Citation
Reapportionment and legislative outputs: a Florida case study. / by

Material Information

Title:
Reapportionment and legislative outputs: a Florida case study. / by
Creator:
Todd, John Richard, 1939-
Place of Publication:
[Gainesville]
Publisher:
University of Florida
Publication Date:
Language:
English
Physical Description:
xvi, 269 leaves : ; 28 cm.

Subjects

Subjects / Keywords:
Cities ( jstor )
Correlation coefficients ( jstor )
Correlations ( jstor )
Counties ( jstor )
Legislators ( jstor )
Legislature ( jstor )
Political parties ( jstor )
Political science ( jstor )
Taxes ( jstor )
Voting ( jstor )
Apportionment (Election law) -- Florida ( lcsh )
Dissertations, Academic -- Political Science -- UF ( lcsh )
Political Science thesis Ph. D ( lcsh )
Miami metropolitan area ( local )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 263-269.
Additional Physical Form:
Also available online.
General Note:
Manuscript copy.
General Note:
Vita.
Statement of Responsibility:
John Richard Todd

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
This item is presumed in the public domain according to the terms of the Retrospective Dissertation Scanning (RDS) policy, which may be viewed at http://ufdc.ufl.edu/AA00007596/00001. The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
13385426 ( OCLC )
0021880043 ( ALEPH )

Downloads

This item has the following downloads:


Full Text









Reapportionment and Legislative Outputs:
A :Forida Case Study

















By

JOHN RICHARD TODD













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






UNIVERSITY OF FLORIDA
December, 1971













AC KNOWLEDGMENTS


Many people have contributed to my intellectual develop-

ment and analytical abilities as a political scientist which

were required to complete this piece of research. The

individual who looms largest in my intellectual background

is the late Charles D. Farris. My own feeble capabilities

were greatly sharpened by association with this outstanding

political scientist. I regret that he died before this

could be completed.

In this particular research effort several people deserve

mention for the constructive roles they played. Dr. Manning

J. Dauer, chairman of my supervisory committee, first

interested me in this particular study. Moreover, he allowed

me to draw upon his years of accumulated expertise in Florida

politics in the task of making political sense out of the

statistical analysis. Dr. Alfred B. Clubok assisted by

providing continuing advice on methodology while the project

was being executed and by giving .a careful evaluation of the

finished product. The remaining members of my supervisory

committee, Dr. David Conradt, Dr. Thomas Henderson, and Dr.

Marvin Shaw, assisted by a. careful reading and by asking


ii









good questions which clarified and improved the final

product. I must, of course, hasten to absolve any of these

gentlemen from complicity in whatever errors exist in this

final version.

My wife Pat deserves commendation for her patience

throughout the whole ordeal and for her encouraging words

to keep me going.









































iii














TABLE OF CONTENTS



Page

ACKOWLEDGMENTS . .. . ii

LIST OF TABLES . . . . v

ABSTRACT . . . . . ... xiii

Chapter

I. INTRODUCTION: REAPPORTIONMENT
AND POLICY . . . . 1

II. SOCIAL4 CLEAVAGE AND LEGISLATIVE
DECISION-MAKING . . .. 34

III. POLICY OUTPUT ANALYSIS--I. .. . 132

IV. POLICY OUTPUT ANALYSIS--II . .. 225

V. SUMMARY AND CONCLUSIONS . . .252

BIBLIOGRAPHY . . . . . .263























iv














LIST OF TABLES


Table Page

II-1 Summary of subject matter issues involved
in roll calls with high loadings on
Factor I, House, 1963 .. . . 47

11-2 Summary of subject matter issues involved
in roll calls with high loadings on
Factor II, House, 1963 . . ... .49

II-3 Summary of subject matter issues involved
in roll calls with high loadings on
Factor II, Senate, 1963 . . ... .50

11-4 Summary of subject matter issues involved
in roll calls with high loadings on
Factor V, Senate, 1963 . . . 51

II-5 Matrix of intercorrelations among measures
of roll call associations and loadings of
roll calls on selected factors, Senate
and House, 1963 . . . ... 54

II-6 Frequency Distribution of Guttman scale scores
produced by scale of roll calls loading high
on Factor I, House, 1963 . . .. 57

1I-7 Relationship of House 1963 scale scores
derived from Factor I to urban-rural and
regional differences . . .. .58

Ii-8 Frequency distribution of Guttman scale scores
produced by scale of roll calls loading high
on Factor V, Senate, 1963 . . .. 60

11-9 Relationship of Senate 1963 scale scores
derived from Factor V to urban-rural and
regional differences . . .. 61



v









LIST OF TABLES (cont.)


