal

PAGE 18

8 distributions of poverty and the need to focus upon the problems of regional or subregional nature which can not be solved on the national level. The new progress in legislation was initiated by a small educated white minority and a group of dedicated Negro leaders. Civil Rights marches, sit-ins, and riots have forced the nation's attention to the problems of her minority poor when other less 12 peaceful efforts failed. The geographic distribution of poverty can readily be seen if the per cent of families earning less than S3, 000 a year is examined by the major divisions of the country (Table 1). Table 1 Per cent of Families Earning Less Than $3,000 a Year by Regions (i960) Region Total White Nonwhite Northeast 9.8 8.4 26.6 North Central 11.5 10.3 29.7 South 24.6 17.9 58.3 West 11.7 11.0 20.7 Source: Louis A. Perman, Joyce L. Kornbluh and Alan Haber (etis„), Povert y in America (Ann Arbor: University of Michigan Press, 1965*77 P« 72. This table shows the enormous variation in income across America. It further points out the plight of the nonwhites. Although the urban centers harbour the majority of the poor, the rural areas have a higher percentage of poor families (Tables 2 and 3).

PAGE 19

Table 2 Distribution of Poverty by Residence (i960) Residence $ of population $ of poor (families earning less than $3,000 per year) Urban 71 54 Rural nonfarm 22 30 Farm 7 16 Source: Louis A. Ferman, Joyce L. Kornbluh and Alan Haber (eds.), Poverty in Ame rica (Ann Arbor: University of Michigan Press, 1965), p. 104. Table 3 Median Income by Residence (i960) Residence Total Po pulatio n White Nonwh ite Urban $6,580 $6,678 $4,469 Rural Nonfarm $5,486 $5,549 $2,645 Rural Farm $3,779 $3,863 $1,32 3 Source: Louis A. Ferman, Joyce L. Kornbluh and Alan Haber (eds.), Pov erty in America (Ann Arbor: University of Michigan Press, 1965"I7~p. 107. In comparing the income between rural and urban residents, one should consider the difference in cost of living. Nevertheless the differences in income are greater than the variation in the cost of living. From these figures a picture arises that reveals small concentrated areas of poverty within cities and a more dispersed form of rural poor. Within the rural sections of the country, there are extreme variations in the concentrations of the poor. Examples of areas with a high rate of unemployment and a concentrated number of low income groups are: Appalachia,

PAGE 20

10 the cutover regions of the Great Lakes, northern New England, the northern Rockies and the Indian reservations. Other regions containing large but less concentrated numbers of poor families are parts of the northern Great Plains, scattered areas in the West and Midwest and extensive sections of the Piedmont and the Atlantic Coastal Plain. This pattern of dispersion not only shows the national scale of the problem, but also the variations from the industrial North to the rural Wast and the South. The distribution of natural resources can not account for all of the variation in poverty. Part of this is due to demographic factors. Minority nonwhite groups are concentrated in many of the distressed economic sections. Some depressed regions have a high percentage of unskilled labour, while many of the predominantly white rural counties have high concentrations of older people. The latter is the result of the out-migration of the young to more attractive economic areas. Continuous out-migration over a period of time, of the highly motivated segment of the population, will tend to economically depress an area. This was illustrated and substantiated by Spinelli in his study of the net out-migration in southeastern Ohio, 1950-1960. Some Approac hes to the Stu d y of Poverty One of the most difficult problems in studying poverty is its definition. What constitutes poverty? Many

PAGE 21

11 popular definitions adopted are based upon incomes. Statistics on family and individual income are readily available and acciirate. The present S3, 000 criterion used for families and the 01,500 for single individuals was defined and used in 1964 by the President's Council of Economic Advisors. The Council estimated that in 1962 there were between 33 and 35 million Americans living at or below this . 16 income level. Economists such as Leon Keyserling believe that this criteria are too low. Keyserling raises the income criteria for families living at the poverty level to 34,000 and for single persons to $2,000 (i960 dollars). In his analysis he includes a second grouping called the deprived category. This group includes all families with incomes between $4,000 and $5,999 and individuals with earnings between $2,000 and $3,999. This latter group can subsist en their incomes, but can not afford many of the luxuries that our present society has to offer. Keyserling estimates that 73 17 million Americans fall within these two categories. One of the weaknesses of these estimates, based upon the above definitions, is that they do not account for the differences in the cost of living throughout the country. They also do not contain flexibility to take into account inflationary periods of the economy which increase the cost of living for individuals on fixed incomes. The social security administration has made some attempt to account

PAGE 22

12 for "the variation in cost of living, but the invariant $3,000 base is still largely used by them. A more flexible standard which distinguishes the differences in need between size of families, rural and urban residence, was published in 1964 by the Department of Health, Education and Welfare. This minimum budget ranges from Si, 100 for a farm family of two, aged 65, to over $5,000 for a nonfarm family of seven or more individuals. This more flexible budget estimated l ft the poor at 34.6 million. More realistic income criteria are those suggested by the Bureau of Labor Statistics. Their index was developed for an urban family of four living in 20 of the major cities. This index varied from 19 $5,370 in Houston to $6,567 in Chicago. The above estimates show the disagreements among experts on what is a proper income level for the definition of poverty. These estimates, furthermore, have not taken into account transfer payments in the form of financial assistance from welfare agencies or relatives. They also have not made any allowance for taxation and therefore, because of these omissions, do not represent true disposable income. Indeed most students of poverty feel that the use of a single index is not sufficient. The trend is to use several indicators such as color, age, income and education. Some social scientists have approached the problem from the sociological and psychological viewpoint. One of

PAGE 23

13 the first to use this definition was Oscar Lewis, who states: Poverty in modern nations is not only a state of economic deprivation, of disorganization or of the absence of something. It is also positive in the sense that it has a structure, a rationale, and defence mechanisms without which the poor could hardly carry on. In short it is a way of life, stable and persistent passed down from generation to generation along family lines. The culture of poverty has its own modalities and distinctive social and psychological consequences for its members. ° Economically retarded areas have within them distinct cultural features. The poor having little money, virtually no savings and no economic security, have evolved unique consumer habits, family structures and a way of life. This subculture, if it may be so called, is composed of the chronically unemployed or those having irregular employment. Those who are employed work in the non-unionized low-paying sectors of the economy. Their low salaries can not provide for the costs of higher education and health services. Often their meager earnings are spent on entertainment luxuries, instead of the basics of food, shelter and clothing, in an attempt to escape their environment for short periods of time. The end result is often a low resistance to disease, ill health and a shorter life span. The everyday problems of this subculture, due to the constant stress of maintaining life without hope, has created in turn psychological manifestations. This has lead to a lack of motivation and imagination among the destitute. Protective shells against the outside world are erected. Within

PAGE 24

14 these shells provincialism and deep-rooted prejudices are maintained. The outside world is viewed with hostility. This hostility is expressed in their distrust of various government agencies and programs. The sociological approach in defining poverty, although accurately describing the subculture of the poor, has certain drawbacks. The major restriction is the lack of sociological data for large sections of the country. To gather the data requires personal interviews, which not only can be costly, but also time consuming. In addition, this definition emphasizes to a great extent the end results rather than the causes. A more recent concept, economic health, is gaining in popularity as it attempts to analyze which indices are responsible for a lower level of living and how these indices may vary through space and time. The term economic health is not universally used, but is represented in a large number of multivariate studies that use numerous social, economic and demographic indices. These indicators are then reduced to a smaller number by elimination or union with other variables into new ones. This new set of indices, which accounts for the largest possible variance within a workable matrix, is then used to explain the difference in economic growth. This matrix can be applied to inter-, or intra-regional studies and to a certain extent

PAGE 25

15 be used to explain the changes through tine. Economic health thus tries, by the use of a large number of variables, to explain why and hov/ different regions of a country, relative to each other, differ in their overall economic prosperity. The method of analysis is not new, but it was not until rapid calculating computers became available that large regions could be examined on the basis of a multitude of variables, and at the same time remove personal bias in regard to the importance of the variables used. This form of analysis is being used successfully by various social scientists engaged in regional studies. The concept of economic health will be used in this study to explain the differences in prosperity, if any, within the five southeastern states in the study area. Twenty-six variables which are indicators of prosperity will be utilized to ascertain what role they play within the study area, and if their importance changes through space and time. Review of the Literature The dependence of the Southeast on agriculture has been documented by geographers (Hart, Vance). This historical reliance upon a few major crops such as cotton and tobacco had its drawbacks. These crops were export orientated and consequently fluctuated greatly in price.

PAGE 26

16 Soil mismanagement often led to the abandonment of farms or a loss in fertility and yields. The plantation form of agriculture was not conducive to the modernization of agricultural practices, as cheap labour after slavery made it economically unfeasible to adopt some of the new techniques until relative prices of the inputs changed. Rural values predominated and eventually evolved into a way of life. Political power came under the control of a small, land based minority to the detriment of education and industrialization. This theme, that the South* s historical development acted as a barrier to progress, pervades the literature on the South. Nichols in a paper "Southern Tradition and Regional Economic Progress" suggests that five social, political, psychological and philosophical factors were respon22 sible for the lag in southern economic growth. These were: the dominance of agrarian values, rigidity of a social class structure which prevented the establishment of a strong middle class, the control of the political parties by a few which led to an undemocratic political structure. Furthermore a lack of social responsibility hindered the development of a first class public school system. This, tied with the conformity of thought and behavior which was expected of a good southerner, left little or no sympathy for dissenters, many of whom migrated North. The loss of this liberal element hindered change and helped maintain the status quo.

PAGE 27

17 In spite of these handicaps the South in the last few decades has made tremendous strides forward. MacDonald "On South's Recent Economic Development " points out that personal income in this region, has increased since 1940, at a faster rate than in the nonsouth. Part of this increase is attributed by MacDonald to a shift in the American economy from coal and steel to plastics and petroleum, and the increased exploitation of other southern resources especially forestry and agriculture. The other main factor responsible for the rapid increase in personal income was a slow population growth rate, due to the heavy out-migration of low income groups. This recent progress of the South was further verified by lassiter. In his study "Education for Males by Region, Race and Age " Lassiter pointed out that at present the investment in education is providing better financial returns for white southern males than white northern males. This rapid rate of grov/th however has not been consistent throughout this region. Moon and McCann, who looked at the levels of adjustment in 30 rural counties in the states of Kentucky, Tennessee, North Carolina, Texas, Louisiana, Mississippi and Alabama, found statistically sig25 nif icant variation. J Similar conclusions were arrived at by Tarver and Beale in "Population Trends of Southern Nonmetropolitan Towns." They concluded that the significant factors explaining the differences in population growth

PAGE 28

18 among 801 southern towns with populations between 2,500 and 9,999 were: the size of the town and its regional location, Fuguitt on the other hand found that county seats in rural nonmetropolitan areas grew at the expense of neighboring ' 27 centers. The importance of urbanization as a significant factor in explaining regional growth in the United States has received considerable attention. Schlesinger has argued that since the Civil War, cities dominated the American 28 growth process. Friedman and Alonso depict a spatial economy in which the processes of economic development are leading to greater specialization and increased polarization of growth in a few urban industrial complexes, each of which 29 is functionally integrated with a surrounding economic area. These centers not only grow so rapidly as to create problems of an entirely new order, but they also act as suction pumps, pulling in the more dynamic elements from the more static regions. The remainder of the country is thus relegated to a second class, peripheral position. It is placed in a quasicolonial relationship to the center, experiencing net outflows of people, capital, and resources, most of which is redound to the advantage of the center where economic growth will tend to be rapid, sustained and cumulative. As a result, income differences between the center and the periphery tend to widen. 30 This viewpoint is also expressed by Schultz, who argues that economic development occurs primarily within an industrial urban matrix where the factor and product markets work best. Agricultural activities located most

PAGE 29

19 favourably in regard to these centers can utilize the factor and product markets more efficiently than those located on the periphery of these urban systems. This was partially substantiated by Bryant who used multiple regression to measure the causes of inter-county variation in •30 farmers' income in the United States. He showed that variation in farm income is largely determined by age, land and capital inputs, education and color. Location in respect to regional urban areas was only significant in the eastern United States, implying that farmers west of the Mississippi were more dependent upon national and international markets. 0. D. Duncan, et al ., in Metropolis and Region , emphasize the domination of hinterland activities by metropolitan centers. Both agricultural and non-metropolitan manufacturing activities were shown to be a function of general accessibility to the urban system of the nation and distance to the local metropolis. Berry concluded in his study of rural poverty in Ontario that areas lacking the dynamic conditions of industrial urban development experience high rates of outmigration and commitantly realize differences in per capita income, even though labour migrations generally appear sufficient to overcome differences attributable to differences in original resource endowments . The factors linking rural poverty to industrial urban growth polarized in urban centers and varying in amplitude according to the rank of >. these centers in an urban hierarchy, are thus clear. This urban influence has also been clearly demonstrated by Thompson, et al .. Berry, Bell and Stevenson,

PAGE 30

20 and Hodge. All of these authors used factor analysis on a large matrix of economic indices in order to derive a smaller number of dimensions or factor loadings which would explain more easily the variations in income levels. In all cases an urban dimension was derived from the analysis that was correlated with the rapidly growing areas. The study by Thompson, et al ., on the economic health of New York State, is of particular interest here. It not only demonstrated that the large metropolitan cities were the nodes of economic growth, but also that the predominantly agricultural counties near these centers fared better than smaller cities, which, had populations of 50,000 or ler.s. Berry arrived at the same conclusions in his analysis of rural poverty in Ontario. Location of the rural townships, in relation to urban areas, played an important role in determining the level of economic achievement as measured by 47 variables. In only a few exceptions, most of which were due to the presence of mineral resources, did a few townships deviate from the pattern of hinterlands dominated by large urban industrial clusters. Using similar procedures Hodge analyzed 473 trade centers in Saskatchewan. He found that trade center viability was largely a function of rank on the urbanism scale; the lower the rank the greater the probability of decline. The apparent importance of the urban or industrial dimension is clearly seen in these studies. That these

PAGE 31

21 complexes do grow at the expense of their hinterlands is readily verified, if one looks at the variables which are correlated with the urban industrial dimensions. These are generally high income groups, youthfulness of the population, a high level of education and a high level of capital expenditures. All are indices of a healthy economy. The converse is true; the dimensions representing poverty have weaker associations with the above indices. Are these indices stable through time? The decline of certain regions relative to others would seem to indicate 39 the contrary. King found that they were not stable. In his examination of Canadian cities he found that the percentage of variance explained by the dimensions arising from his analysis varied through time, although the factor loadings associated with the dimensions remained fairly stable. The Problem The problem involved in this study relates to the variations in the economic environment among 454 counties in Alabama, Georgia, Mississippi, North and South Carolina. Specifically does variation exist? What dimensions account for this variation? Do they vary over time? What percentage of poverty do they explain? Are some counties progressing at the expense of others? The literature reviewed provides a basis from which to examine several useful spatial hypotheses of economic health in the selected southeastern states. The hypotheses to be tested in the study are:

PAGE 32

22 (1) There is variation in economic health among the selected counties and these can be explained in terms of a few dimensions arising out of factor analysis. (2) These dimensions are not stable but change through time as do the conditions required for economic growth, (3) Within the study area there will be nodes of rapidly growing areas which thrive at the expense of their hinterlands. (4) These nodal areas are associated with urbanization and it is postulated that the degree of growth within these sectors, relative to each other, is a function of population size and location, location in relation to local, regional and national metropolitan centers. This study will be successful if it can demonstrate the underlying characteristics of areal variation in economic health in the five selected southeastern states for two points in time, and ascertain if these characteristics have changed between the two time periods (1950 and I960).

PAGE 33

REFERENCES 1 Richard Hartshorne, "Geography and Economic Growth," in Essays on Geography and Economic Development , Norton Ginsburg ( ed. ) (Chicago: The University of Chicago Press, I960), pp. 24-25. 2 Edward L. Ullman, "Geographic Theory and Underdeveloped Areas," in Essay s on Geography and Economic Devel opment , Norton Ginsburg (ed.) (Chicago: The University of Chicago Press, I960), pp. 26-27. 3 Ibid ., p. 29. 4 Hartshorne, op. cit ., pp. 18-2 3. 5 John Galbraith, The Affluent Society (Boston: Houghton-Mifflin Co., 1958). 6 Ibid ., pp. 32 3-333. 7 Michael Harrington, The Other America, Poverty in the United States (Baltimore: Penguin Books, 1966), pp. 176-179. 8 Lar A. Levitan, "Programs in Aid of the Poor," Povert y and Human Resources Abstracts , Vol. I (1966), pp. 11^25~. ^Herman P. Miller, Poverty American St yle (Belmont: Wadsworth Publishing Co. , Inc., 19^6), pp. 2-3. 10 Louis A. Ferman, Joyce L. Kornbluh and Alan Haber (eds.), P overty in America (Ann Arbor: University of Michigan Press, 1965), P« 1. i:L Ibid. 12 Miller, op. cit ., p. 5. 23

PAGE 34

24 Gordon E. Reckord, "The Geography of Poverty in the United States," in Problems and Trends in American Geography , Saul B. Cohen (ed. ) (New York: Basic Books, Inc., 1967), p. 105. 14 Ibid. 15 Joseph Spinelli, "A Study of Net Out-LIigration in Southeastern Ohio, 1950-1960" (Unpublished Master's Thesis, Department of Geography, Ohio State University, 1966). Miller, op. cit . , p. 2. 17 Ibid. 1 8 Harrington, loc. cit . 19 Ibid. 20 Elizabeth Herzog, "Some Assumptions About the Poor," Social Science Review , XXXVII (December, 1963), 390. 21 ~ John Eraser Hart, The Southeastern United State s (Princeton: D. Van Nostrand Co., Inc., 1967) ; ind Rupert B. Vance, Human Geography of the South (Chapel Hill: University of North Carolina Press, 1935 ) . 22 William H. Nichols, "Southern Tradition and Regional Economic Progress," Southern Economic Journal , XXVI (January, I960), 187-198. 23 Stephen L. MacDonald, "On South' s Recent Economic Development," South ern Economic Journal , XXVIII (July, 1961), 30-40. O A Roy L. Lassiter, Jr., "Education for Males by Region, Race and Age," Southern Economic Journal , XXXII (July, 1965), 15-22. 25 Seung Gyu Moon and Glenn C. McCann, Subregional Variability of Adjustment Factors of Ru ral Fam i lies in the South (Raleigh: North Carolina Agricultural Experiment Station, Southern Cooperative Series, Bulletin No. Ill, January, 1966).

PAGE 35

25 James D. Tarver and Calvin L. Beale, "Population Trends of Southern Nonmetropolitan Towns," R ural Sociology, XXXIII (March, 1968), 19-29. 27 'Glenn V. Fuguitt, "County Seat Status as a Factor in Small Town Growth and Decline," Social Forces , XLIII (December, 1965), 245-251. Arthur M. Schle singer, A History of American Life , Vol. XI, The Rise of the City 1878-1 393 (Mew York: The Macmillan Co., 1949). 29 John Friedman and William Alonso (eds.), Regional Dev elopment and Planning (Cambridge: Massachusetts Institute of Technology Press, 19b4), pp. 1-11. 30 Ibid., p. 3. T. W. Schultz, Economic Organization of Ag riculture (New York: Lie Graw~Hill Co . , 19 5 3 ) . 32 Keith W. Bryant, "Inter-County Variation in Farmers 1 Income," Jour nal of Farm Economics , XLVIII (August, 1966), 557-577. Otis Dudly Duncan, et al . , Metropolis and Region (Baltimore: The John Hopkins Press, i960) . Brian J. Berry, Strategies, Models and Economic Theories of Development in Rural Regions . Agricultural Economic Report No. 127 I Washington: Economic Research Service, U.S. Department of Agriculture, 1967). 35 John H. Thompson, et al . , "Toward a Geography of Economic Health: The Case of New York State," in Regi onal De velopment end Planning , John Friedman and William Alonso (eds.) (Cambridge: Massachusetts Institute of Technology Press, 1964), p. 187. Brian J. Berry, "Identification of Declining Regions: An Empirical Study of the Dimensions of Rural Poverty," in Thoman, R. S. (ed.), Areas of Economic Stress in Canada (Kingston: Queens University, 19t>5 ) . 37 W. H. Bell and D. W. Stevenson, "An Index of Economic Health for Ontario Counties and Districts," Ontario Economic Review, II, No. 5 (1964), 1-7.

PAGE 36

26 Gerald Hodge, "Do Villages Grow? Some Perspectives and Predictions," Rural Sociology , XXX, No. 2 (June, 1966), 183-196. 39 Leslie J. King, "Cross Sectional Analysis of Canadian Urban Dimensions 1951 and 1961," Canadian Geographer , X, No. 4 (1966), 205-224.

PAGE 37

CHAPTER II STUDY AREA AND METHODOLOGY The study area is composed of five states, Mississippi, Alabama, Georgia, North and South Carolina. This region covers an area of 241,968 square miles, and in 1968 had a population of 16,326,746. These states are similar in their socio-economic-political characteristics (Hart, 2 3 Vance, Odum ). All five withdrew with the Confederacy and after Reconstruction the states voted solidly Democratic until 1964, with the exception of North Carolina which went Republican once. Perhaps due to the apparent homogenity of the South, few authors have attempted to subclassify it. What regionalization of the South has taken place is largely based upon physical divisions such as the Delta, Piedmont, Appalachia and the Coastal Plain (White and Foscue, Griffin, Young, and Chatham ) . Bogue and Beale however, using 64 variables, have broken the South down into four distinct regions. One of which, the Southeast Coastal Plain and Piedmont, coincides very closely with the area under study (Figure 1). Five distinct characteristics separate this region from the others. It is the least urbanized, has the highest percentage of Negro population, the highest 27

PAGE 38

28

PAGE 39

29 percentage of the labour force in agriculture, and the 7 industrial base is weak and highly dependent upon textiles. These four characteristics are the cause of the fifth, low median income, as the lowest median family income of any region in the United States is found here. Furthermore Bogue and Bcale found enough variation within this distinct region to classify it into 19 subregions, implying that there may be differential growth patterns. The choice of this study area was dependent to a great extent on the previous work done by Bogue and Beale. Although the two areas are not exactly the same, they may be considered as one in their characteristics. The reason for this discrepancy is that state boundaries were maintained so as to reduce the effect of state governments in the analysis. The Settin g Three distinct physiographic features fall within the study area, the Coastal Plain, Piedmont and Appalachia. The boundary between the Piedmont and the Coastal Plain is demarcated by a Fall Line and a change in soils. The Fall Line is characterized by a zone of waterfalls and rapids in the streams which imposed a limit on navigation. Cities such as Raleigh, Columbia, Macon, Columbus and Montgomery are found along this break. In the Carolinas and Georgia the Fall Line is marked by a long, low narrow belt of sand

PAGE 40

30 hills, the remnants of ancient beaches. The Piedmont soils are residual, the product of weathering of the underlying rock over eons of time. The Coastal Plain soils on the other hand were formed of sedimentary materials when the plain was submerged. Both physiographic divisions have a variety of local topography ranging from level to hilly, with the older Piedmont having the greater extremes of relief. A small portion of the study area falls within Appalachia. This section extends from central Alabama to the Blue Ridge of Northwestern Georgia and North Carolina. In the Blue Ridge the mountains rise 3,000 to 6,000 feet above sea level. In the Appalachian Valley of Alabama the relief is much less, varying from 500 to 1,000 feet. The economy was and to a large extent still is centered on agriculture. Three major crops: tobacco, cotton and peanuts provide the majority of the farmers' cash receipts. Industrial growth has been based upon service functions, textiles, food processing, furniture and lumbering. In many respects the economy is similar to that of underdeveloped countries, in that it is dependent on a few staple agricultural products and has a weak industrial base. Historical Backgr ound The retardation of this area, as in the rest of the Souths has largely been due to the historical development of

PAGE 41

31 the region. This historical retardation has been well 8 9 documented by historians and economists (Eaton, Mad do x ). It was the development of an agrarian culture that led to the exclusion of the rise of a strong middle class, and industry. These could have become the impetus for economic growth, when agriculture was diminishing in importance. Once behind, the South was in a poor position to compete with the industrial North, and it became subservient to a centralized economy based upon market potentials. The Southeast was settled by two streams of settlement. The first group occupied the tidewater between Charleston and Savannah. Plantations were established and at first indentured servants, later slaves, were brought in as a source of labour. Rice became the dominant crop along the coast, while indigo was grown on the plantations further inland. As the tidewater was being settled a different group of immigrants were coming into the western Carolinas and Georgia. These were the Scotch-Irish and Germans who had come south from Pennsylvania. Many of these settled permanently in the backcountry and became successful farmers, raising all their requirements. The excess livestock, hogs, corn were sold or converted to other more valuable byproducts such as liquor or hides. The heavy migration into the western area led to a more rapid rate of growth than in the tidev/ater. Eventually this may have led to the political control of the southern colonies by the small farmers, had

PAGE 42

32 not the importance of cotton began to assert itself. Cotton planters began to move into the Piedmont and it became the major crop. The search for more and better cotton soils led to the opening up and settlement of Alabama and Mississippi. The relatively rapid rise in the dominance of cotton resulted from two major events, the industrial revolution in Europe, requiring cotton for its textile mills, and the invention of the cotton gin by Whitney in 1793. The emergence of King Cotton as the dominant crop, controlled by a small group of planters, utilizing slave labor, pre-empted the use of land for other crops. This small aristocratic group gained and maintained political control throughout the area. Furthermore this group managed to impose their values upon the entire southern society. The use of slave labor was a barrier to the establishment of more productive agricultural practices v/hich would utilize a higher share of capital as a factor of production. The social stigma associated with trade and manufacturing hindered the development of a middle class and a more diversified economy. Perhaps the greatest damage done was the concept that education was not for the masses, but for the select few. The education provided the select few, although of high calibre, was heavily oriented toward the liberal arts and was lacking in the field of applied technology.

PAGE 43

33 Other important sources of income were lumber and naval stores, which initially were of prime importance in North Carolina but rapidly spread throughout the South. Tobacco became the second major crop when, in 1850, it was discovered that bright tobacco did well on the poor sandy soils of northwestern North Carolina. The third major crop was the growing of peanuts along the Georgia and Alabama border. Thus in the period preceding the Civil War the southern economy was oriented around the export of a few major staples as sources of revenue with which to pay for the importation of manufactured goods from abroad and the North. This raw material orientation meant that the economy was affected by any severe fluctuations in prices or tariffs. The tariff of abominations established by the North in 1828 was a blow to southern commerce. 10 The Civil War brought ruin upon the South. The abolition of slavery removed a source of cheap labour from the plantations and left many of these large land holdings idle. A system of sharecropping evolved to fill the labour shortage. The system was not satisfactory for it did not encourage any long term agricultural practices by either the landowner or the sharecropper. In addition, development of either industry or a more efficient system of agriculture was strongly hampered by the shortage of investment capital. Inflation during the War had left the

PAGE 44

34 Confederate states with a worthless currency. Deflation after the war left the South with little buying power. Some of the capital problems were attributable to national monetary policies which hurt the South. The National Bank Reforms in 1865 levied a severe tax on state banks, which predominantly were located in the South. Furthermore in these bank reforms, the South did not receive its fair allotment of Federal currency. The establishment of certain industries such as iron and steel was also hampered by the monopolistic action of northern firms who established "Pittsburgh plus basing point" and by this means removed the locational advantage of Birmingham. The industrial growth that did take place occurred mainly in the field of food processing, textiles and cigarette manufacturing. The use of bright tobacco as a major ingredient in cigarettes gave rise to cities like Durham and Winston Salem. The textile industry moved here from the North during the later part of the nineteenth century. Many of the mills were established in towns along the Fall Line where there was a source of power. Despite these attempts at industrialization, the South did not keep pace with the North. As most poor or stagnating areas it suffered from selective out-migration, lack of capital and the establishment of new firms. All of these factors encouraged the location of industry in the heart of the more lucrative northern market.

PAGE 45

35 Thus the South entered the twentieth century in a disadvantaged position. It was an area short of skilled labour, investment capital, and entrepreneurs, technologically and productively lagging behind the other sections of the country. R ecent Trends in the Economy Despite these handicaps, great strides have been made in the last 30 years throughout the study area and the entire South. This has come about due to the mechanization of agriculture, and better utilization of other natural resources. The development of better educational facilities, bi-racial labour force, more internal savings and capital investments has encouraged the location of industry, espe12 cially those that benefit by decentralization. The trend has been to a diversification of both the agricultural and industrial segments of the economy. Examples are the growth of the dairy industry in Mississippi, the establishment of beef cattle on the Black Prairie of Alabama and the Piedmont of Georgia, the location of the broiler industry since the war in northern Alabama and Georgia. Abandoned cropland, formerly in cotton, has been reverting or restocked to pine woodland. Forest management has been initiated and encouraged by the rapid expansion of the pulp and paper industry. Better conservation practices, increased use of fertilizers and mechanization of farms

PAGE 46

36 have resulted in higher yields. Prices of important cash crops such as cotton, tobacco and peanuts have been stabilized by a system of acreage allotments. Industrial growth has been spurred by the establishment of large important industries. Some major examples are: the atomic energy plant near Aiken, South Carolina; the Redstone guided missile project and other space facilities at Huntsville, Alabama, the large Lockheed aircraft plant in Marietta, Georgia. Less dramatic has been the location of electronic firms in the Golden Triangle of North Carolina. The location by northern companies of branch firms in southern cities has aided in the establishment and diversification of industrial growth. Atlanta owes a great deal of its growth to the establishment of these regional branches. It has become an assembly center for Ford and General Motors. Many insurance and airline firms have their regional offices here. In addition the overall economic growth in the South has been aided by the establishment of large military bases and the rapid increase in tourism, both of which have been attracted by the mild climate. Outline of the Analytical Model Briefly the steps undertaken in the study are as follows: (1) transfer raw data into an M x N matrix (M variables and N observations), (2) perform principal components analysis of the M x N intercorrelation matrix of

PAGE 47

37 the M variables and rotation of the resulting eigenvectors to a normal varimax position to satisfy the criterion of a simple structure, (3) compute the factor scores of N observations on the R rotated factors, (4) regress a dependent variable D on the factor scores. The Model The mathematical model used to determine the underlying dimensions of economic health is known as a principal axis factor analysis solution. 13 The choice of the model is based upon several considerations. The study of the covariance of economic, social and demographic characteristics of regions is a problem of a multivariate nature. Data of this nature are often not -normally distributed and are highly interrelated. Regression models using a large number of variables lacking the assumptions of normality and independence will give a coefficient of determination which is biased upwards. In addition, the coefficients and standard errors of the coefficients may be affected. Factor analysis, however, can be used on data lacking these a priori assumptions. This technique can simultaneously manage a large number of variables, compensate for random error, and disentangle complex interrelations into their major and distinct regularities. " ' The essence of this technique is to reduce a large number of variables into a smaller group of distinct orthogonol factors, components,

PAGE 48

38 or dimensions with a minimum loss of information. On the basis of these factors the unique variations of a domain can be discerned. A further step is the obtaining of factor scores which is the location in space of each observation on the particular dimensions. These factor scores can be used as independent variables regressed on any desired 15 dependent variable. This method is suggested by Scott. Scott contends that when the assumptions of independence and normality are not known, this regression technique is more accurate than the classical least squares technique. The essential features of factor analysis are as 17 follows: from a rectangular X matrix of raw data it is possible to compute a R matrix of simple correlation coefficients. X =

PAGE 49

39 between the variables and the dimensions. One can solve for as many characteristic roots as there are variables; however, one generally only solves for the most important roots, that is for the ones that account for the majority of the variance. The equation for solving L is (R » LI) r= where I is the inverse of R. After finding the eigenvalue L one can solve for the eigenvector V which is the column vector of factor loadings, by the following equation (R = LI) V = 0. Repetition of this procedure yields the factor matrix A. A = a11 a12 ... ami a21 a22 ... am2 r is the number of characteristic roots solved for. ami amr Dividing the characteristic roots found by the total communality (communality is the sum of independent common variances of the components on each variable) yields the variance accounted for by each dimension. It is often of interest to find the factor scores, that is the ranking of each observation on each of the derived components. The matrix of factor scores is given by the following equation F:XAL'. In addition the factors are rotated so as to comply with Thurston's concept of simple structure, such that each of the original variables -i o relates highly with one and only one of the new dimensions. This assures that the first factor arising from the analysis

PAGE 50

40 accounts for the maximum variance and each successive factor accounts for decreased proportions of the explained variance. The factor scores have further application in a regression model to explain the per cent of variation of a dependent variable which is accounted for by the orthogonal dimensions. This procedure produces an index which has the largest possible variance of any linear combination. A lower coefficient of determination is obtained in comparison with the least squares method. Factor regression accounts for errors in the variable high intercorrelations and gives regression coefficients which are more accurately 19 representative of economic theory. Variables Used The variables used for the two time periods were chosen to represent social, economic, demographic and locational characteristics of the study area. In all, 26 indices were used. They were obtained from the censuses of population, agriculture, and the city and county data book for the respective time periods. Three locational indices were calculated for the study. The following list is provided to show the nature of the variables.

PAGE 51

41 Table 4 Variables Used 1 Total population 2 Per cent of the population urban 3 Per cent of the population nonwhite 4 Per cent of the population 65 years old or older 5 Per cent of the population 25 years old or over who completed less than five years of school 6 Per cent of the population 25 years old or over who completed high school 7 Per cent of housing units sound 8 Median value of owner-occupied dwelling units 9 Median gross rent 10 Value added by manufacturing 11 Total capital expenditures by manufacturing firms 12 Total retail sales 13 Total wholesale sales 14 Per cent of the population under five years of age 15 Per cent of the labour force unemployed 16 Per cent of the farms operated by tenants 17 Average value of land and buildings per farm 18 Total value of farm products produced (sold) 19 Number of commercial farms earning less than $2,500 annually 20 Number of farmers whose off-farm income is greater than farm income 21 Per cent of families earning between $0 3999 annually 22 Per cent of families earning between $1,000 $1,900 annually 23 Per cent of families earning betv/een $2,000 $2,999 annually 24 Distance in miles from the center of a county to the nearest city with a population of at least 50,000 25 Distance in miles from the center of a county to the nearest city with a population of at least 500,000 26 Distance in miles from the center of a county to Now York City Through these variables an attempt was made to strike a balance betv/een the rural and urban segments of the study area. This choice also took into account that the majority of these indices have been found to be significant by other researchers, in explaining the variation in economic health.