Table Page

II-10 Summary of subject matter issues involved
in roll calls with high loadings on
Factor I, House,1965 . . 54

II-11 Summary of subject matter issues involved
in roll calls with high loadings on
Factor IV, Senate, 1965 . . . 65

11-12 Matrix of intercorrelations among measures
of roll call associations and loadings of
roll calls on Factor I, House, 1965 . 67

II-13 Frequence distribution of Guttman scale scores
produced by scale of roll calls loading high
on Factor I, House, 1965 . . 68

11-14 Relationship of House 1965 scale scores
derived from Factor I to urban-rural
and regional differences . . 70

11-15 Summary of subject matter issues involved
in roll calls with high loadings on
Factor I, House,1967 . . . 72

II-16 Summary of subject matter issues involved
in roll calls with high loadings on
Factor I, Senate,1967 . . ... 74

11-17 Matrix of intercorrelations among measures
of roll call associations and loadings of
roll calls on Factor I, House and Senate,
1967 . . . . . 77

11-18 Frequency distribution of Guttman scale scores
produced by scale of roll calls loading high
on Factor I, House, 1967 . . .. 79

11-19 Frequency distribution of Guttman scale scores
produced by scale of roll calls loading high
on Factor I, Senate, 1967 . . ... 79



vi









LIST OF TABLES (cont.)


Table Page

1I-20 Relationship of House 1967 scale scores
derived from Factor I to regional,
urban-rural, and party differences . .. 81

II-21 Relationship of Senate 1967 scale scores
derived from Factor I to regional,
urban-rural, and party differences. . ... 82

II-22 Summary of subject matter issues involved
in roll calls with high loadings on
Factor I, House, 1969 . . ... 85

II-23 Summary of subject matter issues involved
in roll calls with high loadings on
Factor II, House, 1969 . . .. 86

11-24 Summary of subject matter issues involved
in roll calls with high loadings on
Factor IV, House, 1969 . . . 87

11-25 Summary of subject matter issues involved
in roll calls with high loadings on
Factor I, Senate, 1969 . . .. 88

11-26 Summary of subject matter issues involved
in roll calls with high loadings on
Factor II, Senate, 1969 . . . 89

11-27 Matrix of intercorrelations among measures
of roll call associations and loadings of
roll calls on selected factors, House and
Senate, 1969 . . . . 92

II-28 Frequency distributions of Guttman scale scores
produced on scales of roll calls loading high
on selected factors, House and Senate, 1969 95

II-29 Relationship of House 1969 scale scores
derived from Factor I and Factor IV to
regional, urban-rural, and party
differences . . . . 97


vii









LIST OF TABLES (cont.)


Table Page

11-30 Relationship of Senate 1969 scale scores
derived from Factor I and Factor II
to regional, urban-rural, and party
differences . . . . 99

II-31 House committee memberships as related
to region and urbanism-ruralism,
1963-1969 . . . . 107

II-32 Senate committee memberships as related
to region and urbanism-ruralism,
1963-1969 . . . . ... 107

11-33 Committee assignments per person, by
influence score and urban-rural,
House, 1963-1969 . . . 110

II-34 Committee assignments per person, by
influence score and region. House,
1963-1969 . . . . 111

11-35 Committee assignments per person, by
influence score and urban-rural,
Senate, 1963-1969 . . . 112

II-36 Committee assignments per person, by
influence score and region, Senate,
1963-1969 . . . . 113

11-37 Distribution of over/under-representation
scores for committee assignments by
influence score and by urban-rural,
House, 1963-1969 . . . 114

11-38 Distribution of over/under-representation
scores for committee assignments by
influence score and by region, House,
1963-1969 . . . . ... .115

II-39 Distribution of over/under-representation
scores for committee assignments by
influence score and by urban-rural,
Senate, 1963-1969 . . . 117

viii








LIST OF TABLES (cont.)