PAGE 52

42 The major problem with the data is that the areal units (counties) vary immensely in size. Large counties having extensive variation within them may not accurately reflect this in the data. In spite of this inherent difficulty, county data are at present the best available for such an analysis. Summary This chapter is divided into two sections. The first describes the study area and the historical basis for its poverty. This region was selected for analysis as it is representative of the southeastern United States. Furthermore, it appears to have had a substantial number of economic changes within it in the last few decades, and thus would be an ideal testing area for the analysis of spatial and temporal patterns of economic health. Factor analysis was chosen as an appropriate model because it can handle nonnormal and highly intercorrelated data. In addition it opens up other avenues of research beyond the location and identification of the underlying dimensions.

PAGE 53

REFERENCES John Praser Hart, The Southe astern Uni ted States (Princeton: D. Van Nostrand Co., Inc., 1967). 2 Rupert B. Vance, Human Geog raph y of the South (Chapel Hill: The University of North Carolina Press, 19 35). Howard VV. Odura, Southern Regions of the United Stat es (Chapel Hill: The University of North Carolina~Press, 1936")". A ^G. Langdon White and Edwin J. Poscue, Region al Geography of An glo -A merica, 2nd edition (Englewood Cliffs: Prentice-Hall, 19557. 5 Paul P. Griffin, Robert N. Young and Ronald 1. Chatham, Anglo-Am erica, A Regional Geogra phy of the U nited S t a tes and Canada (San Francisco: Pearon Publishers Inc., 19'GT) . Donald J. Bogue and Calvin L. Beale, E conomic Areas of the United States (Hew York: The Free Press of Glencoe, Inc., 1961), p. an. 7 Ibid., pp. 269-270. o Clement Eaton, A History of the Old Sout h (Hew York: The Macmillan Co,, 1965). 9 James G. Maddox, et al . , The Advancing South: Manpower Prospect s and Problems (New York: The Twentieth" Century Fund, 1967). William E. Laird and James R. Rinehart, "Exogeneous Check on Southern Economic Development," Sout h At] antic .Quarterly, Vol. LXV, No. 4 (1966), pp. 491-5037" Ibid . 12 B. U. Ratchford, "Economic Development in the South," South A tlantic Quarte rly, Vol. LXIV, No. 4 (1965), pp. 496505. 43 '

PAGE 54

44 The type of factor analysis used in this study is the version known as principal component analysis. R. J. Rummcl, "Understanding Factor Analysis," Conflict Resolution , Vol. XI, No. 4 (1967), p. 444. 1 c -Vohn T. Scott, Jr., "Factor Analysis and Regression," Eco nometrica , Vol. 34, No. 3 (July, 1967), pp. 552562. For a more detailed explanation of the model see: H. H. Harmon, Modern Factor Analysis (Chicago: University of Chicago Press, 1961); M. G. Kendall, A_Course in _Multiv ariate Analysi s (London: Charles Griffin, 1957), pp. 10T6; R. J. Rummel, "Understanding Factor Analysis, " Conflict Resolution , Vol. XI, No. A (1967), pp. 444-480; Mary Megee, "£jn Economic Growth and the Factor Analysis Method," Southern Economic Journal, XXXI (July, 1964), 215-228. Qazi Ahmad, Indian Cities: Characte ristics an d Correlates (Chicago: University of Chicago Press, 195*5)7 p. 25. Scott, o p. cit ., p. 552.

PAGE 55

CHAPTER III THE DIMENSIONS OP ECONOMIC HEALTH IN THE SOUTHEASTERN UNITED STATES (1950-1960) The analysis as reported in this chapter as well as in the subsequent chapter, was undertaken at the University of Florida's computing center. The original data, for the two time periods, were transformed into symmetric 26 by 26 correlation matrices. These matrices were then subjected to a principal axis solution which yielded six eigenvectors for the 1950 data and seven eigenvectors for the I960 data, based upon the criteria ofx>|. The components were then rotated to a normal varimax position. The underlying dimensions obtained in this manner were used to identify and to help interpret the areal variation in economic health. The basic dimensions of economic health are identified by the analysis of the derived factor loadings and factor scores. Factor loadings are the correlations of the original variables with the newly derived dimensions. If each of the original variables were plotted in n-dimensional space (n representing the number of derived factors), the coordinates of the variables would represent the correlations or loadings on each principal axis (dimension) by the individual variables. These factor loadings range from 1.0, indicating a high positive correlation with a given dimension 45

PAGE 56

46 (which must he interpreted), to a -1.0, indicating a high negative correlation. Variables with high positive loadings on a factor are expected to be most characteristic of the quality represented in the factor, while high negative loadings on the same factor should indicate a contrasting quality. Factor scores on the other hand measure how closely each observation (county) ranks on each of the individual dimensions. For example, a high positive factor score for a county on a given component can be interpreted to mean that the county is highly representative of that component. With this in mind, it becomes possible to identify, and interpret the areal variation of economic health. In order to facilitate interpretations, of the dimensions only high loadings on each of the components are shown, as are only highly negative and positive factor scores in Tables 7-18, 21-34. The factor scores will also be employed later in this chapter in a stepwise regression as a means of testing how well the derived dimensions explain the variation in poverty. 1950 Factor Analysis Table 5 summarizes the proportion of the total intercounty variance in 1950 accounted for by each of the six derived factors. The six rotated factors together explain 70.5 per cent of the total variance of the 26 variables. Thus these six components are considered representative of the variables that they replaced.

PAGE 57

47 Table 5 Percentage of Total Variance Explained by Each Dimension in 1950 Per cent of total variance Dimension

PAGE 58

78.

PAGE 59

49 functions such as retail and wholesale sales. High loading! are also associated with total population, value added by manufacturing, capital expenditures, poverty and youthful ness of the population. Counties which rank high on this dimension are those which contain large rapier y growing regional centers such as Atlanta, Birmingham, Mobile, and Charlotte (Tabic 8 and Figure 2). These centers, due to their prosperity and rapid rate of economic growth, attract a large number of immigrants from the hinterlan 1, many of whom are unskilled and consequently are employed in low paying industries or seasonal work. Table 7 Dime :s : on I Urban Industrial Counties in 1950 Prima ;y V; ri'.-ibles Fac tor Loadings Total population '92 Per cent of population urban .44 Value added by manufacturing .88 Total Capital expenditures .73 Retail sales »^5 Vyholesale sales «o2 Number of persons under five years of age .83 Per cent of families earning between SO-$999 -7 6 Per cent of families earning between $1,000-31,999 »90 Per cent of families earning between 82, 000-32,999 « 6 5

PAGE 60

50 Table 8 Counties with Extreme Factor Scores on Dimension I in 1950 High Posit ive Jefferson, Ala. Pulton, Ga. Mobile, Ala. Mecklenburg, N. C. Forsythe, N. C. Guilford, N. C. Chatham, Ga. Greenville, S. C. Churlestown, S. C. Spartanburg, S e C. Gaston, N. C. Fayette, Ga. Bibb, Ga. Richland, S. C. Montgomery, Ala. Buncombe, N. C. Hinds, Miss. Wake, N. C. Durham, N. C. Richmond, Ga. 12.32 11.84 3.94 3.66 3.51 3.10 2.78 2.52 2.33 2.29 1.97 1.59 1.59 1.50 1.48 1.48 1.40 1.38 1.35 1.30 Hi gh Negative Ga, Chattahooche, Fannin, Ga. Houston, Ga. Carteret, N. C. Dare, N. C. Onslow, N. C. Lafayette, Miss. Oho wan, N. C. Stone, Miss. Orange, N. C. -2.18 -1.34 -1.03 .82 .81 .78 .76 .71 .71 .70 The second dimension which explains 13.3 per cent of the total variance is a rural poverty indicator. This component is related mainly to types of farming and sources of farm income. High positive loadings are found on total value of farm products produced and number of commercial farms earning less than $2,500 annually, the tenancy rate, and number of farmers whose off-farm income is greater than farm income (Table 9). This dimension isolates the poor tenant and the small individual farmer who, because of lack of finances, education, or the small scale of his holding, has not been able to make a financial success of his

PAGE 61

51 *T\ + CO + A C o *, '•£ (1) 1> s O c p „_ o

PAGE 62

52 enterpris. Factor scores fail to reveal any easily defined grouping of counties. Instead they suggest that this type of poverty is quite widespread, especially in areas outside the dominance of any large metropolitan center (Table 10). There does appear, however, to be a concentration of poor counties in the old cotton belt of the Southeast, suggesting that these rural counties have not fully recovered from the negative effects of the sharecropping system of agriculture (Figure 3). Table 9 Dimension II Tenant and Small Farm Poverty in 1950 Prim ary Var i abl e s Per cent of farmers that are tenants Total value of farm products produced (sold) Number of commerce 1 farms earning less than $2,500 annual ; Number of farmers \ iose off-farm income is greater than their farm income Per cent of families earning between SO-S999 Factor Loadings .49 .83 .93 .39 .55 Table 10 Counties with Extreme Factor Scores on Dimension II in 1950 H igh Positiv e Bolivar, Miss. Sunflower, Miss, Robeson, N. C. Leflore, Miss. Washington, Miss, Coahoma, Miss. Cullman, Ala. De Kalb, Ala. Johnston, N. C. High Ne gative 5.38

PAGE 63

53 Table 10 — Continued High, Positive Sampson, N. C. 2.76 Horry, S. C. 2.74 Orangeburg, S. C. 2.40 Tallahatchie, Miss. 2.36 Columbus, N. C. 2.36 Florence, S. C. 2.34 Williamsburg, S. C. 2.22 Marshall, Ala. 2.15 Pitt, N. C. 2.15 Quitman, Miss. 2.09 Cleveland, N. C. 2.05 Tunica, Miss. 2.02 H igh Negative Long, Ga. -1.41 Bryan , Ga . -1 . 39 Camden, N. C. -1.37 Currituck, N. C. -1.37 Dougherty, Ga. -1.33 Clink, G; . -1.32 Turner, Ga. -1.32 Peach, Ga. -1.31 Dare, N. C. -1.28 Hancock, Miss. -1.28 Jones, Ga. -1.26 Wayne, Ga. -1.26 The third most important factor (Wealthy Urbanized Comities) identifies a pattern of urban wealth. It is associated with variables, such as per cent of the population urban, median value of owner-occupied homes, and per cent of the population that completed high school (Tabic 11). The economic influence of Atlanta is clearly seen in De Kalb, Clayton and Cobb which rank high on this dimension. These counties are to an extent dormitory suburbs of Atlanta and as such contain a high proportion of residents with above average incomes. Many of the other counties listed in Table 12 o\,e their prosperity to the location of state institutions, an example is Orangeburg, South Carolina. Other counties that fal!l into this category are Hinds, Mississippi; Montgomery, Alabama and Richland, South Carolina. The influence of state spe idii g is i elf-evident in these counties as the respective state capitals are located there. Federal payrolls are another important source of prosperity, for examp] e

PAGE 64

54 + c O „ * V O io Q -S.2

PAGE 65

55 Houston, Chattahooche and Muscogee counties in Georgia, the sites of large military bases. Prosperity in these urbanized counties can be traced to concentrations of highly educated people working in professions that pay above average salaries. Table 11 Dimension III Wealthy Urbanized Counties in 1950 Primar y Vari abl es Factor Loadin gs Per cent of the population urban .73 Per cent of the population that completed less than five years of school -.43 Median value of owner-occupied homes .70 Average value of farms (land and buildings) .55 Per cent of the population that completed high school .82 Table 12 Counties with Extreme Factor Scores on Dimension III in 1950 on, Miss High Positive Chattahooche, Ga. De Kalb, Ga. Harri Glynn, Ga. Houston, Ga. New Hanover, N. C. Dougherty, Ga. Muscogee, Ga. Forrest, Miss. Clarke, Ga. Adams, Miss. Hinds, Miss. Montgomery, Ala. Richland, S. C. Jackson, Miss. Orange, N. C. Pasquotank, N. C. Chatham , Ga . 5.54 4.39 3.72 3.71 3.33 3.08 3.06 2.98 2.73 2.71 2.70 2.62 2.44 2.35 2,31 2.24 2.16 2.11 Hi gh Ne g ative Fayette, Ga. Jefferson, Ala. Dawson, Ga. Jasper, S. C. Heard, Ga. Union, Ga. Stokes, N. C. Forsythe, Ga. Yancey, N. C. Berkeley, S. C. Calhoun, S. C. Banks, G . Echols, Ga. Lowndes, Ala. Fulton, Ga. Gilmer, Ga. Greene, Ala. Clebourne, Ala. -2.25 -2.16 -1.85 -1.53 -1.42 -1.40 -1.36 -1.36 -1 . 31 -1.31 -1.31 -1.30 -1 . 30 -1.29 -1.28 -1.27 -1.26 -1.26

PAGE 66

56 Dimension IV (Rural White Counties) shows the contrast in levels of living between primarily white and black counties. Variables that correlate highly negative with this component are: per cent of the population nonwhite, per cent completed less than five years of school and per cent of farmers that are tenants (Table 13). A positive correlation is found between this dimension and number of farmers whose income from off-farm employment is greater than farm income. High positive factor scores are assoc:' ited with the Appalachian counties of Alabama, Georgia, North and South Carolina, with the heaviest concentration in western North Carolina (Table 14). The predominantly Negro counties rank low on this factor (Table 14). These Negro counties are located in a belt which runs from eastern North Carolina to the Mississippi Delta. The largest concentrations are found in the Delta, southwestern Georgia and the tobacco area of eastern North Carolina. The majority of the Negroes in this belt are the descendants of the former slaves who worked on the cotton and tobacco plantations. Thus this area of colored poverty has its roots in the plantation system of land tenure, whose remnants still remain. Although both groups of counties are poor, the Negro comities are more so, indicating the traditional disparity between the black and the white segments of our society.

PAGE 67

57 Table 13 Dimension IV Rural White Counties in 1950 Primary Var iab les Factor Loadings Per cent of population nonwhite Per cent of farmers that are tenants Per cent of population that completed less than five years of school Number of farmers whose income from off-farm employment is greater than farm income -.85 -.72 -.78 .60 Table 14 Counties with Extreme Factor Scores on Dimension IV in 1950 High Posi tiv e Wilkes, N. C. Cullman, Ala. Walker, Ala. Ashe, N. C. Fannin, Ga. Madison, I . C. Caldwell, N. C. Randolph, N. C. Greenvil] , S. C. De Kalb, 11a, Watauga, N. J. Buncombe, N. C. Catawaba N. C. Jackson, If. C, Cherokee, N. C. Spartanburg, S. C Davidson, N. C. Burk e , N . C . Mitchell, N. C. Hig h Negative 2.30

PAGE 68

18 city may not always be beneficial. For a small urban center generally is at a disadvantage competing with a large regional center for a regional market, especially if the regional center contains a large proportion of the regional market. A large city in this instance may cause services and industries to gravitate to it. This may be true of the southwestern Mississippi counties that rank highly negative on Dimension V (Table 16). These counties are depressed economically not only because of their poor agricultural base but also due to the lack of thriving urban centers. In this case the presence of a large urban center is not beneficial, rather it appears that New Orleans is growing at the expense of this area. This in turn suggests that New Orleans is losing its national significance to other more rapidly growing centers such as Atlanta. This is in contrast to the more prosperous counties of North Carolina, whose produce is related more to national markets. Thus this dimension appears to reflect the importance of national markets to economic growth. The last dimension is another locational index as it has negative loadings with all the distance variables (Table 17). The positive loading with median value of owner-occupied dwellings is an indicator of wealth. This factor reflects a locational advantage in regard to regional and local markets. A glance at the factor scores will verify this locational advantage. For example, counties with high positive

PAGE 69

59 scores are found to be centered near large metropolitan centers (Table 18). The economic effect of Atlanta is clearly seen on the Georgia comities. The counties with high negative factor scores tend to be more isolated and economically depressed. Table 15 Dimension V Locational Advantage in Relation to National Market (New York) in 1950 Primary Variables. Factor Loadings Per cent of population 65 years of age or older Median gross rent Avci-ge value of farm (land and building ) Total°value of farm products produced Number of farmers whose incon : from offfarm employment Is greater than farm income Distance in miles fron the center of a county to a city of ;t lear '; 500,000 Distance in miles from the center of a county to Ne.. York -.31 .31 .48 .27 -.22 .66 -.81 Table 16 Counties with Extreme Factor Scores on Dimension V in 1950 High Positiv e Horry, S. C. Yadkin, N. C. Glynn, Ga. Alleghany, N. C. Robeson, N. C. Slurry, N. C. Columbus, N. C. Stokes, N. C. Davie, N. C. Harnett, N.C. Forsythe, N. C. Pitt, N. C. Florence, S. C. Cabarrus, N. C. Dillon, S. C. Camden, Ga. Nash, N. C. Hi gh Ne.gatj.ye_ 2.22

PAGE 70

60 Table 17 Dimension VI Locational Advantage in Relation to Local Regional and National Markets in 1950 Primary Vari ables Median value of owner-occupied dwelling units Median gross rent Distance in mi] es from the cei ber of a county to a city of at least 50,000 Distance in miles from the center of a county to a city of at least 500,000 Distance in miles from the center of a county to New York Factor Loadings .69 .22 -.47 -.43 -.26 Table 18 Counties with Extreme Factor in 1950 High Pos itive Fannin, Ga. Fayette, Ga. De Kalb, Ga. Edgecombe, N. C Clayton, Ga. Wilson, N. C. Turner, Ga. Hertford, N. C. Evans, Ga. Tunica, Miss. Nash, N. C. Wake, N. C. Greene, N. C. De Soto, Miss. Henry, Ga. 12.76 2.90 2.15 2.02 1.95 1.86 1.86 1.81 1.76 1.75 1.69 1.66 1.63 1.55 1.52 Scores on Dimension VI H igh Negative Lowndes, Ga. -2.51 Georgetown, S. C. -1.84 Issaquena, Miss. -1.79 Wayne, Ga. -1.76 Choctow, Ala. -1.76 Washington, Ala. -1.64 Clarke, Miss. -1.62 New Hanover, N. C. -1.55 Columbus, N. C. -1.52 Marion, S. C. -1.46 Louderdale, Miss. -1.44 Mcintosh, Ga. -1.44 Choctow, Miss. -1.43 Jasper, Miss. -1.42 Geneva, Ala. -1.41 I960 Analysis In I960 seven intensions < q r[ d from the factor analysis solution. TY :se seven components explain 77.9 per cent of the total variance (Table 19). The higher percentage

PAGE 71

61 of explanation over the 1950 period is due to the additional dimension and a higher percentage accounted for by the other dimensions. A check of the coramunalities reveals that the higher percentage of explained variation is due to a better factoring of the variables, including old age and unemployment, which v/ere poorly accounted for in 1950 (Table 20). Table 19 Percentage of Tot a]. Variance Explained by Each Dimension in I960 Dimension I II III IV V VI VII Tot; 1 77.9 Table 20 Percentage of Variance (Cornmunality) of Each of the 26 Variables Accounted for by All Seven Components in I960 Variables Co rnmunality in percentag e 1 97.1 2 65.9 3 77.6 4 92.4 5 85.4 6 78.0 7 91.3 8 89.5 9 83.5 10 87.1 11 58.9 12 81.5 13 67.2 14 67.2 Eigenvalue

PAGE 72

62 Table 20 — Continued Variables Communality in p ercentage 15 73.1 16 74.8 17 68.6 18 65.1 19 83.9 20 75.5 21 91.6 22 95.4 23 94.2 24 57.2 25 54.2 26 69.2 The_ I960 Dimensions The largest dimension, as in 1950, is labelled urban industrial counties. It contributes 33.8 per cent of the total variance. High positive correlations are found between this dimension and population, value added, capital expenditures, retail sales, wholesale sales, and poverty (Table 21). The concentration of the poor in the industrial area is not unexpected. As employment diminishes in the rural areas due to increased mechanization and larger land holding the only promise of jobs lies in the growing industrial and service sectors of the economy. Consequently there is a large migration from the country to the cities. At times the flow is too great to be absorbed at once into the economy, resulting in temporary unemployment. On the other hand, due to their lack of skills or color, many of those who are absorbed by the labour market are engaged in low paying occupations, especially in some branches of the service sector of the economy. Consequently the faster a

PAGE 73

63 city grows and the larger it becomes the larger the concentration of urban poverty. For as a city achieves a higher rank in the hierarchy of urban systems, it assumes more service functions, many of which are low paying. In addition the expansion of urban functions often makes the city more attractive to future immigrants; this then compounds the problem of absorbing these migrants into the existing labour force rapidly. Furthermore, if "Die supply of labour greatly exceeds the demand, wages may remain low unless the workers are unionized or covered by wage laws. This notion is suggested by the county factor scores in Table 22. It is observed that the cities ranking highest on the urban industrial poverty dimension are also the largest in population (Figure 4). Table 21 Dimension I Urban Industrial Counties in I960 Primary Variables Factor loading s Total population .9 3 Per cent of the population urban .43 Value added by manufacturing .84 Total capital expenditures .73 Retail sales .80 Wholesale sales .80 Per cent of families earning between SO $999 .80 Per cent of families earning between Si, 000 SI, 999 .90 Per cent of families earning between $2,000 $2,999 .91

PAGE 74

64 Table 22 Counties with Extreme Factor Scores on Dimension I in I960 High Positive Pultan, Ga. Jefferson, Ala. Mecklenburg, N. C, Mobile, Ala. Guilford, N. C. Forsythe, N. C. Greenville, S. C. Chatham, Ga. Charleston, S. C. Spartanburg, S. C, Gaston, N. C. Richland, S. C . Hinds, Miss. De Kalb, Ga. Etowah, Ala. Jackson, Miss. Buncombe, N. C. Wake, N. C. High Negative 11.99

PAGE 75

65
PAGE 76

66 Negro counties. These counties are located in an arc that runs from eastern North Carolina through the old cotton belt to the Mississippi Delta where the heaviest concentration is found. Vrhite counties with low factor scores are generally found in the Appalachian region of the study area. The distribution of the counties is similar to that found in 1950. The third dimension is also an indicator of rural poverty but it is not related to color. This component reflects the residual poverty left over from the sharecropping era. This group of rural poor is made up of tenant and small commercial farmers who have not been able to cope with the changing trends in agricultural production (Table 25). In many instances their deprivation is attributed to the lack of education, capital and land resources. Although this form of poverty is widespread (Table 26), Figure 5 indicates that it is most prevalent in the areas where sharecropping prevailed. Table 2 3 Dimension II Rural Negro Counties in I960 Primary Variables F actor Loadings Per cent of the population nonwhite .82 Per cent of population that completed less than five years of school .72 Per cent of population under five years of age .77 Per cent of farmers who are tenants .74 Average value of farm (land and buildings) .47

PAGE 77

67 Table 23 — Continued Primary Variab les Factor Loadings Total value of farm products produced Number of farmers whose off-farm income is greater than farm income Per cent of dwelling units sound .35 -.48 -.45 Table 24 Counties with Extreme Factor Scores on Dimension II in I960 High Positive

PAGE 78

68

PAGE 79

69 Tabic 26 Counties with Extreme Factor Score;' on Dimension III in I960 High Pos itive Robeson, N. C. 3.37 Cul 1 man , Al a . 3.32 De Kalb, Ala. 3.23 Hinds, Miss. 2.84 Panola, Miss. 2.62 Dallas, Ala. 2.58 Williamsburg, S. C. 2.56 Sunflower, Miss. 2.51 Orangeburg, S. C. 2.50 Horry, S. C. 2.49 Columbus, N. C. 2.48 Sampson, N. C. 2.47 Marshall, Miss. 2.25 Madison, Miss. 2.19 Marshall, Ala. 2.14 Anderson, S. C. 2,12 Florence, S. C. 2.08 High Nega tive Fulton, Ga. Camden, Ga. Mcintosh, Ga. Clinch, Ga. De Kalb, Ga. Jackson, Miss. Charlton, Ga. Peach, Ga. Mecklenburg, N Muscogee, Ga. Bryan, Ga. Chatham, Ga. Currituck, N. ( Liberty, Ga, Dougherty, Ga. Lee, Ga. Jones, Ga. -2.71 -2.53 -2.40 -2.05 -1.86 -1.80 -1.76 -1.74 -1.63 -1.59 -1.58 -1.55 -1.46 -1.42 -1.39 -1 . 39 -1.37 The fourth dimension (Locational Advantage in Relation to National Market) has high correlations with distance to New York and distance to a city of at least 500,000 persons (Table 27). Counties with extreme negative scores on this dimension are generally concentrated in southwestern Mississippi (Table 28). Other low factor scores are found in central Alabama and northern Mississippi, These factor scores, although they do to a certain extent reflect poverty, reflect largely the proximity of these counties to large regional centers such as Memphis, Birmingham and New Orleans. It is of interest to note that no large regional grouping appears around Atlanta. The absence of such a grouping around Atlanta may be attributed to the closer proximity of New York. On the other side of the scale are found the

PAGE 80

70 tobacco -producing and the furniture manufacturing counties of North and South Carolina. Although these counties are not in close proximity to regional markets they are close to the national market. This favourable location is important in regard to their specialization in the growing of tobacco and in the manufacturing of cigarettes and furniture Furthermore these commodities have been less susceptible to mechanization than other enterprises such as cotton growing. This is in contrast to southwe stern Mississippi which suffers from being unable to compete effectively in the agriculture market. As a result there has been a large reversion to forestland, especially with the coming in of large pulp and paper companies. The surplus of labour resulting from the dwindling agricultural base has resulted in a high unemployment rate and out-migration. Table 27 Dimension IV Locational Advantage in Relation to National Market (New York) in I960 Fr imary Va r iabl e s Factor Lo adings Distance in miles to a city of at least 500,000 persons .72 Distance in miles to New York ~»78 Per cent of population 65 years of age or over -.19 Median gross rent .29 Total value of farm products produced .21 Per cent of population nonwhite -.14 Per cent of population unemployed -.17 Distance in miles from center of county to nearest city of at least 50 s 000 .22

PAGE 81

71 Tabic 28 Counties with Extreme Factor Scores on Dimension IV in I960 Hiph Positive Lo adings Gilmer, Ga. 4.61 Horry, S. C. 2.52 Florence, S. C. 2.43 Columbus, S. C 2.20 Dillon, S. C. 2.19 Robeson, N. C. 2.13 Darlington, S. C. 2.10 Alleghany, N. C. 2.10 Marlsboro, N. C. 2.02 Marion, S. C. 1.87 Yadkin, K. C. 1.85 Surry, N. C. 1.83 Scotland, N. C. 1.81 Alexander, N. C. 1.79 Carteret, N. C. 1.71 New Hanover, N. C. 1.71 Wilkes, N. C. 1.71 High_ Nega tive Loadings Pearl River, Miss. -2.42 Harrison, Miss. -2.08 Mobile, Ala. -2.03 Pike, Miss. -2.04 Hancock, Miss. -2.03 Copiah, Miss. -1.9 3 Tuscaloosa, Ala. -1.92 Lincoln, Miss. -1.87 Hinds, Miss. -1.79 Amite, Miss. -1.78 Mar ion , M i s s . -1.76 Wilkinson, Miss. -1.73 Lawrence, Miss. -1.7-1 Jefferson Davis, Miss. -1.70 Walker, Ala. -1.65 Adams, Miss. -1.62 George, Miss. -1.61 Dimension V labelled wealthy urbanized counties, reflects urban prosperity. It has high factor loadings on per cent of the population urban, per cent of population completed high school, per cent of population with sound housing, median value of owner-occupied homes and median gross rent (Table 29). Counties which rank high on this component are prosperous small cities such as Macon, Albany, and Columbus in Georgia, Huntsville, Alabama and Wake Forest in North Carolina (Table 30). Other counties that have high positive scores are Cobb, Clayton and De Kalb in Georgia. These counties contain some of the more prosperous suburbs of Atlanta. Negative factor scores are associated with counties both rural and urban, that have a high percentage of low income groups (Table 30).

PAGE 82

72 Table 29 Dimension V Wealthy Urbanized Counties in I960 Primary Variables Factor Loadin gs Per cent of population urban .67 Per cent of population completed less than five years of school -.52 Per cent of population completed high school .80 Per cent of population with sound housing .80 Median value of owner-occupied homes .70 Median gross rent .75 Retail sales .39 Average value of farm (Land and buildings) .48 Table 30 Counties with Extreme Factor Scores on Dimension V in I960 High Negative Loadings Webster, Ga. Jefferson, Ala. Fulton, Ga. Glascock, Ga. Quitman, Ga. Hancock, Ga. Talbot, Ga. Issaquena, Miss. Stewart, Ga. Bank: Ga. Wheeler, Ga. -2.45 -2.43 -2.43 -1.82 -1.74 -1.67 -1.58 -1.58 -1.54 -1.50 -1.50 Sigh Positive loadings De Kalb s Ga. 4.12 Houston, Ga. 3.69 Clayton, Ga. 3c 52 Dougherty, Ga. 3.31 Harrison, Miss. 3.10 Cobb, Ga. 3.08 Muscogee, Ga. 3.05 Madison, Ala. 2.91 Onslow, N. C. 2.84 Choctow, Ala. 2.69 Clarke, Ga. 2.60 Orange, N. C. 2.35 Montgomery* Ala. 2.31 Wake, N. C. 2.23 Cumberland, N. C. 2.15 Camden, Ga. 2.01 The last two dimensions like those in 1950 are difficult to identify. Dimension VI appears to be another locational index. This component is related to national, regional and local market potentials. Table 32 suggests that

PAGE 83

73 counties which rank high on this factor are located close to urban centers. They tend to have lower unemployment rates and a younger population than counties which have a less favourable location. The last component is difficult to name. Table 33 shows that the salient variables correlated with this factorare old age and median value of owner-occupied dwelling units, and value added by manufacturing. This dimension may isolate areas of declining prosperity, suggesting the out-migration of the younger and better educated labour force. AH though Table 34 is not conclusive, it implies that out-migration may be the key to the explanation of this dimension. Table 31 Dimension VI Locational Advantage in Relation to National, Regional and Local Markets in I960 Primary Variables Factor Loadings Distance in miles from the center of the county to a city of at least 50,000 -.60 Per cent of the labour force unemployed -.83 Distance in miles from the center of the county to a city of at least 500,000 -.10 Di.'tance in miles from the center of the county to New York -.19 Per cent of the population 65 years of age or older -.10

PAGE 84

74 Table 32 Counties with Extreme Factor Scores on Dimension VI in I960 High Positive Loadings Crawford, Ga. 2.08 Edgefield, S. C. 2.06 Guilford, N. C. 2.04 Baker, Ga. 1.84 Hertford, N. C. 1.81 De Kalb, Ga. 1.76 Wake, N. C. 1.74 Payette, Ga. 1.73 Glascock, Ga. 1.71 Webster, Ga. 1.71 Candler, Ga. 1.63 Forsythe, N. C. 1.63 High Negative Loadin gs Bolivar, Miss. Washington, Miss. Pamlico, N. C. Carteret, N. C. Clay, N. C. Hyde, N. C. Avery, N. C. Brunswick , N . C . Rabun, Ga. Pearl River, Miss. Jackson, Miss. Adams, Miss. Jasper, Miss. New Hanover, N. C. Pitt, N. C. Table 33 Dimension VII Areas of Declining Prosperity in I960 -3.85 -3.28 -3.17 -3.10 -2.88 -2.88 -2.72 -2.56 -2.36 -2.31 -2.12 -2.29 -2.02 -2.00 -2.00 Primary Variables Per cent of the population 65 years of age or older Median value of owner-occupied dwelling units Median gross rent Value added by manufacturing Factor Loadings -.89 -.54 .30 -.31 Table 34 Counties with

PAGE 85

75 Table 34 — Continued High Positive Loading s High ne g ative Loadings Houston, Ga.

PAGE 86

76 Although dimension IV in 1950 (Rural White Counties) and dimension II in I960 (Rural Negro Counties) are named differently, they are considered to be similar as both are indicators of the color problem. They have similar factor loadings with the exception that the correlation signs are reversed. Thus one dimension can be considered a reciprocal of the other. The changes in rank order are reflected in dimensions II through V. Negro poverty has moved from fourth in order of importance in 1950 to second in I960 and the per cent of variation contributed has increased from 7.4 to 13.3 per cent, an increase of 5.9 per cent. This change in the importance of the Negro dimension implies that the Negroes have not shared equitably in the economic progress of the Southeast, during this decade. The decrease of 2.7 per cent in explained variation of component II in 1950, tenant and smal3 farm poverty, has reduced it to third in order of importance in I960. The decrease in this form of rural poverty is in accord with the mechanization and the consolidation of farm units that has occurred in the Southeast. In many instances this process of consolidation and mechanization has resulted in a surplus labour supply. Agricultural unemployment is responsible for the increased concentration of Negro poverty in the Mississippi Delta. The mechanization of cotton farming in the Delta has greatly reduced the labour requirements and has either left the Negro unemployed, or has forced the

PAGE 87

77 excess labour supply to migrate elsewhere, particularly to the North. This similar process has been at work in southern Georgia, southern Alabama and southwestern Mississippi, where the pulp and paper companies have become established and have reverted large areas of crop land to pine forest. The shift of dimension III in 1950 (wealthy urbanized counties) to dimension V in I960 is an indication that there has been a narrowing of the range in prosperity between cities. This is interpreted as demonstrating that urban centers have shared in economic growth more equitably than rural areas. This is shown by the increase in rural poverty in certain sections of the study area as expressed by components II and IV in I960. In 1950 these components accounted for 13 per cent of the total variation while in I960 they accounted for 20*5 per cent. The locational factor relating to the national market has moved in this time period from fifth place in 1950 to fourth place in I960. On the other hand the regional and local market locational index has remained stable not only in rank but it also only registered a slight increase in total explained variation. This may be the result of an improved transportation network that has made these markets more accessible. The new dimension that emerged in I960 is not only difficult to identify but it appears to have contributed little to the overall analysis.