Table Page

II-40 Distribution of over/under-representation
scores for committee assignments by
influence score and by region,
Senate, 1963-1969 . . . 118

11-41 PerCent of committee assignments, total and
by influence score, for urban-rural group-
ings, House, 1963-1969 . . .. .120

11-42 PerCent of committee assignments, total and
by influence score, for regional groupings,
House, 1963-1969 . . . ... .121

II-43 PerCent of committee assignments, total and
by influence score, for urban-rural group-
ings, Senate, 1963-1969 . . ... 123

11-44 Per Cent of committee assignments, total and
by influence score, for regional groupings,
Senate, 1963-1969 . . . .124

III-1 Political Variables entered into Statistical
analysis of expenditures . . .. .140

111-2 Correlation and slope coefficients for
adjacent pairs of years of tax and
debt measures . . . . 144

III-3 Correlations between political variables
and tax and debt measures, 1963-1968 .. 146

11-4 Partial correlations between political
variables and tax and debt measures,
controlling for previous year's
expenditures . . . . 150

III-5 Changes in levels of debt and taxation in
Florida as related to trends of change
for all states . . . 153

III-6 Correlation and slope coefficients for
adjacent pairs of years of education
measures ... ..... ........ 156

ix









LIST OF TABLES (cont.)


Table Page

III-7 Correlations between political variables
and education measures, 1963-1968 . 157

III-8 Partial correlations between political
variables and education measures,
controlling for previous year's
expenditures . . . . 163

III-9 Changes in levels of education expenditure
in Florida as related to trends of change
for all states .. . . . 167

III-10 Correlation and slope coefficients for
adjacent pairs of years on welfare
measures . . . . 169

III-11 Correlations between political variables
and welfare measures, 1963-1968 . .. 171

III-12 Partial correlations between political
variables and welfare measures,
controlling for previous year's
expenditures . . . .. 176

III-13 Changes in levels of welfare expenditures
in Florida as related to trends of change
for all states . . . . 181

III-14 Correlation and slope coefficients for
adjacent pairs of years of health and
hospital measures . . . 184

III-15 Correlations between political variables and
expenditures for health and hospitals,
1963-1968 .. . . . . 185

II-16 Partial correlations between political
variables and health and hospitals
measures, controlling for previous
year's expenditures . . . .. 188



x








LIST OF TABLES (cont.)


Table Page

III-17 Changes in levels of health and hospital
expenditures in Florida as related to 191
trends of change for all states . .

III-18 Correlation and slope coefficients for
adjacent pairs of years of highway
measures . . . . 192

1-19 Correlations between political variables
and expenditures for highways, 1963-1968 .. 193

III-20 Partial correlations between political
variables and highway Measures, controlling
for previous year's expenditures . ... .196

III-21 Changes in levels of highway expenditures in
Florida as related to trends of change for
all states . . . . 199

III-22 Relationship between gasoline tax collections
and distributions in the ten most and least
populous counties, 1963 . . . 208

III-23 Relationship between gasoline tax collections
and distributions in the ten most and least
populous counties, 1969-1970 . ... .211

III-24 Relationship between state aid to education
and county population 1963 and 1968 . .. .215

IV-1 Degree and types of change in legislative
operations perceived by Florida
legislators as results of
reapportionment . . . .229

IV-2 Legislative perception of factors influencing
passage of constitutional revision and
executive reorganization . . ... .235

IV-3 Legislative perception of post-reapportionment
response to urban problems and factors
affecting response . . . 237

xi









LIST OF TABLES (cont.)


Table Page'

IV-4 City Administrator perception of
post-reapportionment response to
urban problems and factors
affecting response . . . 238

IV-5 Summary of Florida league of municipalities
legislative activities 1963-1969 .. . .241









































xii









Abstract of Dissertation Presented to the
Graduate Council of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of Doctor of Philosophy

REAPPORTIONMENT AND LEGISLATIVE OUTPUTS:
A FLORIDA CASE STUDY

By

John Richard Todd

December, 1971

Chairman: Manning J. Dauer
Major Department: Political Science

This study attempted to push further into the analysis

of policy output and its determinants. The studies based

on correlational analysis of public expenditures cast doubt

on the importance of reapportionment as a political variable

having influence on policy decisions. At the same time,

many of the analysts responsible for this type of conclusion

have suggested that there was more to be learned about this

question with studies of differing designs. The use of

before and after designs to relate changes in public policy

to changsg in political variables has been suggested by some.