PAGE 88

78 Prom these dimensions it is clear that the variance in economic health throughout the study area in 1950 and I960 is a function of urban ism, color, location, industrialization, and rural poverty. It is observed that the rate and form of urban growth is most important, for not only do the urban areas harbour the majority of the region's poverty, but also influence to some extent the growth rates of their hinterlands. It is suggested that the increased rural poverty that is arising in certain sections, due to increasing technology, will become of greater concern unless an outlet is found for the surplus supply of labour. R egress ion Ana lysis The major underlying patterns of economic health factored from the original 26 variables can be tested for hypotheses by the use of a regression model. As regression analysis requires normal and independent data, the testing of the normality of the factor scores and the independence of the dimensions is of theoretical importance. In addition it is of interest to ascertain how well these dimensions explain the variation in poverty throughout the study area. To test these and additional hypotheses a dependent variable, median family income, was chosen as an index of poverty. This variable was then regressed on the dimensions in a stepwise regression.

PAGE 89

79 Th e Hypothese s Since the dimensions are theoretically representative of the original data, it is hypothesized that a high coefficient of determination between the dimensions and median family income will be obtained. It is hypothesized that the I960 dimensions will result in a higher coefficient of determination, as they have higher communalities than the 1950 components. In addition, hypotheses in regard to the expected correlations between the dimensions and the dependent vsiriable can be tested. The expected relationships are shown in Table 35. Table 35 Expected Correlations Between Median Family Income and the Derived Dimensions Median Family _ Income Urban Industrial Counties Tenant and Small Farm Poverty Wealthy Urbanized Counties + Rural Negro Counties Ideational Advantage in Regard to National Market * Locational Advantage in Regard to Regional and Local Markets + Areas of Declining Prosperity Rural White Counties + The analysis verified the normality of the factor scores and the independence of the dimensions (Tables 36, 37, 39 and 40). The hypothesis regarding the coefficient of determination was found to be correct (Tables 38 and 41), However, the correlations between the dependent and the

PAGE 90

80 independent variables were found as expected with one exception, Urban Industrial dimension had a positive correlation instead of a negative correlation as expected (Tables 37 and 40). Table 36 Normality of 1950 Factor Scores Varia ble Median Family Income Urban Industrial Counties Tenant and Small Farm Poverty Wealthy Urbanized Counties Rural White Comities Locational Advantage in Regard to National Market Locational Advantage in Regard to National, Local and Regional Markets Mean St andard JDeyiat ion 1445.38525 572.95850 -0.00002 .99995 0.00003 *99997 0.00022 .99997 -0.00013 .99997 -0.00000 .99997 -0.0004 .99995 Variable Table 37 1950 Correlation Matrix 1 2 3 4 1 Median Family Income 1.000 0.365 -0.124 0.623 0.334 2 Urban Industrial Counties 1.000 -0.000 0.000 -0.000 3 Tenant and Sma.ll Farm Poverty 1.000 0.000 -0.000 4 Wealthy Urbanised Counties 1.000 -0.000 5 Rural White Counties 1.000 6 Locational Advantage in Relation to National Market 7 Locational Advantage in Relation to National, Regional and Local Markets 0.301 0.2.12 0.000 0.000 -0.000 -0.000 -0.000 -0.000 0.000 0.000 1.000 0.000 1.000

PAGE 91

81

PAGE 92

82 Table 41 I960 Summary Table Step Variable Multiple 2 2 Number Entered R R Increase in R 1

PAGE 93

83 urbanization and rural poverty. Dimension VII contributed very little to the analysis. Although the urban industrial dimension did not correlate negatively as expected, this was probably due to the large percentage of high incomes found in these urban industrial counties which tend to inflate median family incomes, and thus to a degree hide the large number of poor families in these counties. There have, however, been changes in the significance of these dimensions in explaining the differences in medjan family income (Table 41). The influence of the locational dimensions has dropped rapidly from 13.58 per cent of the explained variation in 1950 to 3.62 per cent in I960. Almost all of this loss is accounted for by the dimension representing location in regard to the national market, as this component was not significant in the I960 regression analysis. This may reflect the overall economic growth that was achieved in the South during the two time periods, especially in the cities. In contrast to the decrease in the importance of locs/tion, urban growth played a larger role in I960. The dimension representing urban wealth accounted for an additional 10 per cent in explained variation. On the other hand rural poverty had a slightly higher negative correlation in I960 than in 1950, indicating that in general rural areas are today at a larger disadvantage. This also applies to the Negro, who has lost ground, during the time period studied. This appears also to be borne out by the beta coefficients, with the exception of the beta value, associated with locution in

PAGE 94

84 relation to the national market, as the beta coefficients in I960 are substantially larger than those in 1950. This is an indication that the income gap between the poor and the rich has actually widened in this time period. Summary The factor and regression analysis shows that the urban-rural and the Negro-white dichotomy is increasing. The analysis indicates that economic growth and opportunity are being centered around urban complexes to the disadvantage of rural areas. Furthermore, the economic gap between the Negro and white segments of the population in this region has not narrowed. These conclusions are in accord with the theories regarding poverty and regional growth which were examined in Chapter I.

PAGE 95

REFERENCES "Leslie J. King, "Cross Sectional Analysis on Canadian Urban Dimensions 1951 and 1961." C an ad i ah G e o gra ph e r , X, No. 4 (1966), 205-224. Wayne Hoffman and Gerald Romsa, "Some Temporal Considerations of Basic Urban Dimensions in the Southeast," Paper read before the meeting of the Southeastern Division of the Association of American Geographers, Greenville, North Carolina, November 26, 1968, 85

PAGE 96

CHAPTER IV SUMMARY AND CONCLUSIONS This chapter summarizes the findings of the factor and regression analysis for two time periods of five southeastern states, Alabama, Mississippi, Georgia, South Carolina and North Carolina. These states were selected as they constitute a socio-economic region and are representative of the southeastern United States. The prime objective of the study was to identify the major dimensions of economic health of the study area for two time periods 1950 and I960, and to ascertain if these dimensions had c inged throu ;h til e. In the analysis, the areal variation in econorai ] e; lth was measured by using the multivariate procedure of factor analysis, as the 26 variables selected and used as indices of economic health did not meet the statistical premises of normality and independence. Furthermore, factor scores derived from factor analysis were used in interpreting the areal pattern of the dimensions of economic health., These same factor scores were also used in a regression equation to explain the variance in median family income. 8c

PAGE 97

87 1950 Analysis The 1950 data were grouped into six dimensions which represented 70.5 per cent of the total variance of the original 26 variables. Dimension I was interpreted as being an indicator of industrialization and urbanization. Count: ranking high on this component contained the largest urban centers in the study area. The highest factor scores were associated with the larger and more rapidly growing centers such as Atlanta, Birmingham and Charlotte. Factor II was a low income agricultural component. Variables characterizing this dimension revealed an agricultural structure which was remnant of the sharecropping era. Many of the farmers, in the counties that were identified by this component, are not able to compete successfully with agricultural enterprises that use large inputs of land, capital and managerial skill as substitutes for labour and thus are able to operate on a larger scale. Counties fitting this pattern were generally located in the old cotton belt. The third dimension revealed a pattern of urban counties. These counties contained smaller but wealthier cities than the urban nodes represented by dimension I. These cities contain a high proportion of residents who are employed in certain high paying segments of the service sector, such as education and the military. Other counties which are represented by this dimension contain wealthy dormitory suburbs.

PAGE 98

88 Factor IV was a component defining the largely white population of Appalachia and the predominantly Negro Mississippi Delta. This dimension although reflecting poverty in both areas, identified the disparity in level of education attained and land ownership between white and the nonwhite segments of the population. The last two dimensions were a reflection of the effect of location in regard to local, regional and national markets. No definitive pattern arose, although there is some empirical evidence that poverty in southwestern Mississippi is due to its disadvantageous location. On the basis of the derived dimensions, areal patterns of economic health were derived. In general a picture arose of scattered peaks representing counties with a high proportion of urban residents and or industrial development that are surrounded by a sea of poverty. In the northern portion of the study area are the poor whites of Appalachia. In. the Mississippi delta poverty is largely comprised of Negro agricultural workers, many of whom are unemployed. Poverty i:^ southwestern Mississippi appears to be a reflection of isolation. Throughout the entire area are found the poor tenant and smal] farmer. Factor scores obtained from the analysis were utilized as independent variables in a regression model to test the significance of ill': dimensions in explaining the variation in median family income. The regression analysis did

PAGE 99

89 support the areal pattern of economic health as identified by the six dimensions. The six factor scores obtained from the factor analysis explained 78.45 per cent of the variation in median family income. Median income was found to be positively correlated with urbanization, industrialization, and the white portion of the population. The three dimensions representing these indices accounted for 63.34 per cent of the explained variation. Location and agricultural poverty contributed smaller amounts to the total coefficient of determination. The beta coefficients also substantiated the correlation coefficients. Dimension III had a beta value of 357.09 implying that a one unit increase in this dimension would result in an increase of $357.09 in median family income. Since this dimension is largely a reflection of education, it bears out the economic value of public investment in the schools of this region. A one unit increase in the urban industrial dimension would result in an increased median family income of 3209-26. Being white or located close to urban centers resulted also in higher incomes. Agricultural employment, however, resulted in lower incomes as did a decrease in the white dimension. I960 Analysis The I960 data were reduced by factor analysis to seven dimensions, which explained 77.9 per cent of the variation in economic health. The increase in explained variation

PAGE 100

90 over the 1950 period was clue to the derivation of an additional component and a more complete factoring of the original variables. Except for dimension VII (areas of declining prosperity) the components obtained for this time period were similar to those derived in 1950, although their order of importance had changed. Factor I as in 1950 was associated with urban industrial counties. It had similar factor loadings and was identified with counties containing the larger urban centers. The second component pointed out the economic discrimination against the Negro. This factor differentiates between Negro poverty, of which a large proportion is concentrated in the Mississippi Delta, and the poverty of white Appalachian Although the poverty of Appalachia has been recognized by various levels of government through different aid programs, little has been done for the Delta area. Dimension III represents rural poverty found throughout the Southeast. Rural poverty is found among both the white and the Negro operators of small land holdings or tenants. Dimension IV, a locational index, reflects the effect of proximity to urban markets. The areal pattern of factor scores although difficult to interpret suggests that Mississippi and Alabama suffer by not having closer access to the national market.

PAGE 101

91 Dimension V is another urban component. It represents smaller prosperous cities that have benefited from the location of high paying service functions and wealthy residential suburbs. In general this component reflects the well-educated labour force found in these urbanized counties. The last two dimensions are difficult to interpret; however they ceem to represent a locational index and possible areas of declining prosperity. When the factor scores of the above dimensions were regressed upon median family income for I960 a coefficient of determination of 81.01 was obtained. The most significant dimensions in order of importance were wealthy urbanized counties, urban, industrial counties and rural Negro counties. These three components together explained 73.63 per cent of the total variation in median family income. Less significant were tenant and small farm poverty and location in regard to local, regional and national markets. Dimensions IV and VII were found to be not significant. The beta values were in accord with the correlation coefficients. For example, the component representing wealthy urbanized counties which had the highest correlation also had the highest beta value (686.11). Thus a one unit increase in this dimension would result in an increase of $686.11 in median family income. Conversely a one unit increase in the second most significant dimension, rural Negro counties, would result in a decrease in income of $357.91.

PAGE 102

92 Comparisons of Changes in the Two Time Periods Studi ed" The two time periods revealed similar dimensions of economic health and essentially the same areal patterns when factor scores were mapped. Although the factor loadings varied slightly in the magnitude of the correlations between the variables and the dimensions, the differences were not sufficient to be considered of any significant value. Despite the similarity, there were changes in the rank order of the dimensions. The Negro component moved into second place in 3.960, which is evidence that the nonwhite population in the Southeast had not shared equitably in the overall economic growth. The decrease in the importance of tenant and small farm poverty is in line with the agricultural progress attained in the Southeast during this decade. Similarly, the downward shift of the wealthy urbanized dimension is in accord with the overall progress made in the cities in regard to integration, wages and education. The upward shift of the importance of the national market, from fifth place to fourth, was unexpected. This perhaps is an indication of the more complete integration of the southern economy into the national economy. The results obtained from the regression analysis were even more stable than those obtained from factor analysis with two exceptions. The national market had no significant effect on median family income in I960, implying

PAGE 103

93 that since 1950 the differences in median family income in the study area have narrowed. The beta values for the I960 time period were, however, substantially higher than those obtained in 1950, an indication that the income gaps between the poor and the more affluent had actually widened. Conclus ions The results cf the analysis support empirically the hypotheses outlined in Chapter I. Economic health can be adequately summarized by a small number of factors. As expected, these dimensions did not remain stable in regard to their rank order, for the components reflect the changes in economic development end as such are indicators of change. Factor analysis furthermore revealed that the urban industrial nodes are centers of growth. The counties that were located in close proximity to tliese nodes were more prosperous than those further away. This was substantiated in the regression analysis. If family income is used, as an index of prosperity, then urban areas definitely provide more and better opportunities for higher wages. Higher wages provide a stimulus for migration to these cities. Continued population growth results in an expanding consumer market and consequently economic growth. The decrease in the importance of location relative to the national market may be due to the improved transportation system in the Southeast or it may be due to the fact bhi t the market in the study area

PAGE 104

94 is now large enough to generate its own growth. However, this increase in prosperity has not been achieved by a large segment of the rural and Negro population. Economic growth has been confined largely to the urban centers. Consequently more effort must be directed towards the rural poor if this region is to achieve the national level of median family income. The lag in change in the rural areas suggests that social mores are more difficult to overcome than the lack of capital and technological knowledge. Possibly this is a sign of a slower rate of the diffusion of ideas in rural areas, where there is less personal interaction and thus less pressure to change. Future Research Efforts This study answers some questions and uncovers many problems. The general purpose of the undertaking was realised. The underlying dimensions of economic health were derived and broad areal patterns of economic regions were identified. In addition the importance and the changes of these dimensions for two time periods were established, not only in regard to the variation in economic health but also in how they affected the deviations in median family income. The analysis furthermore revealed that there are subregional variations in economic health. The study is ha ipered somewhat in that it gives a generalized account of the region. Further research should

PAGE 105

S5 be directed to a subregional analysis. One step would be to derive homogeneous economic regions based upon factor scores. Such computer programs are now available but unfortunately were not obtainable for this study. Additional factor and regression analyses could then be performed so that comparisons could be made between the subregions and between the subregions and the study area. This would provide greater insight into local problems concerning economic health. Similar comparisons could be made between this region and other regions. In addition geographers could study the rate of diffusion and acceptance of new ideas in rural and urban areas. The differences in the rate of diffusion and their causes would be of aid in discerning why different levels of economic growth and prosperity exist. One problem that arises in doing a study covering several time periods is the availability of comparable data, as much of the census data can not be used as they were collected in different categories. Some of the variables used in this initial attempt were redundant and should be replaced by other parameters. Direct distances as indices of location might be substituted for by market potential or transportation cost indices. Another major obstacle to the study of economic health in any region is the absence of psychological measures. The rate and form of economic growth is a behavioral response to perceived conditions. Research should be performed on the behavioral aspects of regional growth in order to help

PAGE 106

96 provide a more complete analysis of the reasons for the geographic variations in economic health.

PAGE 107

BIT3LI0GHAPHY Books Ahmad, Qazi. Indi an Cities: Characterist ics and Corr elate I-. . Chicago: University"of Chicago "Press, 196"!?."" Bogue, Donald J., and Beale, Calvin L. Economic Areas of the Un ited States . New York: The Free Press of Glencoe, Inc., 1961. Cohen, Saul B. (ed.). Probl ems and Trends in American Geogr aphy. New York: Basic Books, "inc., 1967. Duncan, Otis Dudly, et__al. Metropolis and Region. Baltimore: The John Hopkins Press, I960,. Eaton, Clement. A History of the Old South . New York: The Mac mil Ian Co., 1*9 6 5 . Perman, Louis A., Kornbluh, Joyce L., and Haber, Alan (ed.). Poverty in America. Ann Arbor: University of Michigan Press, 1965. Pr ie dm an , J o hn an d Al o n s o , Will i am ( e d . ) . Regional Devc l.o J > merit an/ -1 Planning. Cambridge: Massachusetts Institute of Technology Press, 1964. Galbraith, John, The Affluent Soc ioty . Boston: HoughtonMifflin co ; ,"" 1$W. Ginsburg, Norton (ed.). Essays on Geography and EcQ] De velopment. Chicago':"" The "University of Chic, Press, l~§o"0. Griffin, Paul P., Young, Robert N., and Chatham, Ronald L. Anglo -America: A Regi onal G eogr aphy of the United states ana uanaaa . ban Jj'rah'ciscb: U'earon Publishers , lncTi"~T96k'. Harmon; H. H. Modern Factor Ana lysi s. Chicago: University of Chicago Press, l96l. 97

PAGE 108

98 Harrington, Michael. 2^ ie _°^ 1 ± 1 Z.^l£I3-9iL : ™ Poverty in the United States . Bait imo re: Penguin Books, 1966. Hart, John. Fraser. The Southeas tern U nited States. Princeton: D. Van No strand Co . , IncT^"T-9^77 "~~ Kendall, M. G„ A_Course in Mul tivariate Ana lysis. London* Charles GrTffin, Y957. " ' Maddox, James G. , et__al, The Advancing Southj_ Manpower STSi 3 >l£c^s__an d ""-Pro "bl^ems~~( iTe y; York": T?jbe TweniTeth Century Fund, 19 67 ) . Miller, Herman P. Poverty_Amejrican J3tyle, Belmont": Wadsworth PubiTsTfihg Co."7~~Tnc., T966. Odum, Howard W. Sg^^err^RegJjQ^^^f_the JJnited States. Chanel Hill: The University of~North Car o 1 fna" Pr ess 1936. Schlesinger, Arthur M. A History of American Life. Vol. XI: The Rise of""lKe CiTy~T87B-l8"9b r : New~YorkT"" The Macmillan Co., "13T49 . ' Schultz, T. W. Economic Organization ofAgri culture. New York: McGraw Hill Co., 1'953. Thoman, R, S, (cd.). Areas o f Economic Stress in Canada. Kingston: Queens University Pfvss, l96"5.~ Vance, Rupert B, Human Geography of the South. Chapel Hill: University of North Carolina Press, 1935. White, Langdon G, and Foscue, Edwin J. Regional Geography of Angl c -Am er ica . 2 d e d . Engl e wo o d CI if f s : Prentice-Hall, 1956. Articles, Report s and Periodicals Bell, W. H. and Stevenson, P. W. "An Index of Economic Health for Ontario Counties and Districts," Ontario Economic Review , II, No. 5 (1964), 1-7. ' " Fuguitt, Glenn V. "County Seat Status as a Factor in Small Town Growth and Decline," Social Forces, XLIII (December, 1965), 245-251. ' ~~ Herzog, Elizabeth. "Some Assumptions About the Poor," Social Science Review^ XXXVII (December, 1963), 390-4107"

PAGE 109

99 Hodge, Gerald. "The Prediction of Trade Center Viability in the Great Plains," The Re gional Sc ience Association Papers, XV U965), 07-118*. . "Do Villages Grow? Some Perspectives and Predictions," Rural Sociology, XXX, iNo. 2 (June, iy66). 183-196. " ""-" King, Leslie J. "Cross Sectional Analysis on Canadian Urban Dimensions 1951 and 1961," Canadian Geographer, X, No. 4 (1966), 205-224. Laird, William E. and Rinehart, James R. "Exogeneous Check on Southern Economic Development," So uth Atlantic Quarterly, LXV, No. 4 (1966), 491-50B. Lassiter, Ray L., Jr. "Education for Males by Region, Race, and Age," So uthe r n Economic Journal, XXXII (July, 1965), 15-22". Levitan, Lar A. "Programs in Aid of the Poor," Pov ert y and Human Resources A bstracts, I (1966), 11-25. MacDonald, Stephen L. "On South' s Recent Economic Development," Southern Economic Journal, XXVIII (July, 1961), 30-40. ~~" Megee, Mary. "On Economic Growth and the Factor Analysis Method," Sou thern Econo mic Journal, XXXI (July, 1964), 215-228. Moon, Seung Gyu and McCann, Glenn C. Subregipn al Var i ability of Ad ju stme n t F act ors ^ of Rui-al IViimilie s in the S outh . Raleigh: North Carolina Agricultural Experiment"" Station, Southern Cooperative Series, Bulletin No. Ill, January, 1966. Nichols, William II. "Southern Tradition and Regional Economic Progress." Southern Economic Journal, XXVI (January, I960), 187-198. Ratchford, B. U. "Economic Development in the South," South Atlantic Quarterly, LXIV, No. 4 (1965), 496-505. Rummel, R. J. "Understanding Factor Analysis," Conflict Resolution , XI, No. 4 (1967), 444-480. ' Scott, John T., Jr. "Factor Analysis and Regression," Economefrica , XXXIV, No. 3 (July, 1967), 552-562.

PAGE 110

100 Public Documents Berry, Brian J. Stra tegies, Models and Economic Theories of Development in Rural Regions. Agriculti Economic Report No. 127. Washington: Economic Research Service, U.S. Department of Agriculture, 1967. Unpublished Material Hoffman, Wayne and Romsa, Gerald. "Some Temporal Considerations of Basic Urban Dimensions in the Southeast," Paper read before the 23rd annual meeting of the Southeastern Division of the Association of American Geographers, Greenville, North Carolina, November 26, 1968. Spinelli, Joseph. "A Study of Net Out-Migration in Southeastern Ohio, 1950-1960/' Unpublished Master f s Thesis, Department of Geography, Ohio State University, 1966.

PAGE 111

BIOGRAPHICAL SKETCH Gerald Harry Romsa was born May 31 » 1942, at Dolyna, Ukraine. In February, 1948, he immigrated to Canada. In June, I960, he was graduated from Oakville High School. In May, 1963, he received the degree of Bachelor of Science from the University of Manitoba. Prom 1963 until 1964 he served with the Department of National Health and Welfare as a Food and Drug Inspector in Toronto. In 1964 he entered the Graduate School of the University of Waterloo where he received the degree of Master of Arts in 1967. From September, 1966, when he enrolled in the Graduate School of the University of Florida, until the present time he has pursued his work toward the degree of Doctor of Philosophy, Gerald Harry Romsa is married to the former Mary Fraser Reid and they have one son. He is a member of the American Geographical Society, the Association of American Geographers, and Gamma Theta Upsilon, honorary geography fraternity. 101

PAGE 112

This dissertation was prepared under the direction of the chairman of the candidate's supervisory committee and has been approved by all members of that committee. It was submitted to the Dean of the College of Arts and Sciences and to the Graduate Council, and was approved as partial fulfillment of the requirements for the degree of Doctor of Philosophy. June, 1969 Dean, College/ o ;s and Sciences Sup orv i s o ry C ommit t e e Dean, Graduate School <: ^a^&-J. Chairman