This was the approach taken in this study. The Florida

legislature was examined before and after the fact of

reapportionment with the goal of locating changes in legis-

lative behavior which might be associated with this change

in political structure.


xiii








As a prerequisite to relating reapportionment to policy

in Florida, a. roll call analysis was carried out on four

sessions of that state's legislature (two sessions before

and two sessions after reapportionment). The purpose of the

roll call analysis was to see if cleavages underlying roll

call voting were such as to make it probable that policy

would be adjusted after reapportionment. Factor analysis

and Guttman scaling revealed that urban-rural cleavage was

important in three of the four sessions studied. The excep-

tion was the first session after reapportionment, which was

dominated by party voting. Thus, a. type of cleavage was

important during the study period which would make apportion-

ment a potentially significant political variable.

Policy output was analyzed in a two-pronged approach.

One approach involved analysis of taxing and spending

variables over the time period covered by the legislative

analysis. These taxing and spending variables were related

to political variables through correlation and regression

techniques. The analysis of levels of spending and taxation

in Florida showed no significant changes which could

unequivocally be related to the fact of reapportionment.

However, analysis of the distribution of state funds to local

areas did reveal that the formula for the distribution of

xiv








road funds to counties had changed in a, manner which was

beneficial to the more urban, populous counties.

The other part of the output analysis involved an

effort to assess the impact of reapportionment in non-

expenditure policies. These nonexpenditure policies were

studied through a mail questionnaire administered to legis-

lators and city officials. These data indicated that

although the legislators had made some real changes in the

amount of authority extended to cities (especially in the

second session after reapportionment) that the legislature

got a negative evaluation from cities. This negative

evaluation resulted from the city administrator's preoccupa-

tion with the failures of the legislature to provide adequate

fiscal assistance. The questionnaire also revealed that the

legislators believed reapportionment to be closely related

to the passage of what they regarded as their outstanding

legislative achievements in the post-reapportionment sessions.

The study suggests that the political significance of

apportionment may have been excessively downplayed by studies

limited to a correlation or regression analysis of expenditure

variables. It emphasizes the need to expand measures of

policy output beyond easily accessible expenditure data.

An additional point made by the study is that more explicit

attention needs to be given to the time dimension. That is,

xv








the study argues that variability across units cannot always

be substituted for variability over time. The need for

further analysis of the nature of political cleavages and

their relationship to policy output is highlighted by the

analysis of urban-rural cleavage in this study. Comparative

studies using some indicators of the nature of political

cleavages seem to be an important further step in the

analysis of politics and policy.





































xvi













Chapter I


INTRODUCTION: PREAPPORTIONMENT AND POLICY


In a bibliographic essay surveying recent developments

in the study of state and local politics, Herbert Jacob and

Michael Lipsky conclude that "the most marked innovation .

[in the field of state government] has been the investigation

of the relationship of policy outputs to social and economic

variables .. This systematic and comparative analysis

of policy outputs has had the kind of revivifying effect in

the field of state government that power structure studies

had in the field of local politics. It continues to attract

time and talents to its tasks.

The study to be reported here is an attempt to add to

this line of research centering on state politics and policy.

More specifically it is a study of the effects of reappor-

tionment on the behavior and output of the Florida legislature.

To set the framework and background for the report of the

analysis, several preliminary steps need to be carried out

in this chapter. First, to underscore the importance of

reapportionment to Florida politics we will outline the

history of the reapportionment controversy. Second, we will

1






2

briefly sketch the background of policy analysis and its

relationship to the reapportionment issue. Third, a pre-

liminary assessment will be made of the likely contribution

of a study of the kind undertaken here. Finally, an over-

view will be given of the various kinds of data needed and

the analysis techniques to be used in this study. We turn

now to the first of these stage-setting steps.

The Rea.ioortionment Controversy in Florida: An Histor-

ical Sketch--The point to be emphasized in this historical

overview is that reapportionment has been a persistent and

intense controversy in Florida politics. Malapportionment

has not been just a matter of the unnoticed silent gerry-

mander. Instead, it has been a highly visible issue, with

hotly contested debate. The politics of reapportionment has

commonly drawn mostof the political actors of the state into its

vortex. This fact may affect the relationship between policy

and apportionment in Florida.