xml version 1.0 encoding UTF-8 standalone no
fcla fda yes
dl
!-- spatial analysis of the dimensions economic health in southeastern United States (1950 and 1960) ( Book ) --
METS:mets OBJID UF00097777_00001
xmlns:METS http:www.loc.govMETS
xmlns:mods http:www.loc.govmodsv3
xmlns:xlink http:www.w3.org1999xlink
xmlns:xsi http:www.w3.org2001XMLSchema-instance
xmlns:daitss http:www.fcla.edudlsmddaitss
xmlns:sobekcm http:digital.uflib.ufl.edumetadatasobekcm
xsi:schemaLocation
http:www.loc.govstandardsmetsmets.xsd
http:www.loc.govmodsv3mods-3-3.xsd
http:www.fcla.edudlsmddaitssdaitss.xsd
http:digital.uflib.ufl.edumetadatasobekcmsobekcm.xsd
METS:metsHdr CREATEDATE 2010-09-20T08:54:10Z ID LASTMODDATE 2010-02-04T00:00:00Z RECORDSTATUS NEW
METS:agent ROLE CREATOR TYPE ORGANIZATION
METS:name UF
METS:note server=TC
projects=
OTHERTYPE SOFTWARE OTHER
Go UFDC - FDA Preparation Tool
INDIVIDUAL
UFAD\mariner1
METS:dmdSec DMD1
METS:mdWrap MDTYPE MODS MIMETYPE textxml LABEL Metadata
METS:xmlData
mods:mods
mods:genre authority marcgt bibliography
non-fiction
mods:identifier type AlephBibNum 000559424
OCLC 13489737
NOTIS ACY4880
mods:language
mods:languageTerm text English
code iso639-2b eng
mods:location
mods:physicalLocation University of Florida
UF
mods:name personal
mods:namePart Romsa, Gerald Harry
given Gerald Harry
family Romsa
date 1969
mods:role
mods:roleTerm Main Entity
mods:note thesis Thesis--University of Florida.
bibliography Bibliography: leaves 97-100.
additional physical form Also available on World Wide Web
Manuscript copy.
Vita.
mods:originInfo
mods:place
mods:placeTerm marccountry xx
mods:dateIssued marc 1969
point start 1969
mods:copyrightDate 1969
mods:recordInfo
mods:recordIdentifier source ufdc UF00097777_00001
mods:recordCreationDate 860424
mods:recordOrigin Imported from (ALEPH)000559424
mods:recordContentSource University of Florida
marcorg FUG
mods:languageOfCataloging
English
eng
mods:relatedItem original
mods:physicalDescription
mods:extent ix, 101 leaves. : illus. ; 28 cm.
mods:subject SUBJ651_1 lcsh
mods:topic Economic conditions
mods:geographic Southern States
SUBJ690_1
Geography thesis Ph. D
SUBJ690_2
Dissertations, Academic
Geography
UF
mods:titleInfo
mods:nonSort A
mods:title spatial analysis of the dimensions of economic health in the southeastern United States (1950 and 1960)
mods:typeOfResource text
DMD2
OTHERMDTYPE SobekCM Custom
sobekcm:procParam
sobekcm:Collection.Primary UFIR
sobekcm:Collection.Alternate VENDORIA
sobekcm:SubCollection UFETD
sobekcm:MainThumbnail spatialanalysiso00romsrich_Page_001thm.jpg
sobekcm:Download
sobekcm:fptr FILEID UR2
sobekcm:EncodingLevel I
sobekcm:bibDesc
sobekcm:BibID UF00097777
sobekcm:VID 00001
sobekcm:Source
sobekcm:statement UF University of Florida
sobekcm:Type Book
sobekcm:SortDate -1
METS:amdSec
METS:digiprovMD AMD_DAITSS
DAITSS
daitss:daitss
daitss:AGREEMENT_INFO ACCOUNT PROJECT UFDC
METS:fileSec
METS:fileGrp USE reference
METS:file GROUPID G1 J1 imagejpeg SIZE 41713
METS:FLocat LOCTYPE OTHERLOCTYPE SYSTEM xlink:href spatialanalysiso00romsrich_Page_001.jpg
G2 J2 21040
spatialanalysiso00romsrich_Page_002.jpg
G3 J3 14064
spatialanalysiso00romsrich_Page_003.jpg
G4 J4 46194
spatialanalysiso00romsrich_Page_004.jpg
G5 J5 88504
spatialanalysiso00romsrich_Page_005.jpg
G6 J6 36632
spatialanalysiso00romsrich_Page_006.jpg
G7 J7 81337
spatialanalysiso00romsrich_Page_007.jpg
G8 J8 99376
spatialanalysiso00romsrich_Page_008.jpg
G9 J9 89628
spatialanalysiso00romsrich_Page_009.jpg
G10 J10 32193
spatialanalysiso00romsrich_Page_010.jpg
G11 J11 103833
spatialanalysiso00romsrich_Page_011.jpg
G12 J12 121080
spatialanalysiso00romsrich_Page_012.jpg
G13 J13 106211
spatialanalysiso00romsrich_Page_013.jpg
G14 J14 123635
spatialanalysiso00romsrich_Page_014.jpg
G15 J15 113020
spatialanalysiso00romsrich_Page_015.jpg
G16 J16 116576
spatialanalysiso00romsrich_Page_016.jpg
G17 J17 122631
spatialanalysiso00romsrich_Page_017.jpg
G18 J18 104609
spatialanalysiso00romsrich_Page_018.jpg
G19 J19 102790
spatialanalysiso00romsrich_Page_019.jpg
G20 J20 118991
spatialanalysiso00romsrich_Page_020.jpg
G21 J21 119702
spatialanalysiso00romsrich_Page_021.jpg
G22 J22 122481
spatialanalysiso00romsrich_Page_022.jpg
G23 J23 131338
spatialanalysiso00romsrich_Page_023.jpg
G24 J24 115610
spatialanalysiso00romsrich_Page_024.jpg
G25 J25 107432
spatialanalysiso00romsrich_Page_025.jpg
G26 J26 124102
spatialanalysiso00romsrich_Page_026.jpg
G27 J27 119845
spatialanalysiso00romsrich_Page_027.jpg
G28 J28 128627
spatialanalysiso00romsrich_Page_028.jpg
G29 J29 132639
spatialanalysiso00romsrich_Page_029.jpg
G30 J30 123284
spatialanalysiso00romsrich_Page_030.jpg
G31 J31 123660
spatialanalysiso00romsrich_Page_031.jpg
G32 J32 83627
spatialanalysiso00romsrich_Page_032.jpg
G33 J33 89456
spatialanalysiso00romsrich_Page_033.jpg
G34 J34 108718
spatialanalysiso00romsrich_Page_034.jpg
G35 J35 134394
spatialanalysiso00romsrich_Page_035.jpg
G36 J36 31892
spatialanalysiso00romsrich_Page_036.jpg
G37 J37 107400
spatialanalysiso00romsrich_Page_037.jpg
G38 J38 93995
spatialanalysiso00romsrich_Page_038.jpg
G39 J39 111286
spatialanalysiso00romsrich_Page_039.jpg
G40 J40 114835
spatialanalysiso00romsrich_Page_040.jpg
G41 J41 125344
spatialanalysiso00romsrich_Page_041.jpg
G42 J42 116973
spatialanalysiso00romsrich_Page_042.jpg
G43 J43 119965
spatialanalysiso00romsrich_Page_043.jpg
G44 J44 122834
spatialanalysiso00romsrich_Page_044.jpg
G45 J45 118432
spatialanalysiso00romsrich_Page_045.jpg
G46 J46 118890
spatialanalysiso00romsrich_Page_046.jpg
G47 J47 115104
spatialanalysiso00romsrich_Page_047.jpg
G48 J48 99942
spatialanalysiso00romsrich_Page_048.jpg
G49 J49 107353
spatialanalysiso00romsrich_Page_049.jpg
G50 J50 97808
spatialanalysiso00romsrich_Page_050.jpg
G51 J51 134712
spatialanalysiso00romsrich_Page_051.jpg
G52 J52 85626
spatialanalysiso00romsrich_Page_052.jpg
G53 J53 100336
spatialanalysiso00romsrich_Page_053.jpg
G54 J54 78516
spatialanalysiso00romsrich_Page_054.jpg
G55 J55 109458
spatialanalysiso00romsrich_Page_055.jpg
G56 J56 116873
spatialanalysiso00romsrich_Page_056.jpg
G57 J57 64947
spatialanalysiso00romsrich_Page_057.jpg
G58 J58 88719
spatialanalysiso00romsrich_Page_058.jpg
G59 J59 90386
spatialanalysiso00romsrich_Page_059.jpg
G60 J60 104297
spatialanalysiso00romsrich_Page_060.jpg
G61 J61 65635
spatialanalysiso00romsrich_Page_061.jpg
G62 J62 105541
spatialanalysiso00romsrich_Page_062.jpg
G63 J63 119993
spatialanalysiso00romsrich_Page_063.jpg
G64 J64 84693
spatialanalysiso00romsrich_Page_064.jpg
G65 J65 110297
spatialanalysiso00romsrich_Page_065.jpg
G66 J66 114508
spatialanalysiso00romsrich_Page_066.jpg
G67 J67 111519
spatialanalysiso00romsrich_Page_067.jpg
G68 J68 120355
spatialanalysiso00romsrich_Page_068.jpg
G69 J69 116985
spatialanalysiso00romsrich_Page_069.jpg
G70 J70 100212
spatialanalysiso00romsrich_Page_070.jpg
G71 J71 74000
spatialanalysiso00romsrich_Page_071.jpg
G72 J72 95030
spatialanalysiso00romsrich_Page_072.jpg
G73 J73 101198
spatialanalysiso00romsrich_Page_073.jpg
G74 J74 114383
spatialanalysiso00romsrich_Page_074.jpg
G75 J75 62255
spatialanalysiso00romsrich_Page_075.jpg
G76 J76 105970
spatialanalysiso00romsrich_Page_076.jpg
G77 J77 103406
spatialanalysiso00romsrich_Page_077.jpg
G78 J78 92560
spatialanalysiso00romsrich_Page_078.jpg
G79 J79 124028
spatialanalysiso00romsrich_Page_079.jpg
G80 J80 109301
spatialanalysiso00romsrich_Page_080.jpg
G81 J81 124479
spatialanalysiso00romsrich_Page_081.jpg
G82 J82 101909
spatialanalysiso00romsrich_Page_082.jpg
G83 J83 93855
spatialanalysiso00romsrich_Page_083.jpg
G84 J84 107403
spatialanalysiso00romsrich_Page_084.jpg
G85 J85 109561
spatialanalysiso00romsrich_Page_085.jpg
G86 J86 121956
spatialanalysiso00romsrich_Page_086.jpg
G87 J87 116741
spatialanalysiso00romsrich_Page_087.jpg
G88 J88 107217
spatialanalysiso00romsrich_Page_088.jpg
G89 J89 100913
spatialanalysiso00romsrich_Page_089.jpg
G90 J90 96675
spatialanalysiso00romsrich_Page_090.jpg
G91 J91 108665
spatialanalysiso00romsrich_Page_091.jpg
G92 J92 97748
spatialanalysiso00romsrich_Page_092.jpg
G93 J93 124227
spatialanalysiso00romsrich_Page_093.jpg
G94 J94 66261
spatialanalysiso00romsrich_Page_094.jpg
G95 J95 40251
spatialanalysiso00romsrich_Page_095.jpg
G96 J96 90465
spatialanalysiso00romsrich_Page_096.jpg
G97 J97 114372
spatialanalysiso00romsrich_Page_097.jpg
G98 J98 114964
spatialanalysiso00romsrich_Page_098.jpg
G99 J99 112889
spatialanalysiso00romsrich_Page_099.jpg
G100 J100 108001
spatialanalysiso00romsrich_Page_100.jpg
G101 J101 117327
spatialanalysiso00romsrich_Page_101.jpg
G102 J102 113642
spatialanalysiso00romsrich_Page_102.jpg
G103 J103 110208
spatialanalysiso00romsrich_Page_103.jpg
G104 J104 106595
spatialanalysiso00romsrich_Page_104.jpg
G105 J105 122297
spatialanalysiso00romsrich_Page_105.jpg
G106 J106 19916
spatialanalysiso00romsrich_Page_106.jpg
G107 J107 100749
spatialanalysiso00romsrich_Page_107.jpg
G108 J108 124354
spatialanalysiso00romsrich_Page_108.jpg
G109 J109 131865
spatialanalysiso00romsrich_Page_109.jpg
G110 J110 59634
spatialanalysiso00romsrich_Page_110.jpg
G111 J111 81159
spatialanalysiso00romsrich_Page_111.jpg
G112 J112 65831
spatialanalysiso00romsrich_Page_112.jpg
E1 imagejp2 353484
spatialanalysiso00romsrich_Page_001.jp2
E2 220844
spatialanalysiso00romsrich_Page_002.jp2
E3 117514
spatialanalysiso00romsrich_Page_003.jp2
E4 437891
spatialanalysiso00romsrich_Page_004.jp2
E5 795291
spatialanalysiso00romsrich_Page_005.jp2
E6 326142
spatialanalysiso00romsrich_Page_006.jp2
E7 742838
spatialanalysiso00romsrich_Page_007.jp2
E8 911180
spatialanalysiso00romsrich_Page_008.jp2
E9 828758
spatialanalysiso00romsrich_Page_009.jp2
E10 289966
spatialanalysiso00romsrich_Page_010.jp2
E11 911145
spatialanalysiso00romsrich_Page_011.jp2
E12 916741
spatialanalysiso00romsrich_Page_012.jp2
E13 927000
spatialanalysiso00romsrich_Page_013.jp2
E14 916732
spatialanalysiso00romsrich_Page_014.jp2
E15 926989
spatialanalysiso00romsrich_Page_015.jp2
E16 916713
spatialanalysiso00romsrich_Page_016.jp2
E17 926971
spatialanalysiso00romsrich_Page_017.jp2
E18 916726
spatialanalysiso00romsrich_Page_018.jp2
E19 927005
spatialanalysiso00romsrich_Page_019.jp2
E20 916723
spatialanalysiso00romsrich_Page_020.jp2
E21 926979
spatialanalysiso00romsrich_Page_021.jp2
E22 916720
spatialanalysiso00romsrich_Page_022.jp2
E23 926975
spatialanalysiso00romsrich_Page_023.jp2
E24 916694
spatialanalysiso00romsrich_Page_024.jp2
E25 926986
spatialanalysiso00romsrich_Page_025.jp2
E26
spatialanalysiso00romsrich_Page_026.jp2
E27
spatialanalysiso00romsrich_Page_027.jp2
E28 916725
spatialanalysiso00romsrich_Page_028.jp2
E29 927002
spatialanalysiso00romsrich_Page_029.jp2
E30 916688
spatialanalysiso00romsrich_Page_030.jp2
E31 927004
spatialanalysiso00romsrich_Page_031.jp2
E32 816276
spatialanalysiso00romsrich_Page_032.jp2
E33 879421
spatialanalysiso00romsrich_Page_033.jp2
E34 916736
spatialanalysiso00romsrich_Page_034.jp2
E35
spatialanalysiso00romsrich_Page_035.jp2
E36 294392
spatialanalysiso00romsrich_Page_036.jp2
E37 926946
spatialanalysiso00romsrich_Page_037.jp2
E38 916709
spatialanalysiso00romsrich_Page_038.jp2
E39 926999
spatialanalysiso00romsrich_Page_039.jp2
E40 916733
spatialanalysiso00romsrich_Page_040.jp2
E41 926982
spatialanalysiso00romsrich_Page_041.jp2
E42
spatialanalysiso00romsrich_Page_042.jp2
E43 927007
spatialanalysiso00romsrich_Page_043.jp2
E44 916727
spatialanalysiso00romsrich_Page_044.jp2
E45 926995
spatialanalysiso00romsrich_Page_045.jp2
E46 916737
spatialanalysiso00romsrich_Page_046.jp2
E47
spatialanalysiso00romsrich_Page_047.jp2
E48 916712
spatialanalysiso00romsrich_Page_048.jp2
E49
spatialanalysiso00romsrich_Page_049.jp2
E50 916738
spatialanalysiso00romsrich_Page_050.jp2
E51 926945
spatialanalysiso00romsrich_Page_051.jp2
E52 851587
spatialanalysiso00romsrich_Page_052.jp2
E53
spatialanalysiso00romsrich_Page_053.jp2
E54 790356
spatialanalysiso00romsrich_Page_054.jp2
E55 926997
spatialanalysiso00romsrich_Page_055.jp2
E56
spatialanalysiso00romsrich_Page_056.jp2
E57 644813
spatialanalysiso00romsrich_Page_057.jp2
E58 873281
spatialanalysiso00romsrich_Page_058.jp2
E59 891666
spatialanalysiso00romsrich_Page_059.jp2
E60
spatialanalysiso00romsrich_Page_060.jp2
E61 660813
spatialanalysiso00romsrich_Page_061.jp2
E62 916729
spatialanalysiso00romsrich_Page_062.jp2
E63 919145
spatialanalysiso00romsrich_Page_063.jp2
E64 916700
spatialanalysiso00romsrich_Page_064.jp2
E65 919220
spatialanalysiso00romsrich_Page_065.jp2
E66 916693
spatialanalysiso00romsrich_Page_066.jp2
E67 919213
spatialanalysiso00romsrich_Page_067.jp2
E68
spatialanalysiso00romsrich_Page_068.jp2
E69 919177
spatialanalysiso00romsrich_Page_069.jp2
E70 916719
spatialanalysiso00romsrich_Page_070.jp2
E71 716586
spatialanalysiso00romsrich_Page_071.jp2
E72 916740
spatialanalysiso00romsrich_Page_072.jp2
E73 919234
spatialanalysiso00romsrich_Page_073.jp2
E74 916683
spatialanalysiso00romsrich_Page_074.jp2
E75 630873
spatialanalysiso00romsrich_Page_075.jp2
E76 916697
spatialanalysiso00romsrich_Page_076.jp2
E77 919225
spatialanalysiso00romsrich_Page_077.jp2
E78 916703
spatialanalysiso00romsrich_Page_078.jp2
E79 919232
spatialanalysiso00romsrich_Page_079.jp2
E80
spatialanalysiso00romsrich_Page_080.jp2
E81 919231
spatialanalysiso00romsrich_Page_081.jp2
E82
spatialanalysiso00romsrich_Page_082.jp2
E83 913283
spatialanalysiso00romsrich_Page_083.jp2
E84
spatialanalysiso00romsrich_Page_084.jp2
E85 919215
spatialanalysiso00romsrich_Page_085.jp2
E86
spatialanalysiso00romsrich_Page_086.jp2
E87 919207
spatialanalysiso00romsrich_Page_087.jp2
E88
spatialanalysiso00romsrich_Page_088.jp2
E89 919176
spatialanalysiso00romsrich_Page_089.jp2
E90 916685
spatialanalysiso00romsrich_Page_090.jp2
E91 919201
spatialanalysiso00romsrich_Page_091.jp2
E92
spatialanalysiso00romsrich_Page_092.jp2
E93 919236
spatialanalysiso00romsrich_Page_093.jp2
E94 649424
spatialanalysiso00romsrich_Page_094.jp2
E95 388097
spatialanalysiso00romsrich_Page_095.jp2
E96 908349
spatialanalysiso00romsrich_Page_096.jp2
E97
spatialanalysiso00romsrich_Page_097.jp2
E98
spatialanalysiso00romsrich_Page_098.jp2
E99 919216
spatialanalysiso00romsrich_Page_099.jp2
E100
spatialanalysiso00romsrich_Page_100.jp2
E101 919235
spatialanalysiso00romsrich_Page_101.jp2
E102
spatialanalysiso00romsrich_Page_102.jp2
E103 919206
spatialanalysiso00romsrich_Page_103.jp2
E104
spatialanalysiso00romsrich_Page_104.jp2
E105
spatialanalysiso00romsrich_Page_105.jp2
E106 172652
spatialanalysiso00romsrich_Page_106.jp2
E107 919187
spatialanalysiso00romsrich_Page_107.jp2
E108
spatialanalysiso00romsrich_Page_108.jp2
E109 919226
spatialanalysiso00romsrich_Page_109.jp2
E110 595477
spatialanalysiso00romsrich_Page_110.jp2
E111 785847
spatialanalysiso00romsrich_Page_111.jp2
E112 666297
spatialanalysiso00romsrich_Page_112.jp2
archive
F1 imagetiff 6.0 21891324
spatialanalysiso00romsrich_Page_001.tif
F2
spatialanalysiso00romsrich_Page_002.tif
F3
spatialanalysiso00romsrich_Page_003.tif
F4
spatialanalysiso00romsrich_Page_004.tif
F5
spatialanalysiso00romsrich_Page_005.tif
F6
spatialanalysiso00romsrich_Page_006.tif
F7
spatialanalysiso00romsrich_Page_007.tif
F8
spatialanalysiso00romsrich_Page_008.tif
F9
spatialanalysiso00romsrich_Page_009.tif
F10
spatialanalysiso00romsrich_Page_010.tif
F11
spatialanalysiso00romsrich_Page_011.tif
F12 22024964
spatialanalysiso00romsrich_Page_012.tif
F13 22271496
spatialanalysiso00romsrich_Page_013.tif
F14
spatialanalysiso00romsrich_Page_014.tif
F15
spatialanalysiso00romsrich_Page_015.tif
F16
spatialanalysiso00romsrich_Page_016.tif
F17
spatialanalysiso00romsrich_Page_017.tif
F18
spatialanalysiso00romsrich_Page_018.tif
F19
spatialanalysiso00romsrich_Page_019.tif
F20
spatialanalysiso00romsrich_Page_020.tif
F21
spatialanalysiso00romsrich_Page_021.tif
F22
spatialanalysiso00romsrich_Page_022.tif
F23
spatialanalysiso00romsrich_Page_023.tif
F24
spatialanalysiso00romsrich_Page_024.tif
F25
spatialanalysiso00romsrich_Page_025.tif
F26
spatialanalysiso00romsrich_Page_026.tif
F27
spatialanalysiso00romsrich_Page_027.tif
F28
spatialanalysiso00romsrich_Page_028.tif
F29
spatialanalysiso00romsrich_Page_029.tif
F30
spatialanalysiso00romsrich_Page_030.tif
F31
spatialanalysiso00romsrich_Page_031.tif
F32
spatialanalysiso00romsrich_Page_032.tif
F33
spatialanalysiso00romsrich_Page_033.tif
F34
spatialanalysiso00romsrich_Page_034.tif
F35
spatialanalysiso00romsrich_Page_035.tif
F36
spatialanalysiso00romsrich_Page_036.tif
F37
spatialanalysiso00romsrich_Page_037.tif
F38
spatialanalysiso00romsrich_Page_038.tif
F39
spatialanalysiso00romsrich_Page_039.tif
F40
spatialanalysiso00romsrich_Page_040.tif
F41
spatialanalysiso00romsrich_Page_041.tif
F42
spatialanalysiso00romsrich_Page_042.tif
F43
spatialanalysiso00romsrich_Page_043.tif
F44
spatialanalysiso00romsrich_Page_044.tif
F45
spatialanalysiso00romsrich_Page_045.tif
F46
spatialanalysiso00romsrich_Page_046.tif
F47
spatialanalysiso00romsrich_Page_047.tif
F48
spatialanalysiso00romsrich_Page_048.tif
F49
spatialanalysiso00romsrich_Page_049.tif
F50
spatialanalysiso00romsrich_Page_050.tif
F51
spatialanalysiso00romsrich_Page_051.tif
F52
spatialanalysiso00romsrich_Page_052.tif
F53
spatialanalysiso00romsrich_Page_053.tif
F54
spatialanalysiso00romsrich_Page_054.tif
F55
spatialanalysiso00romsrich_Page_055.tif
F56
spatialanalysiso00romsrich_Page_056.tif
F57
spatialanalysiso00romsrich_Page_057.tif
F58
spatialanalysiso00romsrich_Page_058.tif
F59 22085004
spatialanalysiso00romsrich_Page_059.tif
F60
spatialanalysiso00romsrich_Page_060.tif
F61
spatialanalysiso00romsrich_Page_061.tif
F62
spatialanalysiso00romsrich_Page_062.tif
F63
spatialanalysiso00romsrich_Page_063.tif
F64
spatialanalysiso00romsrich_Page_064.tif
F65
spatialanalysiso00romsrich_Page_065.tif
F66
spatialanalysiso00romsrich_Page_066.tif
F67
spatialanalysiso00romsrich_Page_067.tif
F68
spatialanalysiso00romsrich_Page_068.tif
F69
spatialanalysiso00romsrich_Page_069.tif
F70
spatialanalysiso00romsrich_Page_070.tif
F71
spatialanalysiso00romsrich_Page_071.tif
F72
spatialanalysiso00romsrich_Page_072.tif
F73
spatialanalysiso00romsrich_Page_073.tif
F74
spatialanalysiso00romsrich_Page_074.tif
F75
spatialanalysiso00romsrich_Page_075.tif
F76
spatialanalysiso00romsrich_Page_076.tif
F77
spatialanalysiso00romsrich_Page_077.tif
F78
spatialanalysiso00romsrich_Page_078.tif
F79
spatialanalysiso00romsrich_Page_079.tif
F80
spatialanalysiso00romsrich_Page_080.tif
F81
spatialanalysiso00romsrich_Page_081.tif
F82
spatialanalysiso00romsrich_Page_082.tif
F83
spatialanalysiso00romsrich_Page_083.tif
F84
spatialanalysiso00romsrich_Page_084.tif
F85
spatialanalysiso00romsrich_Page_085.tif
F86
spatialanalysiso00romsrich_Page_086.tif
F87
spatialanalysiso00romsrich_Page_087.tif
F88
spatialanalysiso00romsrich_Page_088.tif
F89
spatialanalysiso00romsrich_Page_089.tif
F90
spatialanalysiso00romsrich_Page_090.tif
F91
spatialanalysiso00romsrich_Page_091.tif
F92
spatialanalysiso00romsrich_Page_092.tif
F93
spatialanalysiso00romsrich_Page_093.tif
F94
spatialanalysiso00romsrich_Page_094.tif
F95
spatialanalysiso00romsrich_Page_095.tif
F96
spatialanalysiso00romsrich_Page_096.tif
F97
spatialanalysiso00romsrich_Page_097.tif
F98
spatialanalysiso00romsrich_Page_098.tif
F99
spatialanalysiso00romsrich_Page_099.tif
F100
spatialanalysiso00romsrich_Page_100.tif
F101
spatialanalysiso00romsrich_Page_101.tif
F102
spatialanalysiso00romsrich_Page_102.tif
F103
spatialanalysiso00romsrich_Page_103.tif
F104
spatialanalysiso00romsrich_Page_104.tif
F105
spatialanalysiso00romsrich_Page_105.tif
F106
spatialanalysiso00romsrich_Page_106.tif
F107
spatialanalysiso00romsrich_Page_107.tif
F108
spatialanalysiso00romsrich_Page_108.tif
F109
spatialanalysiso00romsrich_Page_109.tif
F110
spatialanalysiso00romsrich_Page_110.tif
F111
spatialanalysiso00romsrich_Page_111.tif
F112
spatialanalysiso00romsrich_Page_112.tif
R1 textx-pro 9212
spatialanalysiso00romsrich_Page_001.pro
R3 1098
spatialanalysiso00romsrich_Page_003.pro
R4 12758
spatialanalysiso00romsrich_Page_004.pro
R5 43085
spatialanalysiso00romsrich_Page_005.pro
R6 12940
spatialanalysiso00romsrich_Page_006.pro
R7 32020
spatialanalysiso00romsrich_Page_007.pro
R8 40597
spatialanalysiso00romsrich_Page_008.pro
R9 34361
spatialanalysiso00romsrich_Page_009.pro
R10 8684
spatialanalysiso00romsrich_Page_010.pro
R11 32043
spatialanalysiso00romsrich_Page_011.pro
R12 37718
spatialanalysiso00romsrich_Page_012.pro
R13 33291
spatialanalysiso00romsrich_Page_013.pro
R14 38774
spatialanalysiso00romsrich_Page_014.pro
R15 36840
spatialanalysiso00romsrich_Page_015.pro
R16 37126
spatialanalysiso00romsrich_Page_016.pro
R17 38598
spatialanalysiso00romsrich_Page_017.pro
R18 34746
spatialanalysiso00romsrich_Page_018.pro
R19 35617
spatialanalysiso00romsrich_Page_019.pro
R20 37494
spatialanalysiso00romsrich_Page_020.pro
R21 38879
spatialanalysiso00romsrich_Page_021.pro
R22 38099
spatialanalysiso00romsrich_Page_022.pro
R23 44162
spatialanalysiso00romsrich_Page_023.pro
R24 36630
spatialanalysiso00romsrich_Page_024.pro
R25 34013
spatialanalysiso00romsrich_Page_025.pro
R26 39553
spatialanalysiso00romsrich_Page_026.pro
R27 39166
spatialanalysiso00romsrich_Page_027.pro
R28 43019
spatialanalysiso00romsrich_Page_028.pro
R29 43867
spatialanalysiso00romsrich_Page_029.pro
R30 39294
spatialanalysiso00romsrich_Page_030.pro
R31 38951
spatialanalysiso00romsrich_Page_031.pro
R32 25180
spatialanalysiso00romsrich_Page_032.pro
R33 27763
spatialanalysiso00romsrich_Page_033.pro
R34 36823
spatialanalysiso00romsrich_Page_034.pro
R35 45792
spatialanalysiso00romsrich_Page_035.pro
R36 7200
spatialanalysiso00romsrich_Page_036.pro
R37 32139
spatialanalysiso00romsrich_Page_037.pro
R38 15524
spatialanalysiso00romsrich_Page_038.pro
R39 35782
spatialanalysiso00romsrich_Page_039.pro
R40 35636
spatialanalysiso00romsrich_Page_040.pro
R41 39989
spatialanalysiso00romsrich_Page_041.pro
R42 36937
spatialanalysiso00romsrich_Page_042.pro
R43 37742
spatialanalysiso00romsrich_Page_043.pro
R44 38326
spatialanalysiso00romsrich_Page_044.pro
R45 36265
spatialanalysiso00romsrich_Page_045.pro
R46 37099
spatialanalysiso00romsrich_Page_046.pro
R47 36686
spatialanalysiso00romsrich_Page_047.pro
R48 31476
spatialanalysiso00romsrich_Page_048.pro
R49 35629
spatialanalysiso00romsrich_Page_049.pro
R50 29853
spatialanalysiso00romsrich_Page_050.pro
R51 46900
spatialanalysiso00romsrich_Page_051.pro
R52 25991
spatialanalysiso00romsrich_Page_052.pro
R53 34844
spatialanalysiso00romsrich_Page_053.pro
R54 24813
spatialanalysiso00romsrich_Page_054.pro
R55 35550
spatialanalysiso00romsrich_Page_055.pro
R56 37583
spatialanalysiso00romsrich_Page_056.pro
R57 28823
spatialanalysiso00romsrich_Page_057.pro
R58 28255
spatialanalysiso00romsrich_Page_058.pro
R59 29466
spatialanalysiso00romsrich_Page_059.pro
R60 36852
spatialanalysiso00romsrich_Page_060.pro
R61 2515
spatialanalysiso00romsrich_Page_061.pro
R62 38370
spatialanalysiso00romsrich_Page_062.pro
R63 44095
spatialanalysiso00romsrich_Page_063.pro
R64 2895
spatialanalysiso00romsrich_Page_064.pro
R65
spatialanalysiso00romsrich_Page_065.pro
R66 37121
spatialanalysiso00romsrich_Page_066.pro
R67 41422
spatialanalysiso00romsrich_Page_067.pro
R68 39755
spatialanalysiso00romsrich_Page_068.pro
R69 45050
spatialanalysiso00romsrich_Page_069.pro
R70 36714
spatialanalysiso00romsrich_Page_070.pro
R71 34572
spatialanalysiso00romsrich_Page_071.pro
R72 37687
spatialanalysiso00romsrich_Page_072.pro
R73 35492
spatialanalysiso00romsrich_Page_073.pro
R74 41935
spatialanalysiso00romsrich_Page_074.pro
R75 2352
spatialanalysiso00romsrich_Page_075.pro
R76 36042
spatialanalysiso00romsrich_Page_076.pro
R77 38044
spatialanalysiso00romsrich_Page_077.pro
R78 3059
spatialanalysiso00romsrich_Page_078.pro
R79 43614
spatialanalysiso00romsrich_Page_079.pro
R80 38737
spatialanalysiso00romsrich_Page_080.pro
R81 45639
spatialanalysiso00romsrich_Page_081.pro
R82 35941
spatialanalysiso00romsrich_Page_082.pro
R83 31861
spatialanalysiso00romsrich_Page_083.pro
R84 39212
spatialanalysiso00romsrich_Page_084.pro
R85 36318
spatialanalysiso00romsrich_Page_085.pro
R86 39938
spatialanalysiso00romsrich_Page_086.pro
R87 37673
spatialanalysiso00romsrich_Page_087.pro
R88 35052
spatialanalysiso00romsrich_Page_088.pro
R89 34127
spatialanalysiso00romsrich_Page_089.pro
R90 32632
spatialanalysiso00romsrich_Page_090.pro
R91 38235
spatialanalysiso00romsrich_Page_091.pro
R92 35495
spatialanalysiso00romsrich_Page_092.pro
R93 42660
spatialanalysiso00romsrich_Page_093.pro
R94 19141
spatialanalysiso00romsrich_Page_094.pro
R95 11028
spatialanalysiso00romsrich_Page_095.pro
R96 28948
spatialanalysiso00romsrich_Page_096.pro
R97 37628
spatialanalysiso00romsrich_Page_097.pro
R98 38592
spatialanalysiso00romsrich_Page_098.pro
R99 37577
spatialanalysiso00romsrich_Page_099.pro
R100 35206
spatialanalysiso00romsrich_Page_100.pro
R101 39329
spatialanalysiso00romsrich_Page_101.pro
R102 38141
spatialanalysiso00romsrich_Page_102.pro
R103
spatialanalysiso00romsrich_Page_103.pro
R104 36622
spatialanalysiso00romsrich_Page_104.pro
R105 40935
spatialanalysiso00romsrich_Page_105.pro
R106 2926
spatialanalysiso00romsrich_Page_106.pro
R107 33537
spatialanalysiso00romsrich_Page_107.pro
R108 43512
spatialanalysiso00romsrich_Page_108.pro
R109 44919
spatialanalysiso00romsrich_Page_109.pro
R110 18747
spatialanalysiso00romsrich_Page_110.pro
R111 25348
spatialanalysiso00romsrich_Page_111.pro
R112 12763
spatialanalysiso00romsrich_Page_112.pro
T1 textplain 499
spatialanalysiso00romsrich_Page_001.txt
T3 105
spatialanalysiso00romsrich_Page_003.txt
T4 541
spatialanalysiso00romsrich_Page_004.txt
T5 1668
spatialanalysiso00romsrich_Page_005.txt
T6 614
spatialanalysiso00romsrich_Page_006.txt
T7 1419
spatialanalysiso00romsrich_Page_007.txt
T8 1791
spatialanalysiso00romsrich_Page_008.txt
T9 1515
spatialanalysiso00romsrich_Page_009.txt
T10 404
spatialanalysiso00romsrich_Page_010.txt
T11 1396
spatialanalysiso00romsrich_Page_011.txt
T12 1520
spatialanalysiso00romsrich_Page_012.txt
T13 1362
spatialanalysiso00romsrich_Page_013.txt
T14 1551
spatialanalysiso00romsrich_Page_014.txt
T15 1534
spatialanalysiso00romsrich_Page_015.txt
T16 1497
spatialanalysiso00romsrich_Page_016.txt
T17 1544
spatialanalysiso00romsrich_Page_017.txt
T18 1522
spatialanalysiso00romsrich_Page_018.txt
T19 1699
spatialanalysiso00romsrich_Page_019.txt
T20 1561
spatialanalysiso00romsrich_Page_020.txt
T21 1563
spatialanalysiso00romsrich_Page_021.txt
T22 1564
spatialanalysiso00romsrich_Page_022.txt
T23 1804
spatialanalysiso00romsrich_Page_023.txt
T24 1470
spatialanalysiso00romsrich_Page_024.txt
T25 1377
spatialanalysiso00romsrich_Page_025.txt
T26 1618
spatialanalysiso00romsrich_Page_026.txt
T27 1583
spatialanalysiso00romsrich_Page_027.txt
T28 1796
spatialanalysiso00romsrich_Page_028.txt
T29 1779
spatialanalysiso00romsrich_Page_029.txt
T30 1578
spatialanalysiso00romsrich_Page_030.txt
T31 1608
spatialanalysiso00romsrich_Page_031.txt
T32 1055
spatialanalysiso00romsrich_Page_032.txt
T33 1237
spatialanalysiso00romsrich_Page_033.txt
T34 1601
spatialanalysiso00romsrich_Page_034.txt
T35 1902
spatialanalysiso00romsrich_Page_035.txt
T36 369
spatialanalysiso00romsrich_Page_036.txt
T37 1381
spatialanalysiso00romsrich_Page_037.txt
T38 742
spatialanalysiso00romsrich_Page_038.txt
T39 1438
spatialanalysiso00romsrich_Page_039.txt
T40 1448
spatialanalysiso00romsrich_Page_040.txt
T41 1589
spatialanalysiso00romsrich_Page_041.txt
T42 1477
spatialanalysiso00romsrich_Page_042.txt
T43
spatialanalysiso00romsrich_Page_043.txt
T44 1558
spatialanalysiso00romsrich_Page_044.txt
T45 1463
spatialanalysiso00romsrich_Page_045.txt
T46 1491
spatialanalysiso00romsrich_Page_046.txt
T47 1464
spatialanalysiso00romsrich_Page_047.txt
T48 1504
spatialanalysiso00romsrich_Page_048.txt
T49 1480
spatialanalysiso00romsrich_Page_049.txt
T50 1218
spatialanalysiso00romsrich_Page_050.txt
T51 1992
spatialanalysiso00romsrich_Page_051.txt
T52 1071
spatialanalysiso00romsrich_Page_052.txt
T53 1488
spatialanalysiso00romsrich_Page_053.txt
T54 1081
spatialanalysiso00romsrich_Page_054.txt
T55 1498
spatialanalysiso00romsrich_Page_055.txt
T56 1501
spatialanalysiso00romsrich_Page_056.txt
T57 1648
spatialanalysiso00romsrich_Page_057.txt
T58 1451
spatialanalysiso00romsrich_Page_058.txt
T59 1265
spatialanalysiso00romsrich_Page_059.txt
T60 1518
spatialanalysiso00romsrich_Page_060.txt
T61 164
spatialanalysiso00romsrich_Page_061.txt
T62 1663
spatialanalysiso00romsrich_Page_062.txt
T63 1881
spatialanalysiso00romsrich_Page_063.txt
T64 173
spatialanalysiso00romsrich_Page_064.txt
T65 1650
spatialanalysiso00romsrich_Page_065.txt
T66 1483
spatialanalysiso00romsrich_Page_066.txt
T67 1711
spatialanalysiso00romsrich_Page_067.txt
T68 1580
spatialanalysiso00romsrich_Page_068.txt
T69 2006
spatialanalysiso00romsrich_Page_069.txt
T70 1569
spatialanalysiso00romsrich_Page_070.txt
T71 2018
spatialanalysiso00romsrich_Page_071.txt
T72 1768
spatialanalysiso00romsrich_Page_072.txt
T73 1584
spatialanalysiso00romsrich_Page_073.txt
T74 1729
spatialanalysiso00romsrich_Page_074.txt
T75
spatialanalysiso00romsrich_Page_075.txt
T76
spatialanalysiso00romsrich_Page_076.txt
T77 1575
spatialanalysiso00romsrich_Page_077.txt
T78 281
spatialanalysiso00romsrich_Page_078.txt
T79 1786
spatialanalysiso00romsrich_Page_079.txt
T80 1653
spatialanalysiso00romsrich_Page_080.txt
T81 1957
spatialanalysiso00romsrich_Page_081.txt
T82 1506
spatialanalysiso00romsrich_Page_082.txt
T83 1413
spatialanalysiso00romsrich_Page_083.txt
T84 1687
spatialanalysiso00romsrich_Page_084.txt
T85
spatialanalysiso00romsrich_Page_085.txt
T86 1614
spatialanalysiso00romsrich_Page_086.txt
T87 1514
spatialanalysiso00romsrich_Page_087.txt
T88
spatialanalysiso00romsrich_Page_088.txt
T89 1500
spatialanalysiso00romsrich_Page_089.txt
T90 1509
spatialanalysiso00romsrich_Page_090.txt
T91 1657
spatialanalysiso00romsrich_Page_091.txt
T92 1604
spatialanalysiso00romsrich_Page_092.txt
T93 1693
spatialanalysiso00romsrich_Page_093.txt
T94 785
spatialanalysiso00romsrich_Page_094.txt
T95 498
spatialanalysiso00romsrich_Page_095.txt
T96 1225
spatialanalysiso00romsrich_Page_096.txt
T97 1530
spatialanalysiso00romsrich_Page_097.txt
T98 1554
spatialanalysiso00romsrich_Page_098.txt
T99 1516
spatialanalysiso00romsrich_Page_099.txt
T100 1434
spatialanalysiso00romsrich_Page_100.txt
T101 1599
spatialanalysiso00romsrich_Page_101.txt
T102 1536
spatialanalysiso00romsrich_Page_102.txt
T103
spatialanalysiso00romsrich_Page_103.txt
T104 1482
spatialanalysiso00romsrich_Page_104.txt
T105 1642
spatialanalysiso00romsrich_Page_105.txt
T106 174
spatialanalysiso00romsrich_Page_106.txt
T107 1511
spatialanalysiso00romsrich_Page_107.txt
T108 1852
spatialanalysiso00romsrich_Page_108.txt
T109 1962
spatialanalysiso00romsrich_Page_109.txt
T110 870
spatialanalysiso00romsrich_Page_110.txt
T111 1046
spatialanalysiso00romsrich_Page_111.txt
T112
spatialanalysiso00romsrich_Page_112.txt
UR1 3900
spatialanalysiso00romsrich_Page_001thm.jpg
applicationpdf 5032191
spatialanalysiso00romsrich.pdf
AR1 13705
spatialanalysiso00romsrich_Page_001.QC.jpg
AR2 7037
spatialanalysiso00romsrich_Page_002.QC.jpg
AR3 2287
spatialanalysiso00romsrich_Page_002thm.jpg
AR4 5335
spatialanalysiso00romsrich_Page_003.QC.jpg
AR5 1822
spatialanalysiso00romsrich_Page_003thm.jpg
AR6 15917
spatialanalysiso00romsrich_Page_004.QC.jpg
AR7 4260
spatialanalysiso00romsrich_Page_004thm.jpg
AR8 31256
spatialanalysiso00romsrich_Page_005.QC.jpg
AR9 7289
spatialanalysiso00romsrich_Page_005thm.jpg
AR10 13940
spatialanalysiso00romsrich_Page_006.QC.jpg
AR11 3702
spatialanalysiso00romsrich_Page_006thm.jpg
AR12 28374
spatialanalysiso00romsrich_Page_007.QC.jpg
AR13 7438
spatialanalysiso00romsrich_Page_007thm.jpg
AR14 33075
spatialanalysiso00romsrich_Page_008.QC.jpg
AR15 8818
spatialanalysiso00romsrich_Page_008thm.jpg
AR16 31211
spatialanalysiso00romsrich_Page_009.QC.jpg
AR17 8218
spatialanalysiso00romsrich_Page_009thm.jpg
AR18 11811
spatialanalysiso00romsrich_Page_010.QC.jpg
AR19 3759
spatialanalysiso00romsrich_Page_010thm.jpg
AR20 34652
spatialanalysiso00romsrich_Page_011.QC.jpg
AR21 8013
spatialanalysiso00romsrich_Page_011thm.jpg
AR22 40640
spatialanalysiso00romsrich_Page_012.QC.jpg
AR23 10050
spatialanalysiso00romsrich_Page_012thm.jpg
AR24 36583
spatialanalysiso00romsrich_Page_013.QC.jpg
AR25 8664
spatialanalysiso00romsrich_Page_013thm.jpg
AR26 41681
spatialanalysiso00romsrich_Page_014.QC.jpg
AR27 9763
spatialanalysiso00romsrich_Page_014thm.jpg
AR28 38172
spatialanalysiso00romsrich_Page_015.QC.jpg
AR29 8763
spatialanalysiso00romsrich_Page_015thm.jpg
AR30 39315
spatialanalysiso00romsrich_Page_016.QC.jpg
AR31 9084
spatialanalysiso00romsrich_Page_016thm.jpg
AR32 40812
spatialanalysiso00romsrich_Page_017.QC.jpg
AR33 9624
spatialanalysiso00romsrich_Page_017thm.jpg
AR34 34531
spatialanalysiso00romsrich_Page_018.QC.jpg
AR35 8536
spatialanalysiso00romsrich_Page_018thm.jpg
AR36 33968
spatialanalysiso00romsrich_Page_019.QC.jpg
AR37 8120
spatialanalysiso00romsrich_Page_019thm.jpg
AR38 39664
spatialanalysiso00romsrich_Page_020.QC.jpg
AR39 8965
spatialanalysiso00romsrich_Page_020thm.jpg
AR40 39222
spatialanalysiso00romsrich_Page_021.QC.jpg
AR41 9163
spatialanalysiso00romsrich_Page_021thm.jpg
AR42 40855
spatialanalysiso00romsrich_Page_022.QC.jpg
AR43 9706
spatialanalysiso00romsrich_Page_022thm.jpg
AR44 40452
spatialanalysiso00romsrich_Page_023.QC.jpg
AR45 9234
spatialanalysiso00romsrich_Page_023thm.jpg
AR46 39084
spatialanalysiso00romsrich_Page_024.QC.jpg
AR47 8657
spatialanalysiso00romsrich_Page_024thm.jpg
AR48 36976
spatialanalysiso00romsrich_Page_025.QC.jpg
AR49 8549
spatialanalysiso00romsrich_Page_025thm.jpg
AR50 41358
spatialanalysiso00romsrich_Page_026.QC.jpg
AR51 9563
spatialanalysiso00romsrich_Page_026thm.jpg
AR52 39589
spatialanalysiso00romsrich_Page_027.QC.jpg
AR53 8997
spatialanalysiso00romsrich_Page_027thm.jpg
AR54 40224
spatialanalysiso00romsrich_Page_028.QC.jpg
AR55 9108
spatialanalysiso00romsrich_Page_028thm.jpg
AR56 40722
spatialanalysiso00romsrich_Page_029.QC.jpg
AR57 9175
spatialanalysiso00romsrich_Page_029thm.jpg
AR58 41723
spatialanalysiso00romsrich_Page_030.QC.jpg
AR59 9381
spatialanalysiso00romsrich_Page_030thm.jpg
AR60 41457
spatialanalysiso00romsrich_Page_031.QC.jpg
AR61 9511
spatialanalysiso00romsrich_Page_031thm.jpg
AR62 28310
spatialanalysiso00romsrich_Page_032.QC.jpg
AR63 6809
spatialanalysiso00romsrich_Page_032thm.jpg
AR64 26758
spatialanalysiso00romsrich_Page_033.QC.jpg
AR65 6905
spatialanalysiso00romsrich_Page_033thm.jpg
AR66 32221
spatialanalysiso00romsrich_Page_034.QC.jpg
AR67 7726
spatialanalysiso00romsrich_Page_034thm.jpg
AR68 39438
spatialanalysiso00romsrich_Page_035.QC.jpg
AR69 9490
spatialanalysiso00romsrich_Page_035thm.jpg
AR70 9780
spatialanalysiso00romsrich_Page_036.QC.jpg
AR71 3062
spatialanalysiso00romsrich_Page_036thm.jpg
AR72 36134
spatialanalysiso00romsrich_Page_037.QC.jpg
AR73 8434
spatialanalysiso00romsrich_Page_037thm.jpg
AR74 28524
spatialanalysiso00romsrich_Page_038.QC.jpg
AR75 7587
spatialanalysiso00romsrich_Page_038thm.jpg
AR76 37056
spatialanalysiso00romsrich_Page_039.QC.jpg
AR77 8732
spatialanalysiso00romsrich_Page_039thm.jpg
AR78 38554
spatialanalysiso00romsrich_Page_040.QC.jpg
AR79 9450
spatialanalysiso00romsrich_Page_040thm.jpg
AR80 41618
spatialanalysiso00romsrich_Page_041.QC.jpg
AR81 9232
spatialanalysiso00romsrich_Page_041thm.jpg
AR82 39680
spatialanalysiso00romsrich_Page_042.QC.jpg
AR83 9214
spatialanalysiso00romsrich_Page_042thm.jpg
AR84 40400
spatialanalysiso00romsrich_Page_043.QC.jpg
AR85 9330
spatialanalysiso00romsrich_Page_043thm.jpg
AR86 41868
spatialanalysiso00romsrich_Page_044.QC.jpg
AR87 9564
spatialanalysiso00romsrich_Page_044thm.jpg
AR88 40275
spatialanalysiso00romsrich_Page_045.QC.jpg
AR89 9111
spatialanalysiso00romsrich_Page_045thm.jpg
AR90 40700
spatialanalysiso00romsrich_Page_046.QC.jpg
AR91 9606
spatialanalysiso00romsrich_Page_046thm.jpg
AR92 38543
spatialanalysiso00romsrich_Page_047.QC.jpg
AR93 9076
spatialanalysiso00romsrich_Page_047thm.jpg
AR94 33148
spatialanalysiso00romsrich_Page_048.QC.jpg
AR95 8103
spatialanalysiso00romsrich_Page_048thm.jpg
AR96 36861
spatialanalysiso00romsrich_Page_049.QC.jpg
AR97 8784
spatialanalysiso00romsrich_Page_049thm.jpg
AR98 32742
spatialanalysiso00romsrich_Page_050.QC.jpg
AR99 8116
spatialanalysiso00romsrich_Page_050thm.jpg
AR100 39999
spatialanalysiso00romsrich_Page_051.QC.jpg
AR101 8935
spatialanalysiso00romsrich_Page_051thm.jpg
AR102 28267
spatialanalysiso00romsrich_Page_052.QC.jpg
AR103 7052
spatialanalysiso00romsrich_Page_052thm.jpg
AR104 29072
spatialanalysiso00romsrich_Page_053.QC.jpg
AR105 7285
spatialanalysiso00romsrich_Page_053thm.jpg
AR106 24242
spatialanalysiso00romsrich_Page_054.QC.jpg
AR107 5715
spatialanalysiso00romsrich_Page_054thm.jpg
AR108 34789
spatialanalysiso00romsrich_Page_055.QC.jpg
AR109 8217
spatialanalysiso00romsrich_Page_055thm.jpg
AR110 39662
spatialanalysiso00romsrich_Page_056.QC.jpg
AR111 9154
spatialanalysiso00romsrich_Page_056thm.jpg
AR112 20824
spatialanalysiso00romsrich_Page_057.QC.jpg
AR113 5857
spatialanalysiso00romsrich_Page_057thm.jpg
AR114 28989
spatialanalysiso00romsrich_Page_058.QC.jpg
AR115 6895
spatialanalysiso00romsrich_Page_058thm.jpg
AR116 29992
spatialanalysiso00romsrich_Page_059.QC.jpg
AR117 7237
spatialanalysiso00romsrich_Page_059thm.jpg
AR118 32999
spatialanalysiso00romsrich_Page_060.QC.jpg
AR119 7475
spatialanalysiso00romsrich_Page_060thm.jpg
AR120 23139
spatialanalysiso00romsrich_Page_061.QC.jpg
AR121 7429
spatialanalysiso00romsrich_Page_061thm.jpg
AR122 32372
spatialanalysiso00romsrich_Page_062.QC.jpg
AR123 7828
spatialanalysiso00romsrich_Page_062thm.jpg
AR124 37523
spatialanalysiso00romsrich_Page_063.QC.jpg
AR125 8806
spatialanalysiso00romsrich_Page_063thm.jpg
AR126 27271
spatialanalysiso00romsrich_Page_064.QC.jpg
AR127 8179
spatialanalysiso00romsrich_Page_064thm.jpg
AR128 32889
spatialanalysiso00romsrich_Page_065.QC.jpg
AR129 8269
spatialanalysiso00romsrich_Page_065thm.jpg
AR130 38322
spatialanalysiso00romsrich_Page_066.QC.jpg
AR131 8830
spatialanalysiso00romsrich_Page_066thm.jpg
AR132 33144
spatialanalysiso00romsrich_Page_067.QC.jpg
AR133 8057
spatialanalysiso00romsrich_Page_067thm.jpg
AR134 39883
spatialanalysiso00romsrich_Page_068.QC.jpg
AR135 9362
spatialanalysiso00romsrich_Page_068thm.jpg
AR136 35037
spatialanalysiso00romsrich_Page_069.QC.jpg
AR137 8444
spatialanalysiso00romsrich_Page_069thm.jpg
AR138 30101
spatialanalysiso00romsrich_Page_070.QC.jpg
AR139 7281
spatialanalysiso00romsrich_Page_070thm.jpg
AR140 24222
spatialanalysiso00romsrich_Page_071.QC.jpg
AR141 6297
spatialanalysiso00romsrich_Page_071thm.jpg
AR142 31399
spatialanalysiso00romsrich_Page_072.QC.jpg
AR143 7075
spatialanalysiso00romsrich_Page_072thm.jpg
AR144 32134
spatialanalysiso00romsrich_Page_073.QC.jpg
AR145 7759
spatialanalysiso00romsrich_Page_073thm.jpg
AR146 36404
spatialanalysiso00romsrich_Page_074.QC.jpg
AR147 8254
spatialanalysiso00romsrich_Page_074thm.jpg
AR148 22113
spatialanalysiso00romsrich_Page_075.QC.jpg
AR149 7220
spatialanalysiso00romsrich_Page_075thm.jpg
AR150 33901
spatialanalysiso00romsrich_Page_076.QC.jpg
AR151 8160
spatialanalysiso00romsrich_Page_076thm.jpg
AR152 30663
spatialanalysiso00romsrich_Page_077.QC.jpg
AR153 7174
spatialanalysiso00romsrich_Page_077thm.jpg
AR154 28994
spatialanalysiso00romsrich_Page_078.QC.jpg
AR155 8266
spatialanalysiso00romsrich_Page_078thm.jpg
AR156 37681
spatialanalysiso00romsrich_Page_079.QC.jpg
AR157 8939
spatialanalysiso00romsrich_Page_079thm.jpg
AR158 35169
spatialanalysiso00romsrich_Page_080.QC.jpg
AR159 7973
spatialanalysiso00romsrich_Page_080thm.jpg
AR160 39014
spatialanalysiso00romsrich_Page_081.QC.jpg
AR161 8608
spatialanalysiso00romsrich_Page_081thm.jpg
AR162 30871
spatialanalysiso00romsrich_Page_082.QC.jpg
AR163 7448
spatialanalysiso00romsrich_Page_082thm.jpg
AR164 28741
spatialanalysiso00romsrich_Page_083.QC.jpg
AR165 7171
spatialanalysiso00romsrich_Page_083thm.jpg
AR166 31924
spatialanalysiso00romsrich_Page_084.QC.jpg
AR167 7822
spatialanalysiso00romsrich_Page_084thm.jpg
AR168 35767
spatialanalysiso00romsrich_Page_085.QC.jpg
AR169 8198
spatialanalysiso00romsrich_Page_085thm.jpg
AR170 40966
spatialanalysiso00romsrich_Page_086.QC.jpg
AR171 9183
spatialanalysiso00romsrich_Page_086thm.jpg
AR172 39371
spatialanalysiso00romsrich_Page_087.QC.jpg
AR173 9310
spatialanalysiso00romsrich_Page_087thm.jpg
AR174 35006
spatialanalysiso00romsrich_Page_088.QC.jpg
AR175 8530
spatialanalysiso00romsrich_Page_088thm.jpg
AR176 31522
spatialanalysiso00romsrich_Page_089.QC.jpg
AR177 7805
spatialanalysiso00romsrich_Page_089thm.jpg
AR178 30974
spatialanalysiso00romsrich_Page_090.QC.jpg
AR179 7249
spatialanalysiso00romsrich_Page_090thm.jpg
AR180 31859
spatialanalysiso00romsrich_Page_091.QC.jpg
AR181 7775
spatialanalysiso00romsrich_Page_091thm.jpg
AR182 30017
spatialanalysiso00romsrich_Page_092.QC.jpg
AR183 7764
spatialanalysiso00romsrich_Page_092thm.jpg
AR184 41260
spatialanalysiso00romsrich_Page_093.QC.jpg
AR185 9355
spatialanalysiso00romsrich_Page_093thm.jpg
AR186 22300
spatialanalysiso00romsrich_Page_094.QC.jpg
AR187 5433
spatialanalysiso00romsrich_Page_094thm.jpg
AR188 12343
spatialanalysiso00romsrich_Page_095.QC.jpg
AR189 3378
spatialanalysiso00romsrich_Page_095thm.jpg
AR190 29275
spatialanalysiso00romsrich_Page_096.QC.jpg
AR191 6741
spatialanalysiso00romsrich_Page_096thm.jpg
AR192 37233
spatialanalysiso00romsrich_Page_097.QC.jpg
AR193 8607
spatialanalysiso00romsrich_Page_097thm.jpg
AR194 37731
spatialanalysiso00romsrich_Page_098.QC.jpg
AR195 8406
spatialanalysiso00romsrich_Page_098thm.jpg
AR196 38220
spatialanalysiso00romsrich_Page_099.QC.jpg
AR197 8532
spatialanalysiso00romsrich_Page_099thm.jpg
AR198 34834
spatialanalysiso00romsrich_Page_100.QC.jpg
AR199 8251
spatialanalysiso00romsrich_Page_100thm.jpg
AR200 38108
spatialanalysiso00romsrich_Page_101.QC.jpg
AR201 8787
spatialanalysiso00romsrich_Page_101thm.jpg
AR202 37935
spatialanalysiso00romsrich_Page_102.QC.jpg
AR203 8496
spatialanalysiso00romsrich_Page_102thm.jpg
AR204 36255
spatialanalysiso00romsrich_Page_103.QC.jpg
AR205 8344
spatialanalysiso00romsrich_Page_103thm.jpg
AR206 35507
spatialanalysiso00romsrich_Page_104.QC.jpg
AR207 7795
spatialanalysiso00romsrich_Page_104thm.jpg
AR208 39638
spatialanalysiso00romsrich_Page_105.QC.jpg
AR209
spatialanalysiso00romsrich_Page_105thm.jpg
AR210 6329
spatialanalysiso00romsrich_Page_106.QC.jpg
AR211 2108
spatialanalysiso00romsrich_Page_106thm.jpg
AR212 30334
spatialanalysiso00romsrich_Page_107.QC.jpg
AR213 7261
spatialanalysiso00romsrich_Page_107thm.jpg
AR214 36577
spatialanalysiso00romsrich_Page_108.QC.jpg
AR215 8623
spatialanalysiso00romsrich_Page_108thm.jpg
AR216 38630
spatialanalysiso00romsrich_Page_109.QC.jpg
AR217 8953
spatialanalysiso00romsrich_Page_109thm.jpg
AR218 16915
spatialanalysiso00romsrich_Page_110.QC.jpg
AR219 4275
spatialanalysiso00romsrich_Page_110thm.jpg
AR220 27347
spatialanalysiso00romsrich_Page_111.QC.jpg
AR221 6408
spatialanalysiso00romsrich_Page_111thm.jpg
AR222 21325
spatialanalysiso00romsrich_Page_112.QC.jpg
AR223 5611
spatialanalysiso00romsrich_Page_112thm.jpg
AR224 142416
UF00097777_00001.mets
METS:structMap STRUCT1 mixed
METS:div DMDID ORDER 0 main
D1 1 Main
P1 Page i
METS:fptr
P2 i-a 2
P3 ii 3
P4 iii 4
P5 iv 5
P6 v 6
P7 vi 7
P8 vii 8
P9 viii 9
P10 ix 10
P11 11
P12 12
P13 13
P14 14
P15 15
P16 16
P17 17
P18 18
P19 19
P20 20
P21 21
P22 22
P23 23
P24 24
P25 25
P26 26
P27 27
P28 28
P29 29
P30 30
P31 31
P32 32
P33 33
P34 34
P35 35
P36 36
P37 37
P38 38
P39 39
P40 40
P41 41
P42 42
P43 43
P44 44
P45 45
P46 46
P47 47
P48 48
P49 49
P50 50
P51 51
P52 52
P53 53
P54 54
P55 55
P56 56
P57 57
P58 58
P59 59
P60 60
P61 61
P62 62
P63 63
P64 64
P65 65
P66 66
P67 67
P68 68
P69 69
P70 70
P71 71
P72 72
P73 73
P74 74
P75 75
P76 76
P77 77
P78 78
P79 79
P80 80
P81 81
P82 82
P83 83
P84 84
P85 85
P86 86
P87 87
P88 88
P89 89
P90 90
P91 91
P92 92
P93 93
P94 94
P95 95
P96 96
P97 97
P98 98
P99 99
P100 100
P101 101
P102 102
P103 103
P104 104
P105
P106 106
P107 107
P108 108
P109 109
P110 110
P111 111
P112 112
METS:behaviorSec VIEWS Options available to user for viewing this item
METS:behavior VIEW1 STRUCTID Default View
METS:mechanism Viewer JPEGs Procedure xlink:type simple xlink:title JPEG_Viewer()
VIEW2 Alternate
zoomable JPEG2000s JP2_Viewer()
VIEW3
Related image viewer shows thumbnails each Related_Image_Viewer()
INTERFACES Banners or interfaces which resource can appear under
INT1 Interface
UFDC_Interface_Loader


A SPATIAL ANALYSIS OF THE DIMENSIONS OF
ECONOMIC HEALTH IN THE SOUTHEASTERN
UNITED STATES (1950 AND I960)
By
GERALD HARRY ROMSA
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
U OF F LIBRARIES
UNIVERSITY OF FLORIDA
1969

UNIVERSITY OF FLORIDA
3 1262 08552 2190

To my parents and my wife
Mary

ACKNOWLEDGMENTS
I wish to express my gratitude to my advisor Dr.
Clark I. Cross and the other members of my committee, Dr.
Stanley D. Brunn, Dr. James R. Anderson and Dr. John E.
Reynolds for their guidance, thoughts, constructive criti¬
cisms, and patience during the preparation of this disserta¬
tion.
In addition, I want to acknowledge the University
of Florida Computer Center's facilities and services which
were utilized for the data analysis in this study.
iii

TABLE OP CONTENTS
Page
ACKNOWLEDGMENTS .iii
LIST OF TABLES vi
LIST OF FIGURES ix
CHAPTER
I THE CONCEPT OF POVERTY . 1
Introduction 1
Objective 3
Need for Such a Study 3
Poverty in the United States 5
Some Approaches to the Study of Poverty . . 10
Review of the Literature 15
The Problem 21
II STUDY AREA AND METHODOLOGY . 27
The Setting .29
Historical Background 30
Recent Trends in the Economy 35
Outline of the Analytical Model 36
The Model 37
Variables Used 40
Summary . 42
IIITHE DIMENSIONS OF ECONOMIC HEALTH IN THE
SOUTHEASTERN UNITED STATES (1950-1960) .... 45
1950 Factor Analysis ..... 46
The 1950 Dimensions 48
I960 Analysis 60
The I960 Dimensions 62
Comparison of the Two Time Periods 75
Regression Analysis 78
The Hypotheses 79
Summary 84
IV

CHAPTER
Page
IV SUMMARY AND CONCLUSIONS 86
1950 Analysis 87
I960 Analysis 89
Comparison of Changes in the Two Time
Periods Studied 92
Conclusions. 93
Future Research Efforts 94
BIBLIOGRAPHY 97
BIOGRAPHICAL SKETCH 101
v

LIST OF TABLES
Table Page
1 Per Cent of Families Earning Less than
S3,000 a Year by Regions (I960) 8
2 Distribution of Poverty by Residence (I960) . 9
3 Median Income by Residence (i960) 9
4 Variables Used 41
5 Percentage of Total Variance Explained by
Each Dimension in 1990 47
6 Percentage of Variance (Commonality) of
Each of the 26 Variables Accounted for
by All Six Components in 1950 47
7 Dimension I Urban Industrial Counties
in 1950 49
8 Counties with Extreme Factor Scores on
Dimension I in 1950 50
9 Dimension II Tenant and Small Farm
Poverty in 1950 52
10 Counties with Extreme Factor Scores on
Dimension II in 1950 52
11 Dimension III Wealthy Urbanized Counties
in 1950 55
12 Comities with Extreme Factor Scores on
Dimension III in 1950 55
13 Dimension IV Rural White Counties in 1950 . 57
vi

Table
Page
14 Counties with Extreme Factor Scores on
Dimension IV in 1950 57
15 Dimension V Locational Advantage in
Relation to National Market (New York)
in 1950 59
16 Counties with Extreme Factor Scores on
Dimension V in 1950 59
17 Dimension VI Locational Advantage in
Relation to Local, Regional and National
Markets in 1950 60
18 Counties with Extreme Scores on Dimension
VI in 1950 60
19 Percentage of Total Variance Explained by
Each Dimension in I960 61
20 Percentage of Variance (Communality) of
Each of the 26 Variables Accounted for
by All Seven Components in I960 61
21 Dimension I Urban Industrial Counties
in I960 63
22 Counties with Extreme Factor Scores on
Dimension I in I960 64
23 Dimension II Rural Negro Counties 66
24 Counties with Extreme Factor Scores on
Dimension II in I960 67
25 Dimension III Tenant and Small Farm
Poverty in I960 67
26 Counties with Extreme Factor Scores on
Dimension III in I960 69
27 Dimension IV Locational Advantage in
Relation to National Market (New York)
in I960 70
vii

Table Page
28 Counties with Extreme Factor Scores on
Dimension IV in I960 71
29 Dimension V Wealthy Urbanized Counties
in I960 72
30 Counties with Extreme Factor Scores on
Dimension V in I960 72
31 Dimension VI Locational Advantage in
Relation to National, Regional and
Local Markets in I960 73
32 Counties with Extreme Factor Scores on
Dimension VI in I960 74
33 Dimension VII Areas of Declining Prosperity
in I960 74
34 Counties with Extreme Factor Scores on
Dimension VII in I960 74
35 Expected Correlations Between Median
Family Income and the Derived Dimensions. . 79
36 Normality of 1950 Factor Scores 80
37 1950 Correlation Matrix 80
38 1950 Summary Table 81
39 Normality of I960 Factor Scores 81
40 I960 Correlation Matrix 81
41 I960 Summary Table 82
42 Comparison of the Beta, Correlation
Coefficients and the Coefficient of
Determination Between the Two Time Periods 82
viii

LIST OF FIGURES
Figure Page
1 Study Area; the Southeastern Core 28
2 Urban Industrial Counties, 1950 51
3 Tenant and Small Farm Poverty, 1950 54
4 Urban Industrial Counties, I960 65
5 Tenant and Small Farm Poverty, I960 68
IX

CHAPTER 1
THE CONCEPT OP POVERTY
Introduction
The increasing attention given by geographers to the
problems of poverty, economic health and underdeveloped
areas is a healthy sign that they are progressing with the
temporal needs of today’s society. This concern with areal
differentiation in economic growth and levels of economic
wellbeing is not a new geography. It is, instead, a more
specialized topical outlook on the net result of economic
activity as viewed in a spatial framework.1 This approach
attempts to relate regional inequalities in levels of liv¬
ing with man's economic endeavours as they are tied to his
cultural, technological and resource environment.
The trend to look for factors other than resource
endowment is perhaps the major criterion which separates
present geographical research from previous regional stud¬
ies. Ullman has pointed out that natural resources and cul¬
tural differences are not sufficient to explain the dispar-
2
ity in economic growth. Differences in social conditions
and advances in technology often provide additional informa¬
tion in explaining the areal variation in prosperity. The
inclusion of additional variables to explain this
1

2
differential in economic browth has led to the use of multi¬
variate statistical techniques as analytical tools. The
usefulness of these techniques lies in their applicability
in the analysis of large numbers of indices; in their isola¬
tion of significant variables and in the testing of hypoth¬
eses. These are important contributions when dealing with
interrelated data.
The realization by social scientists that the dis¬
tribution of wealth and poverty is uneven has focused their
attention on the need for more regional studies. National
and state data often tend to hide small pockets of high
deviation in levels of income. This has led to economic
regionalization based on counties, and, where possible,
smaller units.
Spatial interaction is another geographic concept
which is increasingly being realized as playing an important
role in explaining the difference in levels of progress.
This concept was introduced by Ullman and refined by Eager-
strand at Lund University. Hagerstrand used simulation
models in an attempt to measure the degree of resistance
that cultural and physical barriers imposed upon the accept¬
ance of new ideas among contiguous areas in Sweden.^
Hartshorne points out that the comparison of eco¬
nomic growth among underdeveloped regions should be rela-
4
tive rather than concrete. A direct comparison between two
culturally different regions may not show the disparity in

3
economic achievement accurately unless the indices utilized
are standardized to represent with less bias the contrast¬
ing cultural and political values. This also is true of
temporal studies. Although this problem at present has not
been resolved, studies within a similar cultural and polit¬
ical unit based upon certain multivariate techniques are
feasible.
Objective
The purpose of this study is to identify the salient
socio-economic dimensions or characteristics of five south¬
eastern states (Mississippi, Alabama, Georgia, South
Carolina and North Carolina) for two time periods, 1950
and I960. These characteristics will be examined to ascer¬
tain if they reflect any areal patterns or changes through
time. If areal variations do exist, they will be analyzed
in regard to their stability, location, effects of urbaniza¬
tion and the importance of push and pull factors.
In addition the extracted dimensions will be tested
individually and as a group for their validity in explain¬
ing the high deviation in median income throughout the
study area.
Need for Such a Study
The concern of some Americans for a more equitable
distribution of this nation's v/ealth has shown itself
through civil legislation and action programs by state,

4
local and federal agencies. Programs such as the Area
Redevelopment Act and the Appalachia Act are examples of the
federal government’s increased participation. To implement
these programs rapidly and to the best advantage requires
a thorough knowledge of the problems. Thus research is
essential if the complex facets of poverty are to be under¬
stood. An example of such a research program in the South
is The Southern Regional Research Project S-44, "Factors in
the Adjustment of Families and Individuals in Low Income
Areas of the South." This is a cooperative study done by
rural sociologists in nine southern states. These studies
were done in selected counties and dealt primarily with the
attitudes of the rural poor.
There is a need, however, for a comprehensive study
of the Southeast which deals with the facets associated with
poverty on a larger scale. The Southeast has often been
classified as a retarded area. Recent studies have sug¬
gested that this blanket statement can not be applied to
the entire region, as it has made some great economic strides
in the past few decades. At present, there is a lack of
geographic research on this region employing multivariate
statistical techniques, which has emphasized the degree of
variation in economic growth through space and time and
classified economic regions accordingly.
In the remainder of this chapter an overview of the
distribution of poverty in the United States and some

approaches to its study are discussed. The pertinent lit¬
erature is reviewed as a basic for formulating hypotheses.
The analytical model used in this study is developed in
Chapter II and a brief description of the study area is
given. Chapter III deals with the analysis of the data,
while in ChapterIV the conclusions reached are examined
in view of the hypotheses presented in Chapter I.
Poverty in the United States
Amid the highest standard of living ever achieved
by the United States there exists a forgotten minority of
poverty stricken people. The problems of the poor are often
masked by the middle class way of life in America. There is
no room in this society for the economically less fortunate.
The middle class viewpoint is well expressed in The Affluent
5
Society by John Galbraith. Galbraith's major theme was how,
with increasing automation, the good life could be continued
for all, with less work and more leisure. Poverty was per¬
ceived only in two forms: case poverty due to physical
and mental handicaps, and insular poverty such as is found
in West Virginia.^
More recent articles have shown that the problems of
poverty are not isolated but nationally widespread. There
are, however, large concentrations of the poor in certain
geographic areas such as the South and in the ghettos of
urban centers. Depending upon the criteria used, there are

6
estimated to be from 36 to 50 million economically deprived
7
persons within this wealthy nation.
This staggering figure is made even more realistic
if one considers the economic, social and psychological
costs of maintaining these individuals in their present way
of life. Not only does the Gross National Product suffer
from the loss of such a large number of human resources,
but there are also heavy social costs. These include wel¬
fare payments, costs of mental and physical health, crime
and juvenile delinquency. The national cost of welfare pro-
g
grams has been estimated at 13 to 15 billion dollars.
Poverty was first recognized as a social problem in
9
the United States approximately 100 years ago. It existed
before but had been regarded as an individual rather than a
social problem. Prior to this period, individuals attempted
to attain economic freedom by leaving the urban ghettos or
the worn out soils of the East and moving westward. This
movement could only be sustained in a highly agrarian soc¬
iety. The slaves on the southern plantations however did
not have this escape valve.
Increasing industrialization and urban growth after
the Civil War brought with it many social evils: poor tene¬
ment housing, inadequate schools, low wages, bad working
conditions and no job security. Henry George in his Pro¬
gress and Poverty referred to the United States of 1869 as

7
a land where "amid the greatest accumulation of wealth, men
die of starvation and puny infants suckle dry breasts."'*'^
The plight of the working classes, many of whom were immi¬
grants, eventually led to some social reforms. Community-
minded citizens led the battle for public housing regula¬
tions, public health, and abolition of child labour. Labour
unions played a major role in obtaining higher wages and
better working conditions.
Major Federal government action on poverty was ini¬
tiated by President Franklin D. Roosevelt in the 1930's.
In the State of the Union message in 1933 the former Presi¬
dent referred to the "one third of the nation that is ill-
clothed, ill-housed and ill-nourished*"^' It was during
this period that the familiar antipoverty tools: social
security, old age assistance, aid to dependent children,
and unemployment compensation were conceived. Furthermore
the Tennessee Valley Authority Act was a recognition that
there were regional inequalities in economic growth, which
resulted in pockets of distressed areas that were resistant
to change and progress.
The next major Federal attack on poverty commenced
in the 1960's. The emphasis was on legislation. The Area
Redevelopment A.ct of 1962, the Economic Opportunity Act and
the Appalachian Regional Development Act of 1964 are a few
examples. The passage of the acts emphasized rehabilitation
find not monetary relief. They also recogiized the areal

8
distributions of poverty and the need to focus upon the
problems of regional or subregional nature which can not
be solved on the national level. The new progress in
legislation was initiated by a small educated white minor¬
ity and a group of dedicated Negro leaders. Civil Rights
marches, sit-ins, and riots have forced the nation's atten¬
tion to the problems of her minority poor when other less
peaceful efforts failed.^
The geographic distribution of poverty can readily
be seen if the per cent of families earning less than
S3,000 a year is examined by the major divisions of the
country (Table 1).
Table 1
Per cent of Families Earning Less Than
S3, 000 a
Year by Regions (1900)
Region
Total
White
Nonwhite
Northeast
9.8
8.4
26.6
North Central
11.5
10.3
29.7
South
24.6
17.9
58.3
We st
11.7
11.0
20.7
Source:
Louis A. Ferman, Joyce L.
Kornbluh and
Alan Haber
(eds„), Poverty in America
(Ann Arbor:
University
of Michigan
Press, 1965),
p. 72.
This table shows the
enormous variation in income across
America
. It further
points out the
plight of the non-
whites.
Although the
urban centers
harbour the
majority
of the poor, the rural areas have a higher percentage of
poor families (Tables 2 and 3).

9
Table 2
Distribution of Poverty by Residence (i960)
Residence $ of population $ of poor (families
earning less than
$3,000 per year)
Urban 71 54
Rural nonfarm 22 30
Farm 7 16
Source: Louis A. Ferman, Joyce L. Kornbluh and Alan Haber
(eds.)> Poverty in America (Ann Arbor: University
of Michigan Press, 1965), p. 104.
Table 3
Median Income by Residence (i960)
Residence Total Population White
Urban $6,580 $6,678
Rural Nonfarm $5,486 $5,549
Rural Farm $3,779 $3,863
Nonwhite
$4,469
$2,645
$1,323
Source: Louis A. Ferman, Joyce L. Kornbluh and Alan Haber
(eds.), Poverty in America (Ann Arbor: University
of Michigan Press, 1965TÍ p. 107.
In comparing the income between rural and urban
residents, one should consider the difference in cost of
living. Nevertheless the differences in income are greater
than the variation in the cost of living. From these
figures a picture arises that reveals small concentrated
areas of poverty within cities and a more dispersed form of
rural poor. Within the rural sections of the country, there
are extreme variations in the concentrations of the poor.
Examples of areas v/ith a high rate of unemployment and a
concentrated number of low income groups are: Appalachia,

10
the cutover regions of the Great Lakes, northern New
England, the northern Rockies and the Indian reservations.^
Other regions containing large but less concentrated numbers
of poor families are parts of the northern Great Plains,
scattered areas in the West and Midwest and extensive sec¬
tions of the Piedmont and the Atlantic Coastal Plain.
This pattern of dispersion not only shows the
national scale of the problem, but also the variations
from the industrial North to the rural West and the South.
The distribution of natural resources can not account for
all of the variation in poverty. Part of this is due to
demographic factors. Minority nonwhite groups are con¬
centrated in many of the distressed economic sections.
Some depressed regions have a high percentage of unskilled
labour, while many of the predominantly white rural coun¬
ties have high concentrations of older people. The latter
is the result of the out-migration of the young to more
attractive economic areas. Continuous out-migration over
a period of time, of the highly motivated segment of the
population, will tend to economically depress an area. This
was illustrated and substantiated by Spinelli in his study
] t3
of the net out-migration in southeastern Ohio, 1950-1960.
Some Approaches to the Study of Poverty
One of the most difficult problems in studying pov¬
erty is its definition. What constitutes poverty? Many

11
popular definitions adopted are based upon incomes. Sta¬
tistics on family and individual income are readily avail¬
able and accurate. The present S3,000 criterion used for
families and the -$1,500 for single individuals was defined
and used in 1964 by the President's Council of Economic
Advisors. The Council estimated that in 1962 there were
between 33 and 35 million Americans living at or below this
, -t 16
income level.
Economists such as Leon Keyserling believe that this
criteria are too low. Keyserling raises the income criteria
for families living at the poverty level to $4,000 and for
single persons to $2,000 (i960 dollars). In his analysis
he includes a second grouping called the deprived category.
This group includes all families with incomes between
$4,000 and $5,999 and individuals with earnings between
$2,000 and $3,999. This latter group can subsist on their
incomes, but can not afford many of the luxuries that our
present society has to offer. Keyserling estimates that 73
17
million Americans fall within these two categories.
One of the weaknesses of these estimates, based upon
the above definitions, is that they do not account for the
differences in the cost of living throughout the country.
They also do not contain flexibility to take into account
inflationary periods of the economy which increase the coot
of living for individuals on fixed incomes. The social
security administration has made some attempt to account

12
for the variation in cost of living, but the invariant
$3,000 base is still largely used by them. A more flexible
standard which distinguishes the differences in need between
size of families, rural and urban residence, was published
in 1964 by the Department of Health, Education and Welfare.
This minimum budget ranges from $1,100 for a farm family of
two, aged 65, to over $5,000 for a nonfarm family of seven
or more individuals. This more flexible budget estimated
Ig
the poor at 34.6 million. More realistic income criteria
are those suggested by the Bureau of Labor Statistics.
Their index was developed for an urban family of four liv¬
ing in 20 of the major cities. This index varied from
19
$5,370 in Houston to $6,567 in Chicago.
The above estimates show the disagreements among
experts on what is a proper income level for the definition
of poverty. These estimates, furthermore, have not taken
into account transfer payments in the form of financial
assistance from welfare agencies or relatives. They also
have not made any allowance for taxation and therefore,
because of these omissions, do not represent true disposable
income. Indeed most students of poverty feel that the use
of a single index is not sufficient. The trend is to use
several indicators such as color, age, income and education.
Some social scientists have approached the problem
from the sociological and psychological viewpoint. One of

13
the first to use this definition was Oscar Lewis, who states:
Poverty in modern nations is not only a state
of economic deprivation, of disorganization or
of the absence of something. It is also positive
in the sense that it has a structure, a rationale,
and defence mechanisms without which the poor
could hardly carry on. In short it is a way of
life, stable and persistent passed down from gen¬
eration to generation along family lines. The
culture of poverty has its own modalities and
distinctive social and psychological consequences
for its members.
Economically retarded areas have within them distinct
cultural features. The poor having little money, virtually
no savings and no economic security, have evolved unique
consumer habits, family structures and a way of life. This
subculture, if it may be so called, is composed of the
chronically unemployed or those having irregular employment.
Those w'ho are employed work in the non-unionized low-paying
sectors of the economy. Their low salaries can not provide
for the costs of higher education and health services.
Often their meager earnings are spent on entertainment
luxuries, instead of the basics of food, shelter and cloth¬
ing, in an attempt to escape their environment for short
periods of time. The end result is often a low resistance
to disease, ill health and a shorter life span.
The everyday problems of this subculture, due to the
constant stress of maintaining life without hope, has created
in turn psychological manifestations. This has lead to a
lack of motivation and imagination among the destitute. Pro¬
tective shells against the outside world are erected. V/ithin

14
these shells provincialism and deep-rooted prejudices are
maintained. The outside world is viewed with hostility.
This hostility is expressed in their distrust of various
government agencies and programs.
The sociological approach in defining poverty,
although accurately describing the subculture of the poor,
has certain drawbacks. The major restriction is the lack
of sociological data for large sections of the country. To
gather the data reo^ires personal interviews, which not only
can be costly, but also time consuming. In addition, this
definition emphasizes to a great extent the end results
rather than the causes.
A more recent concept, economic health, is gaining
in popularity as it attempts to analyze which indices are
responsible for a lower level of living and how these
indices may vary through space and time. The term economic
health is not universally used, but is represented in a
large number of mu.ltivariate studies that use numerous
social, economic and demographic indices. These indicators
are then reduced to a smaller number by elimination, or
union with other variables into new ones. This new set of
indices, which accounts for the largest possible variance
within a workable matrix, is then used to explain the dif¬
ference in economic growth. This matrix can be applied to
inter-, or intra-regiona.1 studies and to a certain extent

15
be used to explain the changes through time. Economic
health thus tries, by the use of a large number of variables,
to explain why and how different regions of a country,
relative to each other, differ in their overall economic
prosperity. The method of analysis is not new, but it was
not until rapid calculating computers became available that
large regions could be examined on the basis of a multitude
of variables, and at the same time remove personal bias in
regard to the importance of the variables used. This form
of analysis is being used successfully by various social
scientists engaged in regional studies.
The concept of economic health will be used in this
study to explain the differences in prosperity, if any,
within the five southeastern states in the study area.
Twenty-six variables which are indicators of prosperity
will be utilized to ascertain what role they play within
the study area, and if their importance changes through
space and time.
Review of_ the Literature
The dependence of the Southeast on agriculture has
23
been documented by geographers (Hart, Vance). This
historical reliance upon a few major crops such as cotton
and tobacco had its drawbacks. These crops were export
orientated and consequently fluctuated greatly in price.

16
Soil mismanagement often lea to the abandonment of farms or
a loss in fertility and yields. The plantation form of
agriculture was not conducive to the modernisation of agri¬
cultural practices, as cheap labour after slavery made it
economically unfeasible to adopt some of the new techniques
until relative prices of the inputs changed. Rural values
predominated and eventually evolved into a way of life.
Political power came under the control of a small, land
based minority to the detriment of education and industrial¬
ization .
This theme, that the South's historical development
acted as a barrier to progress, pervades the literature on
the South. Nichols in a paper "Southern Tradition and
Regional Economic Progress" suggests that five social, polit¬
ical, psychological and philosophical factors were respon-
22
sible for the lag in southern economic growth. These were:
the dominance of agrarian values, rigidity of a social class
structure which prevented the establishment of a strong mid¬
dle class, the control of the political parties by a few which
led to an undemocratic political structure. Furthermore
a lack of social responsibility hindered the development
of a first class public school system. This, tied with the
conformity of thought and behavior which was expected of a
good southerner, left little or no sympathy for dissenters,
many of whom migrated North. The loss of this liberal ele¬
ment hindered change and helped maintain the status quo.

17
In spite of these handicaps the South in the last
few decades has made tremendous strides forward. MacDonald
"On South's Recent Economic Development " points out that
personal income in this region, has increased since 1940,
2 3
at a faster rate than in the nonsouth. Part of this
increase is attributed by MacDonald to a shift in the Amer¬
ican economy from coal and steel to plastics and petroleum,
and the increased exploitation of other southern resources
especially forestry and agriculture. The other main factor
responsible for the rapid increase in personal income was
a slow population growth rate, due to the heavy out-migra¬
tion of low income groups. This recent progress of the
South was further verified by Lassiter. In his study
"Education for Males by Region, Race and Age " Lassiter
pointed out that at present the investment in education is
providing better financial returns for white southern males
than white northern males.
This rapid rate of growth however has not been con¬
sistent throughout this region. Moon and McCann, who looked
at the levels of adjustment in 30 rural counties in the
states of Kentucky, Tennessee, North Carolina, Texas,
Louisiana, Mississippi and Alabama, found statistically sig-
25
nificant variation. Similar conclusions were arrived at
by Tarver and Beale in "Population Trends of Southern Non-
2 6
metropolitan Towns." They concluded that the significant
factors explaining the differences in population growth

16
among 801 southern towns with populations between 2,500 and
9,999 were: the size of the town end its regional location,
Fuguitt on the other hand found that county seats in rural
nonmetropolitan areas grew at the expense of neighboring '
. 27
centers.
The importance of urbanization as a significant fac¬
tor in explaining regional growth in the United States has
received considerable attention. Schlesinger has argued
that since the Civil War, cities dominated the American
28
growth process. Friedman and Alonso depict a spatial
economy in which the processes of economic development are
leading to greater specialization and increased polarization
of growth in a few urban industrial complexes, each of which
2'
is functionally integrated with a surrounding economic area.
These centers not only grow so rapidly as to create
problems of an entirely new order, but they also act
as suction pumps, pulling in the more dynamic ele¬
ments from the more static regions. The remainder
of the country is thus relegated to a second class,
peripheral position. It is placed in a quasi¬
colonial relationship to the center, experiencing
net outflows of people, capital, and resources,
most of which is redound to the advantage of the
center where economic growth will tend to be rapid,
sustained and cumulative. As a result, income dif¬
ferences between the center and the periphery tend
to widen.30
This viewpoint is also expressed by Schultz, who
argues that economic development occurs primarily within
an industrial urban matrix where the factor and product
markets work best. Agricultural activities located most

19
favourably in regard to these centers can utilize the fac¬
tor and product markets more efficiently than those located
on the periphery of these urban systems.^ This was
partially substantiated by Bryant who used multiple regres¬
sion to measure the causes of inter-county variation in
32
farmers’ income in the United States. He showed that
variation in farm income is largely determined by age, land
and capital inputs, education and color. Location in respect
to regional urban areas was only significant in the eastern
United States, implying that farmers west of the Mississippi
were more dependent upon national and international markets.
0. D. Duncan, et al., in Metronolis and Region, emphasize the
domination of hinterland activities by metropolitan centers^
Both agricultural and non-metropolitan manufacturing activ¬
ities were shown to be a function of general accessibilit}'-
to the urban system of the nation and distance to the local
metropolis. Berry concluded in his study of rural poverty
in Ontario that
areas lacking the dynamic conditions of industrial
urban development experience high rates of out¬
migration and commitantly realize differences in
per capita income, even though labour migrations
generally appear sufficient to overcome differences
attributable to differences in original resource
endowments. The factors linking rurai. poverty to
industrial urban growth polarized in urban centers
and varying in amplitude according to the rank of
these centers in an urban hierarchy, are thus clear. 4
This urban influence has also been clearly demon¬
strated by Thompson, et al., Berry,^ Bell and Stevenson,

20
OQ
and Hodge. All of these authors used factor analysis on
a large matrix of economic indices in order to derive a
smaller number of dimensions or factor loadings which would
explain more easily the variations in income levels. In all
cases an urban dimension was derived from the analysis that
was correlated with the rapidly growing areas. The study
by Thompson, et al., on the economic health of New York
State, is of particular interest here. It not only demon¬
strated that the large metropolitan cities were the nodes
of economic growth, but also that the predominantly agri¬
cultural counties near these centers fared better than
smaller cities, which, had populations of 50,000 or less.
Berry arrived at the same conclusions in bis analysis of
rural poverty in Ontario. Location of the rural townships,
in relation to urban areas, played an important role in
determining the level of economic achievement as measured
by 47 variables. In only a few exceptions, most of which
were due to the presence of mineral resources, did a few
townships deviate from the pattern of hinterlands dominated
by large urban industrial clusters. Using similar procedures
Hodge analyzed 473 trade centers in Saskatchewan. He found
that trade center viability was largely a function of rank
on the urbanism scale; the lower the rank the greater the
probability of decline.
The apparent importance of the urban or industrial
dimension is clearly seen in these studies. That these

21
complexes do grow at the expense of their hinterlands is
readily verified, if one looks at the variables which are
correlated with the urban industrial dimensions. These are
general]y high income groups, youthfulness of the popula¬
tion, a high level of education and a high level of capital
expenditures. All are indices of a healthy economy. The
converse is true; the dimensions representing poverty have
weaker associations with the above indices.
Are these indices stable through time? The decline
of certain regions relative to others would seem to indicate
39
the contrary. King found that they were not stable. In
his examination of Canadian cities he found that the per¬
centage of variance explained by the dimensions arising from
his analysis varied through time, although the factor load¬
ings associated with the dimensions remained fairly stable.
The Problem
The problem involved in this study relates to the
variations in the economic environment among 454 comities
in Alabama, Georgia, Mississippi, North and South Carolina.
Specifically does variation exist? What dimensions account
for this variation? Do they vary over time? What percent¬
age of poverty do they explain? Are some counties progres¬
sing at the expense of others? The literature reviewed pro¬
vides a basis from which to examine several useful spatial
hypotheses of economic health in the selected southeastern
states. The hypotheses to be tested in the study are:

22
(1) There is variation in economic health among the
selected counties and these can be explained in terms of a
few dimensions arising out of factor analysis.
(2) These dimensions are not stable but change
through time as do the conditions required for economic growth.
(3) Within the study area there will be nodes of
rapidly growing areas which thrive at the expense of their
hinterlands.
(4) These nodal areas are associated with urbaniza¬
tion and it is postulated that the degree of growth within
these sectors, relative to each other, is a function of popu¬
lation size and location, location in relation to local,
regional and national metropolitan centers.
This study will be successful if it can demonstrate
the underlying characteristics of areal variation in economic
health in the five selected southeastern states for two points
in time, and ascertain if these characteristics have changed
between the two time periods (1950 and I960).

REFERENCES
Richard Hartshorne, "Geography and Economic Growth
in Essays on Geography and Economic Development, Norton
Ginsburg (ed.) (Chicago: The University of Chicago Press,
I960), pp. 24-25.
2
Edward L. Ullman, "Geographic Theory and Under¬
developed Areas," in Essays on Geography and Economic Devel
opment, Norton Gins burg '(ed". ) (Chicago : The University of
Chicago Press, I960), pp. 26-27.
â– ^Ibid., p. 29.
^Hartshorne, op. cit., pp. 18-23.
^John Galbraith, The Affluent Society (Boston:
Houghton-Ivlifflin Co., 195'8 )3
^Ibid., pp. 323-333.
7
Michael Harrington, The Other America, Poverty in
the United States (Baltimore: Penguin Books, 1966), pp.
176-179.
n
°Lar A. Levitan, "Programs in Aid of the Poor,"
Poverty and Human Resources Abstracts, Vol. I (1966), pp.
11-25.
^Herman P. Miller, Poverty American Style (Belmont:
Wadsworth Publishing Co., Inc., 196bJ, pp. 2-32
â– ^Louis A. Ferman, Joyce L. Kornbluh and Alan Haber
(eds.), Poverty in America (Ann Arbor: University of
Michigan Press, 1965 )] pV1.
11Ibid.
12Miller, op. cit., p. 5.
23

24
^Gordon E. Reckord, "The Geography of Poverty in
the United States," in Problems and Trends in American
Geography, Saul B. Cohen (ed.) (New York': Basic Books,
IncY, 1967), p. 105.
14Ibid.
15
Joseph Spinelli, "A Study of let Out-Migration in
Southeastern Ohio, 1950-1960" (Unpublished Master's Thesis,
Department of Geography, Ohio State University, 1966).
16
Miller, op. cit., p,
2.
17
Ibid.
18
Harrington,
loc . cit.
19Ibid.
20
Elizabeth Herzog, "Some Assumptions About the Poor,"
Social Science Review, XXXVII (December, 1963), 390.
21
" John Fraser Hart, The Southeastern United States
(Princeton: D. Van Nostrand Co., Inc., 1967)"; and Rupert ~B.
Vance, Human Geography of the South (Chapel Hill: Univer¬
sity of North Carolina Press, 1935)•
22
William H. Nichols, "Southern Tradition and
Regional Economic Progress," Southern Economic Journal, XXVI
(January, I960), 187-198.
2 3
Stephen L. MacDonald, "On South's Recent Economic
Development," Southern Economic Journal, XXVIII (July, 1961),
30-40.
^4Roy L. Lassiter, Jr., "Education for Males by
Region, Race and Age," Southern Economic Journal, XXXII
(July, 1965), 15-22.
25
Seung Gyu I/Ioon and Glenn C. McCann, Subregional
Variability of Adjustment Factors of Rural Families in the
South (Raleigh: North Carolina Agricultural Experiment
Station, Southern Cooperative Series, Bulletin No. Ill,
January, 1966).

25
26
James D. Tarver and Calvin L. Beale, "Population
Trends of Southern Nonmetropolitan Towns," Rural Sociology.
XXXIII (March, 1563), 19-29.
27
Glenn V. Euguatt, "County Seat Status as a Factor
in Small Town Growth and Beeline," Social Forces, XLIII
(December, 1965), 245-251.
28
Arthur M. Schiesinger, A History of American Life,
Vol. XI, The Rise of the City 1675-1 d5‘6 (New fork: The Mac¬
millan Co., 1949)'.
29
John Friedman and William Alonso (eds.), Regional
Development and Planning (Cambridge: Massachusetts Institute
of Technology Press, 1964), pp. 1-11.
30Ibid., p. 3.
3^T. W. Schultz, Economic Organization of Agri¬
culture (New York: MeGraw“Hill Co., 1953/.
32
Keith W. Bryant, "Inter-County Variation in
Farmers’ Income," Journal of Farm Economics, XLVIII
(August, 1966), 557-577.
33
Otis Dudly Duncan, et al., Metropolis and Region
(Baltimore: The John Hopkins Press, I960)V
3^Bria,n J. Berry, Strategies, Models and Economic
Theories of DeveD opment in Rural Regions. Agricult ural
Economic' lie port No, 127 (Washington: Economic Research
Service, U.S. Department of Agriculture, 1967).
35
John H. Thompson, et al., "Toward a Geography of
Economic Health: The Case of New York State," in Regional
Development and Planning, John Friedman and William Alonso
Teds.) (Cambridge: Massachusetts Institute of Technology
Press, 1964), p. 187.
r
JUBrian J. Berry, "Identification of Declining
Regions: An Empirical Study of the Dimensions of Rural
Poverty," in Thoraan, R. S. (ed.), Areas cf Economic Stress
in Canada (Kingston: Queens University, 1965).
37W. H. Bell and D. W. Stevenson, "An Index of Eco¬
nomic Health for Ontario Counties and Districts," Ontario
Economic Review, II, No. 5 (1964), 1-7.

26
â– )0
'50Gerald Hodge, "Do Villages Grow? Some Perspectives
and Predictions," Rural Sociology, XXX, No. 2 (June, 1966),
183-196.
39
Leslie J. King, "Cross Sectional Analysis of
Canadian Urban Dimensions 1951 and 1961," Canadian Geographer,
X, No. 4 (1966), 205-224.

CHAPTER II
STUDY AREA AND METHODOLOGY
The study area is composed of five states, Missis¬
sippi, Alabama, Georgia, North and South Carolina. This
region covers an area of 241,968 square miles, and in 1968
had a population of 16,326,746. These states are similar
in their socio-economic-political characteristics (Hart,’1’
2 3
Vance, Odum ). All five withdrew with the Confederacy and
after Reconstruction the states voted solidly Democratic
until 1964, with the exception of North Carolina which went
Republican once.
Perhaps due to the apparent homogenity of the South,
few authors have attempted to subclassify it. What region¬
alization of the South has taken place is largely based
upon physical divisions such as the Delta, Piedmont,
Appalachia and the Coastal Plain (White and Foscue,^
Griffin, Young, and Chatham ). Bogue and Beale however,
using 64 variables, have broken the South down into four
distinct regions.^ One of which, the Southeast Coastal Plain
and Piedmont, coincides very closely with the area under
study (Figure 1). Five distinct characteristics separate
this region from the others. It is the least urbanized, has
the highest percentage of Negro population, the highest
27

r\o
oo

29
percentage of the labour force in agriculture, and the
7
industrial base is weak and highly dependent upon textiles.
These four characteristics are the cause of the fifth, low
median income, as the lowest median family income of any
region in the United States is found here. Furthermore
Bogue and Beale found enough variation within this distinct
region to classify it into 19 subregions, implying that
there may be differential growth patterns.
The choice of this study area was dependent to a
great extent on the previous work done by Bogue and Beale.
Although the two areas are not exactly the same, they may
be considered as one in their characteristics. The reason
for this discrepancy is that state boundaries were maintained
so as to reduce the effect of state governments in the anal¬
ysis.
The Setting
Three distinct physiographic features fall within
the study area, the Coastal Plain, Piedmont and Appalachia.
The boundary between the Piedmont and the Coastal Plain is
demarcated by a Fall Line and a change in soils. The Fall
Line is characterized by a zone of waterfalls and rapids in
the streams which imposed a limit on navigation. Cities
such as Raleigh, Columbia, Macon, Columbus and Montgomery
arc found along this break. In the Carolinas and Georgia
the Pall Line is marked by a long, low narrow belt of sand

30
mils, the remnants of ancient beaches. The Piedmont soils
are residual, the product of weathering of the underlying
rock over eons of time. The Coastal Plain soils on the
other hand were formed of sedimentary materials when the
plain was submerged.
Both physiographic divisions have a variety of local
topography ranging from level to hilly, with the older Pied¬
mont having the greater extremes of relief. A small portion
of the study area falls within Appalachia. This section
extends from central Alabama to the Blue Ridge of North¬
western Georgia and North Carolina. In the Blue Ridge the
mountains rise 3,000 to 6,000 feet above sea level. In the
Appalachian Valley of Alabama the relief is much less, varying
from 500 to 1,000 feet.
The economy was and to a large extent still is
centered on agriculture. Three major crops: tobacco, cot¬
ton and peanuts provide the majority of the farmers' cash
receipts. Industrial growth has been based upon service
functions, textiles, food processing, furniture and lumber¬
ing. In many respects the economy is similar to that of
underdeveloped countries, in that it is dependent on a few
staple agricultural products and has a weak industrial base.
Historical Background
The retardation of this area,as in the rest of the
South, has largely been due to the historical development of

31
the region. This historical retardation has been well
8 9
documented by historians and economists (Eaton, Maddox ).
It was the development of an agrarian culture that led to
the exclusion of the rise of a strong middle class, and
industry. These could have become the impetus for economic
growth, when agriculture was diminishing in importance.
Once behind, the South was in a poor position to compete
with the industrial North, and it became subservient to a
centralized economy based upon market potentials.
The Southeast was settled by two streams of settle¬
ment. The first group occupied the tidewater between
Charleston and Savannah. Plantations were established and
at first indentured servants, later slaves, were brought in
as a source of labour. Rice became the dominant crop along
the coast, while indigo was grown on the plantations further
inland. As the tidewater was being settled a different
group of immigrants were coming into the western Carolinas
and Georgia. These were the Scotch-Irish and Germans who
had come south from Pennsylvania. Many of these settled
permanently in the backcountry and became successful farmers,
raising all their requirements. The excess livestock, hogs,
com were sold or converted to other more valuable by¬
products such as liquor or hides. The heavy migration into
the western area led to a more rapid rate of growth than in
the tidewater. Eventually this may have led to the political
control of the southern colonies by the small farmers, had

32
not the importance of cotton began to assert itself. Cotton
planters began to move into the Piedmont and it became the
major crop. The search for more and better cotton soils led
to the opening up and settlement of Alabama and Mississippi.
The relatively rapid rise in the dominance of cotton resulted
from two major events, the industrial revolution in Europe,
requiring cotton for its textile mills, and the invention of
the cotton gin by Whitney in 1793.
The emergence of King Cotton as the dominant crop,
controlled by a small group of planters, utilizing slave
labor, pre-empted the use of land for other crops. This
small aristocratic group gained and maintained political
control throughout the area. Furthermore this group man¬
aged to impose their values upon the entire southern soci¬
ety. The use of slave labor was a barrier to the establish¬
ment of more productive agricultural practices which would
utilize a higher share of capital as a factor of production.
The social stigma associated with trade and manufacturing
hindered the development of a middle class and a more diver¬
sified economy. Perhaps the greatest damage done was the
concept that education was not for the masses, but for the
select few. The education provided the select few, although
of high calibre, was heavily oriented toward the liberal
arts and was lacking in the field of applied technology.

33
Other important sources of income were lumber and
naval stores, which initially were of prime importance in
North Carolina but rapidly spread throughout the South.
Tobacco became the second major crop when, in 1850, it was
discovered that bright tobacco did well on the poor sandy
soils of northwestern North Carolina. The third major crop
was the growing of peanuts along the Georgia and Alabama
border.
Thus in the period preceding the Civil War the
southern economy was oriented around the export of a few
major staples as sources of revenue with which to pay for
the importation of manufactured goods from abroad and the
North. This raw material orientation meant that the econ¬
omy was affected by any severe fluctuations in prices or
tariffs. The tariff of abominations established by the
North in 1828 was a blow to southern commerce.^
The Civil War brought ruin upon the South. The
abolition of slavery removed a source of cheap labour from
the plantations and left many of these large land holdings
idle. A system of sharecropping evolved to fill the labour
shortage. The system was not satisfactory for it did not
encourage any long term agricultural practices by either
the landowner or the sharecropper. In addition, develop¬
ment of either industry or a more efficient system of
agriculture was strongly hampered by the shortage of invest¬
ment capital. Inflation during the War had left the

34
Confederate states with a worthless currency. Deflation
after the war left the South with little buying power.
Some of the capital problems were attributable to national
monetary policies which hurt the South. The National Bank
Reforms in 1865 levied a severe tax on state banks, which
predominantly were located in the South. Furthermore in
these bank reforms, the South did not receive its fair
11
allotment of Federal currency.
The establishment of certain industries such as
iron and steel was also hampered by the monopolistic action
of northern firms who established ,:Pittsburgh plus basing
point" and by this means removed the locational advantage
of Birmingham. The industrial growth that did take place
occurred mainly in the field of food processing, textiles
and cigarette manufacturing. The use of bright tobacco as
a major ingredient in cigarettes gave rise to cities like
Durham and Winston Salem. The textile industry moved here
from the North during the later part of the nineteenth
century. Many of the mills were established in towns along
the Fall Line where there was a source of power. Despite
these attempts at industrialization, the South did not keep
pace with the North. As most poor or stagnating areas it
suffered from selective out-migration, lack of capital and
the establishment of new firms. All of these factors
encouraged the location of industry in the heart of the
more lucrative northern market.

35
Thus the South entered the twentieth century in a
disadvantaged position. It was an area short of skilled
labour, investment capital, and entrepreneurs, technolog¬
ically and productively lagging behind the other sections
of the country.
Recent Trends in the Economy
Despite these handicaps, great strides have been
made in the last 30 years throughout the study area and
the entire South. This has come about due to the mechaniza¬
tion of agriculture, and better utilization of other natural
resources. The development of better educational facilities,
bi-racial labour force, more internal savings and capital
investments has encouraged the location of industry, espe-
12
cially those that benefit by decentralization.
The trend has been to a diversification of both the
agricultural and industrial segnents of the economy. Exam¬
ples are the growth of the dairy industry in Mississippi,
the establishment of beef cattle on the Black Prairie of
Alabama and the Piedmont of Georgia, the location of the
broiler industry since the war in northern Alabama and
Georgia. Abandoned cropland, formerly in cotton, has been
reverting or restocked to pine woodland. Forest management
has been initiated and encouraged by the rapid expansion of
the pulp and paper industry. Better conservation practices,
increased use of fertilizers and mechanization of farms

36
have resulted in higher yields. Prices of important cash
crops such as cotton, tobacco and peanuts have been stabi¬
lized by a system of acreage allotments.
Industrial growth has been spurred by the establish¬
ment of large important industries. Some major examples
are: the atomic energy plant near Aiken, South Carolina;
the Redstone guided missile project and other space facilities
at Huntsville, Alabama, the large Lockheed aircraft plant in
Marietta, Georgia. Less dramatic has been the location of
electronic firms in the Golden Triangle of North Carolina.
The location by northern companies of branch firms in south¬
ern cities has aided in the establishment and diversifica¬
tion of industrial growth. Atlanta owes a great deal of
its grev/th to the establishment of these regional branches.
It has become an assembly center for Ford and General Motors.
Many insurance and airline firms have their regional offices
here. In addition the overall economic growth in the South
has been aided by the establishment of large military bases
and the rapid increase in tourism, both of which have been
attracted by the mild climate.
Outline of the Analytical Model
Briefly the steps undertaken in the study are as
follows: (1) transfer raw data into an M x H matrix (M
variables and N observations), (2) perform principal com¬
ponents analysis of the M x N intercorrelation matrix of

37
the M variables and rotation of the resulting eigenvectors
to a normal varimax position to satisfy the criterion of a
simple structure, (3) compute the factor scores of N obser¬
vations on the R rotated factors, (4) regress a dependent
variable D on the factor scores.
The Model
The mathematical model used to determine the under¬
lying dimensions of economic health is known as a principal
axis factor analysis solution.1-^ The choice of the model
is based upon several considerations. The study of the
covariance of economic, social and demographic character¬
istics of regions is a problem of a multivariate nature.
Data of this nature are often not-normally distributed and
are highly interrelated. Regression models using a large
number of variables lacking the assumptions of normality
and independence will give a coefficient of determination
which is biased upwards. In addition, the coefficients and
standard errors of the coefficients may be affected. Fac¬
tor analysis, however, can be used on data lacking these a
priori assumptions. This technique can simultaneously
manage a large number of variables, compensate for random
error, and disentangle complex interrelations into their
major and distinct regularities."^ The essence of this
technique is to reduce a large number of variables into a
smaller group of distinct orthogonol factors, components,

38
or dimensions with a minimum loss of information. On the
basis of these factors the unique variations of a domain can
be discerned. A further step is the obtaining of factor
scores which is the location in space of each observation
on the particular dimensions. These factor scores can be
used as independent variables regressed on any desired
15
dependent variable. This method is suggested by Scott.
Scott contends that when the assumptions of independence
and normality are not known, this regression technique is
more accurate than the classical least squares technique.
The essential features of factor analysis are as
17
follows: from a rectangular X matrix of raw data it is
possible to
compute
a R matrix
of s
imple
correlation coef
ficients.
xll
xl2
• • •
xln
X =
x21
x22
• • •
x2n
xml
xm2
• * •
xrnn
XX* - R
rll
rl2
* • •
rln
R =
r21
r22
• • *
r2n
• • *
rn1
• * •
m2
• • •
• • «
• • •
rnn
The problem in factor analysis is to find a matrix A such
that R ^ AA*. This is done by solving for, using an iterative
process, the eigenvalue or characteristic roots L, where L
is the sum of the squared factor loadings of the M varia.bl.es
on a dimension. The factor loadings are the correlations

39
between the variables and the dimensions. One can solve
for as many characteristic roots as there are variables;
however, one generally only solves for the most important
roots, that is for the ones that account for the majority
of the variance. The equation for solving L is (R « LI) » 0
where I is the inverse of R. After finding the eigenvalue
L one can solve for the eigenvector V which is the column
vector of factor loadings, by the following equation (R =
LI) V = 0. Repetition of this procedure yields the factor
matrix A.
all a12 ... ami
. a21 a22 ... am2 r is the number of character-
A - istic roots solved for.
ami amr
Dividing the characteristic roots found by the total com-
munality (communality is the sum of independent common
variances of the components on each variable) yields the
variance accounted for by each dimension.
It is often of interest to find the factor scores,
that is the ranking of each observation on each of the
derived components. The matrix of factor scores is given
by the following equation F- XAL*. In addition the factors
are rotated so as to comply with Thurston's concept of sim¬
ple structure, such that each of the original variables
l8
relates highly with one and only one of the new dimensions.
This assures that the first factor arising from the analysis

40
accounts for the maximum variance and each successive factor
accounts for decreased proportions of the explained variance.
The factor scores have further application in a
regression model to explain the per cent of variation of
a dependent variable which is accounted for by the orthogonal
dimensions. This procedure produces an index which has the
largest possible variance of any linear combination. A
lower coefficient of determination is obtained in compari¬
son with the least squares method. Factor regression
accounts for errors in the variable high intercorrelations
and gives regression coefficients which are more accurately
19
representative of economic theory.
Variables Used
The variables used for the two time periods were
chosen to represent social, economic, demographic and
locational characteristics of the study area. In all, 26
indices were used. They were obtained from the censuses
of population, agriculture, and the city and county data
book for the respective time periods. Three locational
indices were calculated for the study. The following list
is provided to show the nature of the variables.

41
Table 4
Variables Used
1 Total population
2 Per cent oí' the population urban
3 Per cent of the population nonwhite
4 Per cent of the population 65 years old or older
5 Per cent of the population 25 years old or over who
completed less than five years of school
6 Per cent of the population 25 years old or over who
completed high school
7 Per cent of housing units sound
8 Median value of owner-occupied dwelling units
9 Median gross rent
10 Value added by manufacturing
11 Total capital expenditures by manufacturing firms
12 Total retail sales
13 Total wholesale sales
14 Per cent of the population under five years of age
15 Per cent of the labour force unemployed
16 Per cent of the farms operated by tenants
17 Average value of land and buildings per farm
18 Total value of farm products produced (sold)
19 Number of commercial farms earning less than $2,500
annually
20 Number of farmers whose off-farm income is greater
than farm income
21 Per cent of families earning between $0 - $999 annually
22 Per cent of families earning between $1,000 - $1,900
annually
23 Per cent of families earning between $2,000 - $2,999
annually
24 Distance in miles from the center of a county to the
nearest city with a population of at least 50,000
25 Distance in miles from the center of a county to the
nearest city with a population of at least 500,000
26 Distance in miles from the center of a county to New
York City
Through these variables an attempt was made to strike
a balance between the rural and urban segments of the study
area. This choice also took into account that the majority
of these indices have been found to be significant by other
researchers, in explaining the variation in economic health.

42
The major problem with the data is that the areal units
(counties) vary immensely in size. Large counties having
extensive variation within them may not accurately reflect
this in the data. In spite of this inherent difficulty,
county data are at present the best available for such an
analysis.
Summary
This chapter is divided into two sections. The
first describes the study area and the historical basis
for its poverty. This region was selected for analysis
as it is representative of the southeastern United States.
Furthermore, it appears to have had a substantial number
of economic changes within it in the last few decades, and
thus would be an ideal testing area for the analysis of
spatial and temporal patterns of economic health.
Factor analysis was chosen as an appropriate model
because it can handle nonnormal and highly intercorrelated
data. In addition it opens up other avenues of research
beyond the location and identification of the underlying
dimensions.

REFERENCES
"'â– John Fraser Hart, The Southeastern United States
(Princeton: D. Van Nostrand Co., Inc., 1967).
2
Rupert B. Vance, Human Geography of the South
(Chapel Hill: The University of North Carolina Press, 1936).
^Howard W. Odum, Southern Regions of the United States
(Chapel Hill: The University of North Carolina Press, 1936).
^G. Langdon White and Edwin J. Foscue, Regional Geog¬
raphy of Anglo-America, 2nd edition (Englewood Cliffs:
Prentice-Hall, 1956*).
5
Paul F. Griffin, Robert N. Young and Ronald L. Chatham,
Anglo-America, A Regional Geography o f the United S t a tes and
Canada (San Francisco: Fearon Publishers Inc., 1962*).
°Donald J. Eogue and Calvin L. Beale, Economic Areas
of the United States (New York: The Free Press of Glencoe,
Inc., 19617) p. XLII.
7Ibid., pp. 269-270.
O
°Clement Eaton, A History of the Old South (New fork:
The Macmillan Co., 196577"
g
James G. Maddox, et al., The Advancing South: Man¬
power Prospects and Problems (New York: The Twentieth.
Century Fund, 1967).
^William E. Laird and James R. Rinehart, "Exogeneous
Check on Southern Economic Development," South Atlantic
Quarterly, Vol. LXV, No. 4 (1966), pp. 491-5035
3 3
JJ*Ibid.
-*-^B. U. Ratchford, "Economic Development in the South,"
South Atlantic Quarterly, Vol. IjXIV, No. 4 (1965), pp. 496-
505.
43 *

44
"^The type of factor analysis used in this study is
the version known as principal component analysis.
â– ^R. J. Rummel, "Understanding Factor Analysis,"
Conflict Resolution, Vol. XI, No. 4 (1967), p. 444.
] R
'Xohn T. Scott, Jr., "Factor Analysis and Regres¬
sion," Econometrica, Vol. 34, No. 3 (July, 1967), pp. 552-
562.
16,.. . .
Ibid.
17
For a more detailed explanation of the model see:
H. H. Harmon, Modern Factor Analysis (Chicago: University
of Chicago Press, Ib’olj; W7~G. Kendall, A_Cqurse in Mu.lti-
variate Analysis (London: Charles Griff in, ~195 7)',' pp. 10-
*56f R• ~J« Rummel, "Understanding Factor Analysis, " Conflict
Resolution, Vol. XI, No. 4 (1967), pp. 444-480; Mary Ivlegee,
"On Economic Growth and the Factor Analysis Method," South¬
ern Economic Journal, XXXI (July, 1964)? 215-228.
-i O
‘LOQazi Ahmad, Indian Cities: Characteristics and
Correlates (Chicago: University of Chicago Press, 1965*77
p. 25.
â– ^Scott, op. cit., p. 552,

CHAPTER III
THE DIMENSIONS OF ECONOMIC HEALTH IN THE
SOUTHEASTERN UNITED STATES (1950-1960)
The analysis as reported in this chapter as well
as in the subsequent chapter, was undertaken at the Univer¬
sity of Florida's computing center. The original data, for
the two time periods, were transformed into symmetric 26 by
26 correlation matrices. These matrices were then subjected
to a principal axis solution which yielded six eigenvectors
for the 1950 data and seven eigenvectors for the I960 data,
based upon the criteria ofx>|. The components were then
rotated to a normal varimax position. The underlying dimen¬
sions obtained in this manner were used to identify and to
help interpret the areal variation in economic health.
The basic dimensions of economic health are identi¬
fied by the analysis of the derived factor loadings and fac¬
tor scores. Factor loadings are the correlations of the
original variables with the newly derived dimensions. If
each of the original variables were plotted in n-dimensional
space (n representing the number of derived factors), the
coordinates of the variables would represent the correlations
or loadings on each principal axis (dimension) by the indi¬
vidual variables. These factor loadings range from 1.0,
indicating a high positive correlation with a given dimension
45

46
(which must he interpreted), to a -1.0, indicating a high
negative correlation. Variables with high positive loadings
on a factor are expected to be most characteristic of the
quality represented in the factor, while high negative load¬
ings on the same factor should indicate a contrasting qual¬
ity. Factor scores on the other hand measure how closely
each observation (county) ranks on each of the individual
dimensions. For example, a high positive factor score for
a county on a given component can be interpreted to mean
that the county is highly representative of that component.
With this in mind, it becomes possible to identify, and
interpret the areal variation of economic health. In order
to facilitate interpretations, of the dimensions only high
loadings on each of the components are shown, as are only '
highly negative and positive factor scores in Tables 7--18,
21-34. The factor scores will also be employed later in
this chapter in a stepwise regression as a means of testing
how well the derived dimensions explain the variation in
poverty.
1950 Factor Analysis
Table 5 summarizes the proportion of the total inter¬
county variance in 1950 accounted for by each of the six
derived factors. The six rotated factors together explain
70.5 per cent of the iotal variance of the 26 variables.
Thus these six components are considered representative of
the variables that they replaced.

47
Table 5
Percentage of Total Variance Explained by Each
Dimension in 1950
Dimension Eigenvalue
I
8.71
II
3.09
III
2.14
IV
1.91
V
1.45
VI
1.02
Total
Per cent of total variance
33.5
11.9
8.2
7.4
5.6
3.9
70.5
Another indication of the representation of the
original variables by the new dimensions is given by the
communalities. These values are the total sum of squares
of the correlations of a particular variable with all six
components. As would be expected, all variables are not
equally represented (Table 6).
Table 6
Percentage of Variance (Communality) of each of the 26
Variables Accounted for by All Six Components in 1950
Variables
Communality in Percentage
1
96.6
2
74.5
3
79.1
4
19.3
5
83.0
6
82.2
7
87.1
8
53.4
9
75.5
10
84.2
11
59.6
12
77.2
13
69.4

48
Table 6—Continued
Variables
Communal it,y in Percentage
14
15
16
17
18
19
20
21
22
23
24
25
26
78.1
10.9
80.2
67.9
84.7
87.9
69.6
93.2
95.5
46.3
39.4
63.6
75.1
Two .variables, per cent of the population 65 years of age
and older and per cent unemployed, are poorly accounted for
as only 19.3 and 10.9 per cent of their total sums of squares
are explained. This implies that these variables were not
significant in the analysis. On the other hand the remain¬
ing variables are summarized quite adequately by the new
dimensions. Thus the variations of the selected indices
are effectively captured by this statistical technique, a
desirable trait in such an analysis.
The 1950 Dimensions
Factor I (Urban Industrial Counties), the most
significant dimension, contributing 33.5 per cent of the
total variance, isolates the urban ghettos of the Southeast.
This component is associated with poverty found in some of
the urban industrial complexes within this region. Table 7
shows the correlations between this dimension and urban

49
functions such as retail and wholesale sales. High loadings
are also associated with total population, value added by
manufacturing, capital expenditures, poverty and youthful¬
ness of the population. Counties which rank high on this
dimension are those which contain large rapid*1 y growing
regional centers such as Atlanta, Birmingham, Mobile, and
Charlotte (Table 8 and Figure 2). These centers, due to
their prosperity and rapid rate of economic growth, attract
a large number of immigrants from the hint crian 1, many of
whom are unskilled and consequently are employed in low
paying industries or seasonal work.
Table 7
Dime is ion I Urban Industrial Counties in 1950
Prima jy Vi rtables Factor Loadings
Total population
.92
Per cent of population urban
.44
Value added by manufacturing
.88
Total Capital expenditures
.73
Retail sales
.85
Ydiolesale sales
.82
Number of persons under five
years of age
•
CO
Per cent of families earning
$0-8999
between
.76
Per cent of families earning
$1,000-81,999
between
.90
Per cent of families earning
$2,000-82,999
between
«65

50
Table 8
Counties with Extreme Factor Scores on Dimension I
in 1950
High Positive
High Negative
Jefferson, Ala.
12.32
Chattahooehe, Ga.
-2.18
Fulton, Ga.
11.84
Fannin, Ga.
-1.34
Mobile, Ala.
3.94
Houston, Ga.
-1.03
Me c kl enljur g, N. C.
3.66
Carteret, N. C.
- .82
Forsythe, N. C.
3.51
Dare, N. C.
- .81
Guilford, N. C.
3.10
Onslow, N. C.
- .78
Chatham, Ga.
2.78
Lafayette, Miss.
- .76
Greenville, S. C.
2.52
Ohowan, N. C.
- .71
Charlestown, S. C«
2.33
Stone, Miss.
- .71
Spartanburg, S, C.
2.29
Orange, N. C.
- .70
Gaston, N. C.
1.97
Fayette, Ga.
1.59
Bibb, Ga.
1.59
Richland, S. C.
1.50
Montgomery, Ala.
1.48
Buncombe, N. C.
1.48
Hinds, Miss.
1.40
Wak e, N. C.
1.38
Durham, N. C.
1.35
Richmond, Ga.
1.30
The second
dimension which explains 13.3 per
cent
of the total variance
is a rural poverty indicator.
This
component is related
mainly to
types of farming and
sources
of farm income. High
positive
loadings are found on
total
value of farm products produced
and number of commercial
farms earning less
than $2,500
annually, the tenancy
rate,
and number of farmers
whose off
-farm income is greater than
farm income (Table
9)
. This dimension isolates the ;
poor
tenant and the small
individual
farmer who, because of lack
of finances, education, or the small scale of his holding,
has not been able to make a financial success of his


52
enterprise. Factor scores fail to reveal any easily defined
grouping of counties. Instead they suggest that this type
of poverty is quite widespread, especially in areas outside
the dominance of any large metropolitan center (Table 10).
There does appear, however, to be a concentration of poor
counties in the old cotton belt of the Southeast, suggesting
that these rural counties have not fully recovered from the
negative effects of the sharecropping system of agriculture
(Figure 3).
Table 9
Dimension II Tenant and Small Farm Poverty in 1950
Primary Variables Factor Loadings
Per cent of farmers that are tenants .49
Total vs,lue of farm products produced (sold) .83
Number of commercicl farms earning less than
$2,500 annual!? .93
Number of farmers \hose off-farm income is
greater than their farm income .39
Per cent of families earning between
$0-8999 .55
Table 10
Counties with Extreme Factor Scores on Dimension II
High Positive
in 1950
High Negative
Bolivar, Miss.
Sunflower, I.Iiss.
5.38
Evans, Ga,
-2.43
5.30
Glynn, Ga.
-1.93
Robeson, N. C«
3.61
Camden, Ga.
Muscogee, Ga.
-1.89
Leflore, Miss.
3.38
-1.64
Washington, Miss.
3.28
McIntosh, Ga.
-1.60
Coahoma, Miss.
3.15
Chatham, Ga.
-1.59
Cullman, Ala.
3.09
Liberty, Ga.
-1.52
De Kalb, Ala.
3.08
Fulton, Ga.
-1.51
Johnston, N. C.
3.00
Charlton, Ga.
-1.43

53
Table 10—Continued
High Positive
High Negative
Sampson, N. C.
2.76
Long, Ga.
-1.41
Horry, S. C.
2.74
Bryan, Ga.
-1.39
Orangeburg, S. C.
2.40
Camden, N. C.
-1.37
Tallahatchie, Miss.
2.36
Currituck, N. C.
-1.37
Columbus, N. C.
2.36
Doughertj, Ga.
-1.33
Florence, S. C.
2.34
Clin]:, Ga.
-1.32
Williamsburg, S. C.
2.22
Turner, Ga.
-1.32
Marshall, Ala.
2.15
Peach, Ga.
-1.31
Pitt, N. C.
2.15
Dare, N. C.
-1.28
Quitman, Miss.
2.09
Hancock, Miss.
-1.28
Cleveland, N. C.
2.05
Jones, Ga.
-1.26
Tunica, Miss.
2.02
Wayne, Ga.
-1.26
The third most important factor (Wealthy Urbanized
Counties) identifies a pattern of urban wealth. It is
associated with variables, such as per cent of the popula¬
tion urban, median value of owner-occupied homes, and per
cent of the population that completed high school (Table 11).
The economic influence of Atlanta is clearly seen in De Kalb,
Clayton and Cobb which rank high on this dimension. These
counties are to an extent dormitory suburbs of Atlanta and
as su.ch contain a high proportion of residents with above
average incomes. Many of the other comities listed in Table
12 owe their prosperity to the location of state institutions,
an example is Orangeburg, South Carolina. Other counties
that fain into this category are Hinds, Mississippi; Mont¬
gomery, Alabama and Richland, South Carolira. The influence
of state spe id.ii g is r elf-evident in these comities as the
respective state capitals are loerted there. Federal pay¬
rolls are another important source of prosperity, for exampie


55
Houston, Chattahooche and Muscogee counties in Georgia, the
sites of large military bases. Prosperity in these urban¬
ized counties can be traced to concentrations of highly edu¬
cated people working in professions that pay above average
salaries.
Table 11
Dimension III Wealthy Urbanized Counties in 1950
Primary Variables Factor Loadings
Per cent of the population urban .73
Per cent of the population that completed
less than five years of school -.43
Median value of owner-occupied homes .70
Average value of farms (land and buildings) .55
Per cent of the population that completed
high school .82
Table 12
Counties with Extreme Factor Scores on Dimension III
in 1950
High Positive
Chattahooche, Ga.
5.54
De Kalb, Ga.
4.39
Harrison, Miss.
3.72
Glynn, Ga.
3.71
Houston, Ga.
3.33
New Hanover, N. C.
3.08
Dougherty, Ga.
3.06
Muscogee, Ga.
2.98
Forrest, Miss.
2.73
Clarke, Ga.
2.71
Adams, Miss.
2.70
Hinds, Miss.
2.62
Montgomery, Ala.
2.44
Richland, S. C.
2.35
Jackson, Miss.
2,31
Orange, N. C.
2.24
Pasquotank, N. C.
2.16
Chatham, Ga.
2.11
High Negative
Fayette, Ga.
-2.25
Jefferson, Ala.
-2.16
Dawson, Ga.
-1.85
Jasper, S. C.
-1.53
Heard, Ga.
-1.42
Union, Ga.
-1.40
Stokes, N. C.
-1.35
Forsythe, Ga.
-1.35
Yancey, N. C.
-1.31
Berkeley, S. C.
-1.31
Calhoun, S. C.
-1.31
Banks, Ga.
-1.30
Echols, Ga.
-1.30
Lowndes, Ala.
-1.29
Fulton, Ga.
-1.28
Gilmer, Ga.
-1.27
Greene, Ala.
-1.26
Clebourne, Ala.
-1.26

56
Dimension IV (Rural If/hite Counties) shows the con¬
trast in levels of living between primarily white and black
counties. Variables that correlate highly negative with
this component are: per cent of the population nonwhite,
per cent completed less than five years of school and per
cent of farmers that are tenants (Table 13). A positive
correlation is found between this dimension and number of
farmers whose income from off-farm employment is greater
than fain income. High positive factor scores are asso¬
ciated with the Appalachian counties of Alabama, Georgia,
North and South Carolina-, with the heaviest concentration
in western North Carolina (Table 14). The predominantly
Negro counties rank low on this factor (Table 14). These
Negro counties are located in a belt which runs from east¬
ern North Carolina to the Mississippi Delta. The largest
concentrations are found in the Delta, southwestern Georgia
and the tobacco area of eastern North Carolina. The major¬
ity of the Negroes in this belt are the descendants of the
former slaves who worked on the cotton and tobacco planta¬
tions. Thus this area of colored poverty has its roots in
the plantation system of land tenure, whose remnants still
remain. Although both groups of counties are poor, the
Negro comities are more so, indicating the traditional
disparity between the black and the white segments of our
society.

57
Table 13
Dimension IV Rural White Counties in 1950
Primary Variables Factor Loadings
Per cent of population nonwhite -.85
Per cent of farmers that are tenants -.72
Per cent of population that completed less
than five years of school -.78
Number of farmers whose income from off-farm
employment is greater than farm income .60
Table 14
Counties with Extreme Factor Scores on Dimension IV
in 1950
High Positive High Negative
Wilkes, N. C.
2.30
Lee, Ga.
-2.56
Cullman, Ala.
2.25
Union, Miss.
-2.54
Walker, Ala.
2.13
Calhoun, Ga.
-2.20
Ashe, N. 0.
2.06
Burke, Ga.
-2.02
Fannin, Ga.
2.01
Turner, Ga.
-2.01
Madison, N. C.
1.99
Coahoma, Miss.
-2.01
Caldwell, N. C.
1.89
Terrel, Ga.
-1.92
Rendo.’'ph, N. C.
1.89
Peach, Ga.
-1.84
Greenvilí , S. C.
1.82
Lowndes, Ala.
-1.81
Do Kalb, Ala.
1.80
Stewart, Miss.
-1.79
Watauga, N. ;.
1.77
Issaquena, Miss.
-1.78
Buncombe, N. C.
1.66
Greene, Ala.
-1.77
Catawaba, N. C.
1.66
Baker, Go..
-1.73
Jackson, N. C.
1.65
Washington, Miss.
-1.71
Cherokee, N. C.
1.63
Bullock, Ala.
-1.71
Spartanburg, S. C,
1.59
Edgecombe, N. C,
-1.70
Davidson, N. C.
1.58
Quitman, Ga.
-1.66
Burk e, N. C.
1.55
Webster, Ga.
-1.63
Mitchell, N. C.
1.55
Adams, Miss.
-1.62
The last two components together contribute 10.1 per
cent to the total explained variation. These two dimensions
are difficult to identify. They appear to be indicators of
urban influence upon their hinterlands (Tables 15 through
18). Factor V indicates that a close proximity to a large

58
city may not always be beneficial. For a small urban center
generally is at a disadvantage competing with a large
regional center for a regional market, especially if the
regional center contains a large proportion of the regional
market. A large city in this instance may cause services
and industries to gravitate to it. This may be true of the
southwestern Mississippi counties that rank highly negative
on Dimension V (Table 16). These counties are depressed
economically not only because of their poor agricultural
base but also due to the lack of thriving urban centers.
In this case the presence of a large urban center is not
beneficial, rather it appears that New Orleans is growing
at the expense of this area. This in turn suggests that New
Orleans is losing its national significance to other more
rapidly growing centers such as Atlanta. This is in con¬
trast to the more prosperous counties of North Carolina,
whose produce is related more to national markets. Thus
this dimension appears to reflect the importance of national
markets to economic growth.
The last dimension is another locational index as it
has negative loadings with all the distance variables (Table
17). The positive loading with median va.lue of owner-occupied
dwellings is an indicator of wealth. This factor reflects
a locational advantage in regard to regional and local mar¬
kets. A glance at the factor scores will verify this loca¬
tional advantage. For example, counties with high positive

59
scores arevfound to be centered near large metropolitan cen¬
ters (Table 18). The economic effect of Atlanta is clearly-
seen on the Georgia counties. The counties with high nega¬
tive factor scores tend to be more isolated and economically-
depressed.
Table 15
Dimension V Locational Advantage in Relation to
National Market (New York) in 1950
Primary Variables ' factor Loadings
Per cent of population 65 years of
age or older -.31
Median gross rent .31
Avei^ge value of farm (land and building:.) .48
Total value of farm' products produced .27
Number of farmers
whose incone from off-
farm employment
is greate
r than farm income
-.22
Distance in miles
from the
center of a county
to a city of at
lea: b 900
,000
.66
Distance in miles
from the
center of a county
to New York
1
•
00
H
Table 16
Counties with
Extreme Factor Scores on Dimension V
in
1950
High Positive
High Negative
Horry, S. C.
2.22
Evans, Ga.
-2.55
Yadkin, N. C.
2.18
Pike, Miss.
-2.08
Glynn, Ga.
2.05
Hinds, Miss.
-2.07
Alleghany, N. C.
2.05
Harrison, Miss.
-1.99
Robeson, N. C.
2.00
Lincoln, Miss.
-1.97
Surry, N. C.
1.97
Copiah, Miss.
-1.89
Columbus, N. C.
1.95
Amite, Miss.
-1.80
Stokes, N. C.
1.94
Wilkinson, Miss.
-1.74
Davie, N. C.
1.81
Walthall, Miss.
-1.73
Harnett, N.C.
1.79
Marion, Miss.
-1.73
Forsythe, N. C.
1.79
Forrest, Miss.
-1.71
Pitt, N. C.
1.78
Jefferson, Miss.
-1.68
Florence, S. C.
1.74
Lafayette, Miss.
-1.68
Cabarrus, N. C.
1.72
Marshall, Miss.
-1.61
Dillon, S. C.
1.68
Tuscaloosa, Ala.
-1.61
Camden, Ga.
1.68
Jefferson Davis,
Miss r-1.56
Nash, N. C.
1.63
Dallas, Ala.
-1.56

60
Table 17
Dimension VI Locational Advantage in Relation to
Local Regional and National Markets in 1950
Primary Variables Factor Loadings
Median value of owner-occupied
dwelling units .69
Median gross rent .22
Distance in miles from the center of a
county to a city of at least 50,000 -.47
Distance in miles from the center of a
county to a city of at least 500,000 -.43
Distance in miles from the center of a
county to New York -.26
Table 18
Counties with Extreme Factor Scores on Dimension VI
in 1950
High Positive
High Negative
Fannin, Ga.
12.76
Lowndes, Ga.
-2.51
Fayette, Ga.
2.90
Georgetown, S. C.
-1.84
De Kalb, Ga.
2.15
Is s aquena, Mis s.
-1.79
Edgecombe, N. C.
2.02
Wayne, Ga.
-1.76
Clayton, Ga.
1.95
Choctow, Ala.
-1.76
Wilson, N. C.
1.86
Washington, Ala.
-1.64
Turner, Ga.
1.86
Clarke, Miss.
-1.62
Hertford, N. C.
1.81
New Hanover, N. C.
-1.55
Evans, Ga.
1.76
Columbus, N. C.
-1.52
Tunica, Miss.
1.75
Marion, S. C.
-1.46
Nash, N. C.
1.69
Louderdale, Miss.
-1.44
Wake, N. C.
1.66
McIntosh, Ga.
-1.44
Greene, N. C.
1.63
Choctow, Miss.
-1.43
De Soto, Miss.
1.55
Jasper, Miss.
-1.42
Henry, Ga.
1.52
Geneva, Ala.
-1.41
I960 Analysis
In I960 seven 'imensions ( mrrg d from the factor
analysis solution. Th :se seven components explain 77.9 per
cent of the total variance (Table 19). The higher percentage

61
of explanation over the 1950 period is due to the additional
dimension and a higher percentage accouited for by the other
dimensions. A check of the communalities reveals that the
higher percentage of explained variation is due to a better
factoring of the variables, including old age and unemploy¬
ment, which were poorly accounted for in 1950 (Table 20).
Table 19
Percentage of Total Variance Explained by Each
Dimension in I960
Dimension
Tot; 1
Eigenvalue
Per cent of total variance
I
8.80
33.8
II
3.44
13.3
III
2.40
9.2
IV
1.86
7.2
V
1.47
5.7
VI
1.1"3
4.4
VII
1.12
4.3
1
77.9
Table 20
Percentage of Variance (Communality) of Each of the 26
Variables Accounted for by All Seven Components in I960
Variable:
Communality in percentage
1
97.1
2
65.9
3
77.6
4
92.4
5
85.4
6
78.0
7
91.3
8
89.5
9
83.5
10
87.1
11
58.9
12
81.5
13
67.2
14
67.2

62
Table 20—Continued
Variables
Communality in percentage
15
16
17
18
19
20
21
22
23
24
25
26
73.1
74.8
68.6
65.1
83.9
75.5
91.6
95.4
94.2
57.2
54.2
69.2
The I960 Dimensions
The largest dimension, as in 1950, is labelled
urban industrial counties. It contributes 33.8 per cent of
the total variance. High positive correlations are found
between this dimension and population, value added, capital
expenditures, retail sales, wholesale sales, and poverty
(Table 21). The concentration of the poor in the industrial
area is not unexpected. As employment diminishes in the
rural areas due to increased mechanization and larger land
holding the only promise of jobs lies in the growing indus¬
trial and service sectors of the economy. Consequently
there is a large migration from the country to the cities.
At times the flow is too great to be absorbed at once into
the economy, resulting in temporary unemployment. On the
other hand, due to their lack of skills or color, many of
those who are absorbed by the labour market are engaged in
low paying occupations, especially in some branches of the
service sector of the economy. Consequently the faster a

63
city grows and the larger it becomes the larger the concen¬
tration of urban poverty. For as a city achieves a higher
rank in the hierarchy of urban systems, it assumes more
service functions, many of which are low paying. In addi¬
tion the expansion of urban functions often makes the city
more attractive to future immigrants; this then compounds
the problem of absorbing these migrants into the existing
labour force rapidly. Furthermore, if the supply of labour
greatly exceeds the demand, wages may remain low unless the
workers are unionized or covered by wage laws. This notion
is suggested by the county factor scores in Table 22. It
is observed that the cities ranking highest on the urban
industrial poverty dimension are also the largest in popula¬
tion (Figure 4).
Table 21
Dimension I Urban Industrial Counties in I960
Primary Variables
Factor Loadings
Total population .93
Per cent of the population urban .43
Value added by manufacturing .84
Total capital expenditures .73
Retail sales .80
Wholesale sales .80
Per cent of families earning between
$0 - $999 .80
Per cent of families earning between
$1,000 - $1,999 .90
Per cent of families earning between
$2,000 - $2,999
.91

64
Table 22
Counties with Extreme Factor Scores on Dimension 1
in I960
High Positive
High Negative
Fultan, Ga.
11.99
Houston, Ga.
-1
.17
Jefferson, Ala.
11.33
Chattahoochee, Ga.
—
.99
keck1enburg, N. C.
4.89
Clayton, Ga.
—
.97
KotSle, Ala.
4.13
Catoosa, Ga.
—
.92
Guilford, N. C.
3.63
Lafayette, Miss.
—
.86
Forsythe, N« C.
3.12
Hancock, Miss.
—
.75
Greenville, S. C.
2.90
Lumpkin, Gs .
—
.71
Chatham, Ga.
2.58
Stone, Miss.
—
.70
Charleston, S. C.
2.53
Columbia, Ca.
—
.70
Spartanburg, S. C.
2.48
Walthall, Miss.
.67
Gaston, N. C.
2.08
Candler, Ga.
—
.67
Richland, S. C .
1.90
Liberty, Ga.
—
.67
Hinds, Miss.
1.82
Watauga, N. C.
-
.67
De Kalb, Ga.
1.62
Lamar, Ga.
—
. 66
Etowah, Ala.
1.61
George, Miss.
—
.62
Jackson, Miss.
1.57
Oconee, Ga.
—
. 61
Buncombe, N. C.
1.51
Wayne, Ga.
—
. 60
Wake, N. C.
1.46
Polk, N. C.
-
. 60
The second most important dimension contributes 13.3
per cent to the total variance. This component is called
Rural Negro Counties as it relates to the poverty found in
black rural counties. It has high factor loadings on per
cent of the population nonwhite, per cent of the population
who have completed less than five years of school, per cent
of the population under five years of age and per cent of
farmers who are tenants (Table 23). The other factor load¬
ings algo emphasize the rural nature of this dimension. The
ranking of factor scores clearly indicates the differential
in levels of living between the white and nonwhite (Table
24). High positive scores are associated with predominantly

FIGURE JV
URBAN - INDUSTRIAL
COUNTIES
1960
Stondord Deviation
of Foctor Scores
lüif3
--------- -2
-1 to V1
Hwii

66
Negro counties. These counties are located in an arc that
runs from eastern North Carolina through the old cotton belt
to the Mississippi Delta where the heaviest concentration
is found. White counties with low factor scores are gen¬
erally found in the Appalachian region of the study area.
The distribution of the counties is similar to that found
in 1950.
The third dimension is also an indicator of rural
poverty but it is not related to color. This component
reflects the residual poverty left over from the sharecrop¬
ping era. This group of rural poor is made up of tenant
and small commercial farmers who have not been able to cope
with the changing trends in agricultura] production (Table
25). In many instances their deprivation is attributed to
the lack of education, capital and land resources. Although
this form of poverty is widespread (Table 26), Figure 5
indicates that it is most prevalent in the areas where share-
cropping prevailed.
Table 23
Dimension II Rural Negro Counties in I960
Primary Variables Factor Loadings
Per cent of the population nonwhite .82
Per cent of population that completed
less than five years of school .72
Per cent of population under five years
of age
Per cent of farmers who are tenants
Average value of farm (land and buildings)
.77
.74
.47

67
Table 23—Continued
Primary Variables Factor Loadings
Total value of farm products produced .35
Number of farmers whose off-farm income
is greater than farm income -.48
Per cent of dwelling units sound -.45
Table 24
Counties with Extreme Factor Scores on Dimension II
in I960
High Positive
High Negative
Tunica, Miss.
3.55
Ashe, N. C.
-1.92
Coahoma, Miss.
3.15
V/autauga, N. C.
-1.92
Bolivar, Miss.
3.10
Jackson, N. C.
-1.87
Washington, Miss.
3.09
Dare, N. C.
-1.83
Sunflower, Miss.
3.02
Chattahoochee, Ga.
-1.81
Leflore, Miss.
2.91
Avery, N. C.
-1.79
Sharkey, Miss.
2.66
Macon, N. C.
-1.79
Quitman, Miss.
2.44
T i shorningo , Mi s s.
-1.78
Humphre y s, Mis s.
2.23
Buncombe, N. C.
-1.77
Tallahatchie, Miss.
2.20
Towns, Ga.
-1.74
Lee, Ga.
2.04
Fannin, Ga.
-1.73
De Soto, Miss.
2.03
Marion, Ala.
Alleghany, N. C.
-1.71
Lowndes, Ala.
2.01
-1.66
Burke, Ga.
1.94
Mitchell, N. C.
-1.59
Pitt, N. C.
1.91
Darlington, S. C.
-1.58
Table 25
Dimension III Tenant and Small Farm Poverty in I960
Primary Variables
Factor Loadings
Per cent tenants . 38
Average value of farms -.38
Value of farm products produced .65
Number of commercial farms earning less
than $2,500 annually .89
Per cent of families earning between $0-$999 .45
Per cent of families earning between $1,000-
$1,999 .26
Number of farmers whose off-farm income is
greater than farm income
.65


69
Table 26
Counties with Extreme Factor Scorer on Dimension III
in I960
High Positive
Robeson, N. C. 3.37
Cullman, Ala. 3.32
De Kalb, Ala. 3.23
Hinds, Miss. 2.84
Panola, Miss. 2.62
Dallas, Ala. 2.58
Williamsburg, S. C. 2.56
Sunflower, Miss. 2.51
Orangeburg, S. C. 2.50
Horry, S. C. 2.49
Columbus, N. C. 2.48
Sampson, N. C. 2.47
Marshall, Miss. 2.25
Madison, Miss. 2.19
Marshall, A1a. 2.14
Anderson, S. C. 2.12
Florence, S. C. 2.08
High Negative
Fulton, Ga.
-2.71
Camden, Ga.
-2.53
McIntosh, Ga.
-2.40
Clinch, Ga.
-2.05
De Kalb, Ga.
-1.86
Jackson, Miss.
-1.80
Charlton, Ga.
-1.76
Peach, Ga.
-1.74
Mecklenburg, N. C.
-1.63
Muscogee, Ga.
-1.59
Bryan, Ga.
-1.58
Chatham, Ga.
-1.55
Currituck, N. C.
-1.46
Liberty, Ga.
-1.42
Dougherty, Ga.
-1.39
Lee, Ga.
-1.39
Jones, Ga.
-1.37
The fourth dimension (Locational Advantage in Rela¬
tion to National Market) has high correlations with distance
to New York and distance to a city of at least 500,000 per¬
sons (Table 27). Counties with extreme negative scores on
this dimension are generally concentrated in southwestern
Mississippi (Table 28). Other low factor scores are found
in central Alabama and northern Mississippi* These factor
scores, although they do to a certain extent reflect poverty,
reflect largely the proximity of these counties to large
regional centers such, as Memphis, Birmingham and New Orleans.
It is of interest to note that no large regional grouping
appears around Atlanta. The absence of such a grouping
around Atlanta may be attributed to the closer proximity
of New York. On the other side of the scale are found the

70
tobacco-producing and the furniture manufacturing counties
of North and South Carolina. Although these counties are
not in close proximity to regional markets they are close
to the national market. This favourable location is impor¬
tant in regard to their specialization in the growing of
tobacco and in the manufacturing of cigarettes and furniture.
Furthermore these commodities have been less susceptible to
mechanization than other enterprises such as cotton growing.
This is in contrast to southwestern Mississippi which suf¬
fers from being unable to compete effectively in the agri¬
culture market. As a result there has been a large rever¬
sion to forestland, especially with the coming in of large
pulp and paper companies. The surplus of labour resulting
from the dwindling agricultural base has resulted in a high
unemployment rate and out-migration.
Table 27
Dimension IV Locational Advantage in Relation to
National Market (New York) in I960
Primary Variables Factor Loadings
Distance In miles to a city of at
least 500,000 persons .72
Distance in miles to New York -.78
Pei1 cent of population 65 years of age
or over -.19
Median gross rent • .29
Total value of farm products produced .21
Per cent of population nonwhite -.14
Per cent of population unemployed -.17
Distance in miles from center of county
to nearest city of at least 50,000
.22

71
Tabic 20
Counties with Extreme Factor Scores on Dimension IV
in I960
High Positive Loadings
High Negative Loadings
Gilmer, Ga.
4.61
Pearl River, Miss.
-2.42
Horry, S. C.
2.52
Harrison, Miss.
-2.08
Florence, S. C.
2.43
Mobile, Ala.
-2.03
Columbus, S. C.
2.20
Pike, Miss.
-2.04
Dillon, S. C.
2.19
Hancock, Miss.
-2.03
Robeson, N. C.
2.13
Copiah, Miss.
-1.93
Darlington, S. C.
2.10
Tuscaloosa, Ala.
-1.92
Alleghany, N. C.
2.10
Lincoln, Miss.
-1.87
Marlsboro, N. C.
2.02
Hinds, Miss.
-1.79
Marion, S. C.
1.87
Amite, Miss.
-1.78
Yadkin, N. C.
1.85
Marion, Miss.
-1.76
Surry, N. C.
1.83
Wilkinson, Miss.
-1.73
Scotland, N. C.
1.81
Law'rence, Miss.
-1.71
Alexander, N. C.
1.79
Jefferson Davis, Miss.
-1.70
Carteret, N. C.
1.71
Walker, Ala.
-1.65
Nev,' Hanover, N. C.
1.71
Adams, Miss.
-1.62
Wilkes, N. C.
1.71
George, Miss.
-1.61
Dimension V labelled wealthy urbanized counties,
reflects urban prosperity. It has high factor loadings on
per cent of the population urban, per cent of population
completed high school, per cent of population with sound
housing, median value of owner-occupied homes and median
gross rent (Table 29). Counties which rank high on this
component are prosperous small cities such, as Macon, Albany,
and Columbus in Georgia, Huntsville, Alabama and Wake Forest
in North Carolina (Table 30). Other counties that have high-
positive scores are Cobb, Clayton and De Kalb in Georgia.
These counties contain some of the more prosperous suburbs
of Atlanta. Negative factor scores arc associated with
counties both rural and urban, that have a high percentage
of low income groups (Table 30).

72
Table 29
Dimension V Wealthy Urbanized Counties in I960
Primary Variables
Factor Loadings
Per cent of population urban .67
Per cent of population completed less
than five years of school -.52
Per cent of population completed high
school .80
Per cent of population with sound housing .80
Median value of owner-occupied homes .70
Median gross rent .75
Retail sales .39
Average value of farm (Land and buildings) .48
Table 30
Counties with Extreme Factor Scores on Dimension V
in I960
High Positive Loadings
High Negative Loadings
De Kalb, Ga.
4.12
Webster, Ga.
-2.45
Houston, Ga.
3.69
Jefferson, Ala,.
-2.43
Clayton, Ga.
3.52
Fulton, Ga.
-2.4 3
Dougherty, Ga.
3.31
Glascock, Ga.
-1.82
Harrison, Miss.
3.10
Quitman, Ga.
-1.74
Cobb, Ga.
3.08
Hancock, Ga.
-1.67
Muscogee, Ga.
3.05
Talbot, Ga.
-1.58
Madison, Ala.
2.91
Issaquena, Miss.
-1.58
Onslow, N. C.
2.84
Stewart, Ga.
-1.54
Choctow, Ala,
2.69
Banks, Ga.
-1.50
Clarke, Ga.
2.60
Wheeler, Ga.
-1,50
Orange, N. C.
2,35
Montgomery, Ala,
2.31
Wake, N, C.
2.23
Cumberland, N. C.
2.15
Camden, Ga.
2.01
The last two dimensions like those in 1950 are dif¬
ficult to identify. Dimension VI appears to be another
locational index. This component is related to national, re¬
gional and local market potentials. Table 32 suggests that

73
counties which rank high on this factor are located close
to urban centers. They tend to have lower unemployment
rates and a younger population than counties which have a
less favourable location.
The last component is difficult to name. Table 33
shows that the salient variables correlated with this factor
are old age and median value of owner-occupied dwelling units,
and value added by manufacturing. This dimension may isolate
areas of declining prosperity, suggesting the out-migration
of the younger and better educated labour force. Although
Table 34 is not conclusive, it implies that out-migration
may be the key to the explanation of this dimension.
Table 31
Dimension VI Locational Advantage in Relation to
National, Regional and Local Markets in
I960
Primary Variables Factor Loadings
.60
.83
.10
.19
Distance in miles from the center of the
county to a city of at least 90,000
Per cent of the labour force unemployed
Distance in miles from the center of the
county to a city of at least 900,000
Distance in miles from the center of the
county to New York
Per cent of the population 69 years of
age or older
-.10

74
Table 32
Counties with Extreme Factor Scores on Dimension VI
in I960
High Positive Loadings
High Negative loadings
Crawford, Ga.
2.08
Bolivar, Miss.
-3.85
Edgefield, S. C.
2.06
Washington, Miss.
-3.28
Guilford, N. C.
2.04
Pamlico, N. C.
-3.17
Baker, Ga.
1.84
Carteret, N. C.
-3.10
Hertford, N. C.
1.81
Clay, N. C.
-2.88
De Kalb, Ga.
1.76
Hyde, N. C.
-2.88
Wake, N. C.
Fayette, Ga.
1.74
Avery, N. C.
-2.72
1.73
Brunswick, N. C.
-2.56
Glascock, Ga.
1.71
Rabun, Ga.
-2.36
Webster, Ga.
1.71
Pearl River, Miss.
-2.31
Candler, Ga.
1.63
Jackson, Miss.
-2.12
Forsythe, N. C.
1.63
Adams, Miss.
-2.29
Jasper, Miss.
-2.02
New Hanover, N. C.
-2.00
Pitt, N. C.
-2.00
Table 33
Dimension VII Areas of Declining Prosperity-
in I960
Primary Variables Factor Loadings
Per cent of the population 65 years
of age or older -.89
Median value of owner-occupied
dwelling units -.54
Median gross rent .30
Value added by manufacturing -.31
Table 34
Counties with Extreme Factor Scores on Dimension VII
in I960
High Negative Loadings
High Positive
Berkeley, S. C
Liberty, Ga.
Hinds, Miss.
Beaufort, S. C
Cumberland, N.
George, Miss.
Camden, Ga,
Chattahoochee,
Onslow, N. C.
Jackson, Miss.
Charlton, Ga.
Loadings
1.77
1.73
1.71
1.70
C. 1.58
1.34
1.27
Ga. 1.25
1.24
I .16
1.16
Choctaw, Ala
Forsythe, H.
Baldwin, Ga.
Fulton, Ga.
Hyde, N. C.
Alleghany, N
-18.77
C. -1.38
-1.28
-1.21
-1.04
C. -1.00

75
Table 34—Continued
High Positive Loadings High Negative Loadings
Houston, Ga. 1.16
Mobile, Ala. 1.09
Charleston, S. C. 1.04
Marengo, Miss. 1.02
Muscogee, Ga. 1.00
Pearl River, Miss. 1.00
Comparison of the Two Time Periods
The purpose of examining two time periods is to
interpret changes through time, in the overall economic
growth of the southeastern United States, to analyze whe¬
ther there has been a change in the importance of the dimen¬
sions of economic health, and whether new factors have arisen
over the decade 1950-1960. Such a comparative study, however,
is limited to a descriptive analysis as tests of significance
between dimensions over two time periods have not yet been
derived.
The 1950 and I960 factor ana?yses revealed essentially
the same patterns of economic health, although in I960 an
additional dimension emerged. This dimension, however, con¬
tributed little to the explanation of the total variation.
In spite of this similarity important changes have occurred
in the rank order of the components, indicating that the
dimensions are not stable through time, and that changes have
occurred in the overall factor pattern. Similar results were
noted by King'1' in his study of Canadian cities and by Hoffman
2
and Romsa in their analysis of urban dimensions in the south¬
eastern United States.

76
Although dimension IV in 1950 (Rural White Counties)
and dimension II in I960 (Rural Negro Counties) are named
differently, they are considered to be similar as both are
indicators of the color problem. They have similar factor
loadings with the exception that the correlation signs are
reversed. Thus one dimension can be considered a reciprocal
of the other.
The changes in rank order are reflected in dimensions
II through V. Negro poverty has moved from fourth in order
of importance in 1950 to second in I960 and the per cent of
variation contributed has increased from 7.4 to 13.3 per
cent, an increase of 5.9 per cent. This change in the impor¬
tance of the Negro dimension implies that the Negroes have
not shared equitably in the economic progress of the South¬
east, during this decade. The decrease of 2.7 per cent in
explained variation of component II in 1950, tenant and small
farm poverty, has reduced it to third in order of importance
in I960, The decrease in this form of rural poverty is in
accord with the mechanization end the consolidation of farm
units that has occurred in the Southeast. In many instances
this process of consolidation and mechanization has resulted
in a surplus labour supply. Agricultural unemployment is
responsible for the increased concentration of Negro poverty
in the Mississippi Delta. The mechanization of cotton farm¬
ing in the Delta has greatly reduced the labour requirements
and has either left the Negro unemployed, or has forced the

77
excess labour supply to migrate elsewhere, particularly to
the North. This similar process has been at work in south¬
ern Georgia, southern Alabama and southwestern Mississippi,
where the pulp and paper companies have become established
and have reverted large areas of crop land to pine forest.
The shift of dimension III in 1950 (wealthy urban¬
ized counties) to dimension V in I960 is an indication that
there has been a narrowing of the range in prosperity
between cities. This is interpreted as demonstrating that
urban centers have shared in economic growth more equitably
than rural areas. This is shown, by the increase in rural
poverty in certain sections of the study area as expressed
by components II and IV in I960. In 1950 these components
accounted for 13 per cent of the total variation while in
I960 they accounted for 20.5 per cent.
The locational factor relating to the national
market has moved in this time period from fifth place in
1950 to fourth place in I960. On the other hand the regional
and local market locational index has remained stable not
only in rank but it also only registered a slight increase
in total explained variation. This may be the result of an
improved transportation network that has made these markets
more accessible. The new dimension that emerged in I960 is
not only difficult to identify but it appears to have con¬
tributed little to the overall analysis.

78
From these dimensions it is clear that the variance
in economic health throughout the study area in 1950 and
I960 is a function of urbanism, color, location, industrial¬
ization, and rural poverty. It is observed that the rate
and form of urban growth is most important, for not only do
the urban areas harbour the majority of the region's poverty,
but also influence to some extent the growth rates of their
hinterlands. It is suggested that the increased rural pov¬
erty that is arising in certain sections, due to increasing
technology, fell become of greater concern unless an outlet
is found for the surplus supply of labour.
Regression Analysis
The major underlying patterns of economic health
factored from the original 26 variables can be tested for
hypotheses by the use of a regression model. As regression
analysis requires normal and independent data, the testing
of the normality of the factor scores and the independence
of the dimensions is of theoretical importance. In addition
it is of interest to ascertain how well these dimensions
explain the variation in poverty throughout the study area.
To test these and additional hypotheses a dependent variable,
median family income, was chosen as an index of poverty.
This variable was then regressed on the dimensions in a
stepwise regression.

79
The Hypotheses
Since the dimensions are theoretically representative
of the original data, it is hypothesized that a high coef¬
ficient of determination between the dimensions and median
family income will be obtained. It is hypothesized that the
3.960 dimensions will result in a higher coefficient of deter¬
mination, as they have higher communalities than the 1950
components. In addition, hypotheses in regard to the expected
correlations between the dimensions and the dependent variable
can be tested. The expected relationships are shown in Table
35.
Table 35
Expected Correlations Between Median Family
Income and the Derived Dimensions
Median Family Income
Urban Industrial Counties
Tenant and Small Farm Poverty -
Wealthy Urbanized Counties 4-
Rural Negro Counties -
Locational Advantage in Regard to
National Market L
Locational Advantage in Regard to
Regional and Local Markets +
Areas of Dec3.ining Prosperity
Rura.1 White Counties •{
The analysis verified the normality of the factor
scores and the independence of the dimensions (Tables 36,
37, 39 and 40). The hypothesis regarding the coefficient of
determination was found to be correct (Tables 38 and 41).
However, the correlations between the dependent and the

80
independent variables were found as expected with one
exception, Urban Industrial dimension had a positive cor¬
relation instead of a negative correlation as expected
(Tables 37 and 40).
Table 36
Normality of 1950 Factor Scores
Variable
Mean Standard Deviation
Median Family Income
Urban Industrial Counties
Tenant and Small Farm Poverty
Wealthy Urbanized Counties
Rural White Counties
Locational Advantage in Regard
to.National Market
Locational Advantage in Regard
to National, Local aid
Regional Markets
1445.38525
-0.00002
0.00003
0.00022
-0.00013
-0.00000
-0.0004
572.95850
.99995
.-99997
.99997
.99997
.99997
.99995
Table 37
1950 Correlation Matrix
Variable 1 2 3 4 5 6 7
1
Median Family
Income 1.000
0.365
-0.124
2
Urban Industrial
Counties
1.000
-0.000
3
Tenant and Small
Farm Poverty
1.000
4
Wealthy Urbanized
Counties
5 Rural White Counties
6 Locational Advantage in Relation
to National Market
7 Locational Advantage in Relation
to National, Regional and Local
0.623
0.334
0.301
0.212
0.000
-0.000
0.000
0.000
0.000
-0.000
-0.000
-0.000
1.000
-0.000
1.000
-0.000
0.000
-0.000
0.000
1.000
0.000
Markets
1.000

81
Table 38
1950 Summary Table
Step
Number
Variable
Entered
Mult
R
iple 9
R¿
_ . 2
Increase in R
1
4
0.6232
0.3884
0.3884
2
2
0.7223
0.5218
0.1334
3
5
0.7959
0.6334
0.1116
4
6
0.8510
0.7243
0.0909
5
7
0.8770
0.7692
0.0449
6
3
0.8857
0.7845
0,0153
Table
39
Normality of I960
Factor Scores
Variable
Mean
Standard Deviat
Median. '
Family Income
3072.64746
986.50854
Urban Industrial
Counties
-0.00003
.99992
Tenant ¡
and Small '
Farm Poverty
0.00007
.99998
Wealthy
Urbanized
Counties
0.00018
.99995
Rural Negro Poverty
0.00013
.99999
Locational Advantage in Relation
to National Market
0.00002
.99995
Loca.tional Advantage in Relation
to National, Regional and
Local Markets
-0.00006
.99996
Areas of Declining Prosperity
-0.00011
.99995
Table 40
I960 Correlation Matrix
Variable
1
2
7
1 Median Family
Income 1.000 0.349 -0.363 -0.187 0.046 0.695 0.186
2 Urban Industrial
Counties
1.000 -0.000 0.000 0.000 -0.000-0,000 -
1.000 0.000 -0.000 -0.000-0.000
1.000 -0.000 0.000 -0,000
3 Rural Negro
Counties
4 Tenant and Small
Farm Poverty
5 Locational Advantage in
Relation to National Market 1.000 -0.000 0.000
6 Wealthy Urbanized Counties 1,000 -0.000
7 Locational Advantage in Relation to
National, Regional and Loca] Markets 1.000
8 Areas of Dec]ining Prosperity
8
0.042
0.000
-0.000
-0.000
0.000
-0.000
0.000
1.000

82
Table 41
I960 Summary Table
Step
Number
Variable
Entered
R
Multiple 9
rr
Increase in
1
6
0.6954
0.4836
0.4836
2
3
0.7844
0.6153
0.1316
3
2
0.8584
0.7368
0.1215
4
4
0.8785
0.7717
0.0349
5
7
0.8979
0.8061
0.0344
6
5
0.8991
0.8083
0.0022
7
8
0.9001
0.8101
0.0018
Table 42
Comparison of the Beta, Correlation Coefficients
and the Coefficient of Determination Betv/een the
Two Time Periods
1950
I960
Beta
R
R2
Beta
R
r2
x
X
-X
X
Urban Industrial
209.26
.365
.3 33
343.94
. 349
.121
Counties
Tenant and Small
X
Farm Poverty
-70.92
-.124
.015
-184.26
-.187
.034
Wealthy Urbanized
X-
X
X
Counties
357.09
.623
.388
686.11*
.695,,
.483
Rural Negro Counties
—
—
-357.91
-.363
.131
Locational Advantage
in Relation to
the *
*
*
National Market
172.72
. 301
.090
45.80
.046
,002
Locational Advantage
in Relation to
National
)
Regional and Local *
-X-
„ *
Markets
121.44
.212
.044
183.09
. 136
.034
Areas of Declining
„ X
Prosperity
—-
—
—
-41.85
-.042
.001
Rural White
X
X-
Counties
191.42
. 334
.111
—
—
*—
-x
Significant at .05 level.
The Beta and Correlation Coefficients (Table 42)
indicate that the variance in median family income in both
periods was a function of location, color, industrialization,

83
urbanization and rural poverty. Dimension VII contributed
very little to the analysis. Although the urban industrial
dimension did not correlate negatively as expected, this
was probably due to the large percentage of high incomes
found in these urban industrial counties which tend to inflate
median family incomes, and thus to a degree hide the large
number of poor families in these counties.
There have, however, been changes in the significance
of these dimensions in explaining the differences in meojan
family income (Table 41). The influence of the locational
dimensions has dropped rapidly from 13.58 per cent of the
explained variation in 1950 to 3.62 per cent in I960. Almost
all of this loss is accounted for by the dimension represent¬
ing location in regard to the national market, as this com¬
ponent was not significant in the I960 regression analysis.
This may reflect the overall economic growth that was achieved
in the South during the two time periods, especially in the
cities. In contrast to the decrease in the importance of
location, urban growth played a larger role in I960. The
dimension representing urban wealth accounted for an addi¬
tional 10 per cent in explained variation. On the other hand
rural, poverty had a slightly higher negative correlation in
I960 than in 1950, indicating that in general rural areas are
today at a larger disadvantage. This also applies to the
Negro, who has lost ground during the time period studied.
This appears also to be borne out by the beta coefficients, with
the exception of the beta value, associated with location in

84
relation to the national market, as the beta coefficients
in I960 are substantially larger than those in 1950. This
is an indication that the income gap between the poor and
the rich lias actually widened in this time period.
Summary
The factor and regression analysis shows that the
urban-rural and the Negro-white dichotomy is increasing.
The analysis indicates that economic growth and opportunity
are being centered around urban complexes to the disadvantage
of rural areas. Furthermore, the economic gap between the
Negro and white segments of the population in this region
has not narrowed. These conclusions are in accord with the
theories regarding poverty and regional growth which were
examined in Chapter I.

REFERENCES
^'Leslie J. King, "Cross Sectional Analysis on
Canadian Urban Dimensions 1951 and 1961," Canadian Geographer.
X, No. 4 (1966), 205-224.
'Wayne Hoffman and Gerald Romsa, "Some Temporal Con¬
siderations of Basic Urban Dimensions in the Southeast,"
Paper read before the meeting of the Southeastern Division
of the Association of American Geographers, Greenville, North
Carolina, November 26, 1968.
85

CHAPTER IV
SUMMARY AND CONCLUSIONS
This chapter summarizes the findings of the factor
and regression analysis for two time periods of five south¬
eastern states, Alabama, Mississippi, Georgia, South Carolina
and North Carolina. These states were selected as they
constitute a socio-economic region and are rcpresentative of
the southeastern United States.
The prime objective of the study was to identify the
major dimensions of economic health of the study area for
two time periods 1950 and I960, and to ascertain if these
dimensions had c1 anged throu ;h tiT o. In the analysis, the
areal variation in economi - ]ei 1th was measured by using the
multivariate procedure of factor analysis, as the 26 variables
selected and used as indices of economic health did not meet
the statistical premises of normality and independence.
Furthermore, factor scores derived from factor analysis were
used in interpreting the areal pattern of the dimensions of
economic health., These same factor scores were also used in
a regression equation to explain the variance in median
family income.
86

87
1950 Analysis
The 1950 data were grouped into six dimensions which
represented 70.5 per cent of the total variance of the
original 26 variables. Dimension I was interpreted as being
an indicator of industrialization and urbanization. Counties
ranking high on this component contained the largest urban
centers in the study area. The highest factor scores were
associated with the larger and more rapidly growing centers
such as Atlanta, Birmingham and Charlotte.
Factor II was a low’ income agricultural component.
Variables characterizing this dimension revealed an agricul¬
tural structure which was remnant of the sharecropping era.
Many of the farmers, in the counties that were identified by
this component, are not able to compete successfully with
agricultural enterprises that use large inputs of land,
capital and managerial skill as substitutes for labour and
thus are able to operate on a larger scale. Counties fitting
this pattern were generally located in the old cotton belt.
The third dimension revealed a pattern of urban
counties. These counties contained smaller but wealthier
cities than the urban nodes represented by dimension I.
These cities contain a high proportion of residents who are
employed in certain high paying segments of the service sec¬
tor, such as education and the military. Other counties
which arc represented by this dimension contain wealthy
dormitory suburbs.

88
Factor IV was a component defining the largely
white population of Appalachia and the predominantly Negro
Mississippi Delta. This dimension although refl ecting
poverty in Doth cireas, identified the disparity in level of
education attained and land ownership between white and the
nonwhite segments of the population. The last two dimen¬
sions were a reflection of the effect of location in regard
to local, regional and national markets. No definitive pat¬
tern arose, although there is some empirical evidence that
poverty in southwestern Mississippi is due to its disadvan¬
tageous location.
On the basis of the derived dimensions, areal pat¬
terns of economic health were derived. In general a picture
arose of scattered peaks representing counties with a high
proportion of urban residents and or industrial development
that are surrounded by a sea of poverty. In the northern
portion of the study area are the poor whites of Appalachia.
In. the Mississippi delta poverty is largely comprised of
Negro agricultural workers, many of whom are unemployed.
Poverty in southwestern Mississippi appears to be a reflec¬
tion of isolation. Throughout the entire area are found the
poor tenant and small farmer.
Factor scores obtained from the analysis were util¬
ised as independent variables in a regression model to test
the significance of the dimensions in explaining the varia¬
tion in median family income. The regression analysis did

89
support the areal pattern of economic health as identified
by the six dimensions. The six factor scores obtained from
the factor analysis explained 78.49 per cent of the varia¬
tion in median family income. Median income was found to be
positively correlated with urbanization, industrialization,
and the white portion of the population. The three dimensions
representing these indices accounted for 63.34 per cent of the
explained variation. Location and agricultural poverty con¬
tributed smaller amounts to the total coefficient of deter¬
mination.
The beta coefficients also substantiated the correla¬
tion coefficients. Dimension III liad a beta value of 397.09
implying that a one unit increase in this dimension -would
result in an increase of $397»09 in median family income.
Since this dimension is largely a reflection of education,
it bears out the economic value of public investment in the
schools of this region. A one unit increase in the urban
industrial dimension would result in an increased median
family income of $209.26. Being white or .located close to
urban centers resulted also in higher incomes. Agricultural
employment, however, resulted in lower incomes as did a
decrease in the white dimension.
I960 Analysis
The I960 data were reduced by factor analysis to
seven dimensions, which explained 77.9 per cent of the varia¬
tion in economic health. The increase in explained variation

90
over the 1990 period was due to the derivation of an addi¬
tional component and a more complete factoring of the
original variables. Except for dimension VII (areas of
declining prosperity) the components obtained for this time
period were similar to those derived in 1990, although their
order of importance had changed.
Factor I as in 1990 was associated with urban
industrial counties. It had similar factor loadings and
was identified with counties containing the larger urban
centers. The second component pointed out the economic
discrimination against the Negro. This factor differentiates
between Negro poverty, of which a large proportion is con¬
centrated in the Mississippi Delta, and the poverty of white
Appalachia. Although the poverty of Appalachia has been
recognized by various levels of government through different
aid programs, little has been done for the Delta area.
Dimension III represents rural poverty found
throughout the Southeast. Rural poverty is found among
both the white and the Negro operators of small land hold¬
ings or tenants.
Dimension IV, a locational index, reflects the
effect of proximity to urban markets. The areal pattern
of factor scores although difficult to interpret suggests
that Mississippi and Alabama suffer by not having closer
access to the national ma'cket.

91
Dimension V is another urban component. It repre¬
sents smaller prosperous cities that have benefited from the
location of high paying service functions and wealthy resi¬
dential suburbs. In general this component reflects the
well-educated labour force found in these urbanized counties.
The last two dimensions are difficult to interpret; however,
they seem to represent a locational index and possible areas
of declining prosperity.
When the factor scores of the above dimensions were
regressed upon median family income for I960 a coefficient
of determination of 81.01 was obtained. The most significant
dimensions in order of importance were wealthy urbanized
counties, urban industrial counties and rural Negro counties.
These three components together explained 73.68 per cent of
the total variation in median family income. Less signif¬
icant were tenant and small farm poverty and location in
regard to local, regional and national markets. Dimensions
IV and VII were found to be not significant.
The beta values were in accord with the correlation
coefficients. For example, the component representing
wealthy urbanized counties which had the highest correlation
also had the highest beta value (686.11). Thus a one unit
increase in this dimension would result in an increase of
$686.11 in median family income. Conversely a one unit
increase in the second most significant dimension, rural
Negro counties, would result in a decrease in income of $357.91.

92
Comparisons of Changes in the Two
Time Periods Studied
The two time periods revealed similar dimensions of
economic health and essentially the same areal patterns when
factor scores were mapped. Although the factor loadings
varied slightly in the magnitude of the correlations between
the variables and the dimensions, the differences were not
sufficient to be considered of any significant value.
Despite the similarity, there were changes in the
rank order of the dimensions. The Negro component moved into
second place in I960, which is evidence that the nonwhite
population in the Southeast had not shared equitably in the
overall economic growth. The decrease in the importance of
tenant and small farm poverty is in line with the agricul¬
tural progress attained in the Southeast during this decade.
Similarly, the downward shift of the wealthy urbanized dimen¬
sion is in accord with the overall progress made in the
cities in regard to integration, wages and education. The
upward shift of the importance of the national market, from
fifth place to fourth, was unexpected. This perhaps is an
indication of the more complete integration of the southern
economy into the national economy.
The results obtained from the regression analysis
were even more stable than those obtained from factor anal¬
ysis with two exceptions. The national market had no sig¬
nificant effect on median family income in I960, implying

93
that since 1950 the differences in median family income in
the study area have narrowed. The beta values for the I960
time period were, however, substantially higher than those
obtained in 1950, an indication that the income gaps between
the poor and the more affluent had actually widened.
Conclusions
The results of the analysis support empirically the
hypotheses outlined in Chapter I. Economic health can be
adequately summarized by a small number of factors. As
expected, these dimensions did not remain stable in regard
to their rank order, for the components reflect the changes
in economic development and as such are indicators of change
Factor analysis furthermore revealed that the urban
industrial nodes are centers of growth. The counties that
were located in close proximity to tliese nodes were more
prosperous than those further away. This was substantiated
in the regression analysis. If family income is used as an
index of prosperity, then urban areas definitely provide
more and better opportunities for higher wages. Higher wage
provide a stimulus for migration to these cities. Continued
population growth results in an expanding consumer market
and consequently economic growth-.. The decrease in the impor
tunee of location relative io the national market may be due
to the improved transportation system in the Southeast or
it may bo due to the fact that the market in the study area

94
is now large enough to generate its own growth. However,
this increase in prosperity has not been achieved by a large
segment of the rural and Ilegro population. Economic growth
has been confined largely to the urban centers. Consequently
more effort must be directed towards the rural poor if this
region is to achieve the national level of median family
income.
The lag in change in the rural areas suggests that
social mores are more difficult to overcome than the lack of
capital and technological knowledge. Possibly this is a sign
of a slower rate of the diffusion of ideas in rural areas,
where there is less personal interaction and thus less pres¬
sure to change.
Future Besearoh Efforts
This study answers some questions and uncovers many
problems, The general purpose of the undertaking was real¬
ised. The underlying dimensions of economic health were
derived and broad areal patterns of economic regions were
identified. In addition the importance and the changes of
these dimensions for two time periods were established, not
only in regard to the variation in economic health but also
in how they affected the deviations in median family income.
The analysis furthermore revealed that there are subregional
variations in economic health.
The study is hampered somewhat in that it gives a
generalized account of the region. Further research should

S5
be directed to a subregional analysis. One step would be to
derive homogeneous economic regions based upon factor scores.
Such computer programs are now available but unfortunately
were not obtainable for this study. Additional factor and
regression analyses could then be performed so that compari¬
sons could be made between the subregions and between the
subregions and the study area. This would provide greater
insight into local problems concerning economic health.
Similar comparisons could be made between this region and
other regions. In addition geographers could study the rate
of diffusion and acceptance of new ideas in rural and urban
areas. The differences in the rate of diffusion and their
causes would be of aid in discerning why different levels of
economic growth and prosperity exist.
One problem that arises in doing a study covering
several time periods is the availability of comparable data,
as much of the census data can not be used as they were col¬
lected in different categories. Some of the variables used
in this initial attempt were redundant and should be replaced
by other parameters. Direct distances as indices of location
might be substituted for by market potential or transporta¬
tion cost indices. Another major obstacle to the study of
economic health in any region is the absence of psychological
measures. The rate and form of economic growth is a behavioral
response to perceived conditions. Research should be performed
on the behavioral aspects of regional growth in order to help

96
provide a more complete analysis of the reasons for the
geographic variations in economic health.

BIBLIOGRAPHY
Books
Ahmad, Qazi. Indian Cities: Characteristics and Correlates.
Chicago: University~of Chicago Press, lSi6"5T
Bogue, Donald J., and Beale, Calvin L. Economic Areas of
the United States. New York: The Pree Press of
Glencoe, Inc., 1961.
Cohen, Saul B. (ed.). Problems and Trends in American
Geography. New York: Basic Books, "inc., 1967.
Duncan, Otis Dudly, et al. Metropolis and Region. Baltimore
The John. Hopkins Press, 19o0.
Eaton, Clement. A History of the Old South. New York: The
Macmillan Co ., 1969m!
Herman, Louis A., Kornbluh, Joyce L., and Haber, Alan ( ed.).
Poverty in America. Ann Arbor; University of
Michigan Press, 1965.
Friedman, John and Alonso, William (ed.). Regional Develop¬
ment and Planning. Cambridge: Massachusetts Insti¬
tute of Technology Press, 1964.
Galbraith, John. The Affluent Society. Boston: Houghton-
Mifflin Co., 19*59.’
Ginsburg, Norton (ed.). Essays on Geography and Economic
Development. Chicago':’" The University of Chicago
Press, i960.
Griffin, Paul F., Young, Robert N., and Chatham, Ronald L.
Anglo--America: A Regional Geography of the United
STafes~"and lanada! Sari Francisco": FearbrT'Puo'J.isEers
lrTc“~T96H“.
Harmon, H. H. Modern Factor Analysis. Chicago; University
of Chicago Press, 1961.
97

98
Harrington, Michael. The Other America: Poverty in the
United States. 'Baltimore: Penguin Boo’ksT" 19dUT
Hart, John Fraser. The Southeastern United States. France-
ton: I). Van**KostrandH*Bo., Inc., '19'5’V.
Kendall, M. G. A Coarse in Multivariate Analvsis. London:
Charies Griffin7*l957.' “
Maddox, James G., et al. The Advancing South: Manpower
Prospects and "Problems' (New York: T?he Twentieth
C entury“P\md ,”1967").
Milder, Herman P. Poverty American Style. Belmont-:
Wadsworth PuBTisTiíñg Co., The., 1966.
Odum, Howard W. Southern Regions of the United States.
Chapel Hill: ihe University of“Ttorth"Carol‘ina Press,
1936.
Schlesinger, Arthur M, A History of American Life,, Vol.
XI: The Rise of'tile 071^1678-189 d New"York”""
The~l’5acmillan Co., 1949. *
Schultz, T. W. Economic Organization of Agriculture. New
York: McGraw Hill Co.,"’ T9*5"3.
Thoman, R. S. (ed.). Areas of Economic Stress in Canada.
Kingston: Queens University Press, YsHad
Vance, Rupert B. Human Geography of the South. Chapel Hill:
University of North Carolina Press,’ 1*935.
White, Langdon G. and Foscue, Edwin J. Regional Geography
of Anglo-America. 2d ed. Englewood Cliffs:
Pr entic ¿-Hall, 1956.
Artides, Reports and Periodicals
Bell, W, H. and Stevenson, I). W. "An Index of Economic
Health for Ontario Counties and Districts," Ontario
Economic Review, II, No. 5 (1964), 1-7.
Fuguitt, Glenn V. "County Seat Status as a Factor in Small
Town Growth and Decline," Social Forces, XLIII
(December, 1965), 245-251.
Herzog, Elizabeth. "Some Assumptions About the Poor," Social
Science Review, XXXVII (December, 1963), 390-410,

99
Hodge, Gerald. "The Prediction of Trade Center Viability in
the Great Plains," The Regional Science Association
Papers, XV (1965), By-llb.
__ . "Do Villages Grow? Some Perspectives and
Predictions," Rural Sociology, XXX, Wo. 2 (dune, iy66),
183-196.
King, Leslie J, "Cross Sectional Analysis on Canadian Urban
Dimensions 195-1 and 1961," Canadian Geographer, X,
No. 4 (1966), 205-224. “
Laird, William E. and Rinehart, James R. "Exogeneous Check
on Southern Economic Development," South Atlantic
Quarterly, LXV, No. 4 (1966), 491-508, '
Lassiter, Ray L., Jr. "Education for Males by Region, Race,
and Age," Southern Economic Journal, XXXII (July,
1965), 15-22. ~ ~~
Levitan, Lar A. "Programs in Aid of the Poor," Poverty and
Human Resources Abstracts, I (1966), 11-25,
MacDonald, Stephen L. "On South’s Recent Economic Develop¬
ment," Southern Economic Journal, XXVIII (July, 1961),
30-40.
Megee, Mary. "On Economic Growth and the Factor Analysis
Method," Southern Economic Journal, XXXI (July, 1964),
215-228.
Moon, Seung Gyu and McCann, Glenn C. Subregional Variability
of Adjustment Factors of Rural Families in the South,
Raleigh: ÑortH Carolina Agricultural Experiment*
Station, Southern Cooperative Series, Bulletin No. Ill,
January, 1966.
Nichols, William K. "Southern Tradition and Regional Eco¬
nomic Progress." Southern Economic Journal, XXVI
(January, I960), 18*7-19*8*.
Ratchford, B. U. "Economic Development in the South," South
Atlantic Quarterly, LXIV, No. 4 (1965), 496-505'.
Rummel, R. J. "Understanding Factor Analysis," Conflict
Resolution, XI, No. 4 (1967), 444-480.
Scott, John T., Jr. "Factor Analysis ¿aid Regression,"
Bconometrjca, XXXIV, No. 3 (July, 1967), 552-562.

100
Public Documents
Berry, Brian J. Strategies, Models and Economic Theories of
Development in Rural Regions. Agricultural Economic
Report No. 12'/. Washington: Economic Research
Service, U.S. Department of Agriculture, 1967.
Unpublished Material
Hoffman, Wayne and Romsa, Gerald. "Some Temporal Considera¬
tions of Basic Urban Dimensions in the Southeast,"
Paper read before the 23rd annual meeting of the
Southeastern Division of the Association of American
Geographers, Greenville, North Carolina, November 26,
1968.
Spinelli, Joseph. "A Study of Net Out-Migration in South¬
eastern Ohio, 1950-1960," Unpublished Master's
Thesis, Department of Geography, Ohio State Univer¬
sity, 1966.

BIOGRAPHICAL SKETCH
Gerald Harry Romsa, was born Hay 31» 1942, at Dolyna,
Ukraine„ In February, 1948, he immigrated to Canada» In
June, I960, he v/as graduated from Oakville High School. In
Hay, 1963, he received the degree of Bachelor of Science
from the University of Manitoba» From 1963 until 1964 he
served with the Department of National Health and Welfare
as a Food and Drug Inspector in Toronto. In 1964 he entered
the Graduate School of the University of Waterloo where he
received the degree of Master of Arts in 1967» From Sep¬
tember, 1966, when he enrolled in the Graduate School of
the University of Florida, until the present time he has
pursued his work toward the degree of Doctor of Philosophy,
Gerald Harry Rornsa is married to the former Mary
Fraser Reid and they have one son. He is a member of the
American Geographical Society, the Association of American
Geographers, and Gamma Theta Upsilon, honorary geography
fraternity.
101

This dissertation was prepared under the direction
of the chairman of the candidate's supervisory committee and
has been approved by all members of that committee. It was
submitted to the Dean of the College of Arts and Sciences
and to the Graduate Council, and was approved as partial
fulfillment of the requirements for the degree of Doctor of
Philosophy.
June, 1969
Dean, Graduate School
Supervisory Committee:
Chairman