Two scholars who have given probably the most attention

to understanding the malapportionment problem in Florida

suggest that the allocation of seats in the legislative body

has always been a problem in Florida. They note that "even

in the period before the state was admitted to the union

difficulties were involved in the apportionment of the

territorial council. 2 In order to examine that notion in






3

comprehensible bites we shall divide that long historical

period into three parts--from statehood to 1900, from 1900

to 1955, and from 1955 to 1967. We believe that these

divisions will give both the scope we desire and, at the

same time, focus our attention upon those events of the

reapportionment controversy most likely to be relevant to

our study period in the 1960's.

From statehood to 1900--Constitutional provisions have

always been intimately involved with the malapportionment

problem in Florida. The consitutions of the state have

tended to virtually guarantee that malapportionment would

be the outcome of any attempt to allocate legislative seats.

In addition, reapportionment has been accomplished most

commonly through changing constitutional provisions (either

by changing constitutions or amending existing ones) rather

than by statutory authority. We will, therefore, devote

some attention at the outset to the constitutional provisions

relating to apportionment.

The precedent for restrictive apportionment provisions

which prevent a genuine population-based apportionment was

set from the very beginning with the constitution of 1838.

This constitution, which served as the constitution when

Florida became a state in 1.845, set out guidelines for

apportionment whiich ultimately would make it impossible to






4

allocate enough members to the most populous counties to

give their citizens adequate repr:esentation. The document

(1) limited the total number of members of the House of

Representatives, (2) guaranteed at least one representative

to each county, (3) stipulated that senatorial districts

should be single-member districts, and (4) prohibited the

subdividing of counties for senatorial districts. The

guarantee of one representative per county and the prohibi-

tion against subdividing counties for single-member

senatorial, districts were sufficient to make population-

based apportionment impossible.

Although Florida had several constitutions in the

remainder of the nineteenth century--with new constitutions

in 1861, 1865, and 1885--the basic constraining principles

of apportionment laid down in the constitution of 1838

(or 1845) were not overturned. The constitutions of 1868

and 1885 substituted a limit on the maximum number of house

seats to be allocated to one county for a limit on the total

number of seats. Nevertheless, if the most populous counties

can, at most, have three or four members of the house and

all counties are guaranteed one member, rural over-

representation is virtually bound to occur. The provision

for single-member senatorial seats which must encompass at

least one whole county persisted throughout the nineteenth






5

century. This guaranteed a. malapportionment problem for

the senate. The senate provisions concerning apportionment

are even more striking when we note that it is clearer that

the constitution makers intended for senatorial districts

to be based on an equal population base than it is for house

districts.

Only once in the nineteenth century did the legislature

reapportion itself by statute. This came in 1887 in response

to provisions of the constitution of 1885 which virtually

required action by the legislature. The reapportionment then

was quite limited in effect. Every other instance of

reapportioning seats in this period came as a result of the

adoption of a new constitution. Each one had an apportion-

ment of seats written into it. While these were intended

only as temporary apportionments until the legislature could

act, with the exception of the 1887 reapportionment,the

legislature never acted to revise these temporary apportion-

ments.

From 1900 to 1955--In the period from 1900 to 1955 the

legislature submitted constitutional amendments dealing with

apportionment to the people of the state seven times--in

1915, 1921, 1923, 1941, 1943, 1947, and 1951. Yet, none of

these proposed amendments overturned the constrictive

provisio(ns of earlier constitutions. In fact, one of them--






6

the 1915 proposal--attempted to overturn the population

basis for the senate and substitute a more restrictive

provision based on the federal principle, i.e., one senator

per county. Of all these proposals only that of 1923 was

accepted by the voters at the polls. Its appeal seems to

have derived from the fact that it slightly increased urban

representation in the house, although it came nowhere near

an equal population basis.

During the 1900 to 1955 period the legislature enacted

reapportionment statutes twice. In 1925 it simply applied

the formula of the 1923 amendment to the House and modified

the Senate slightly. The Senate apportionment fell far

short of an equal population apportionment since the largest

district had over 130,000 people and the smallest had only

4700. The other case of statutory reapportionment in this

period came in 1945 when the governor used his constitutional

authority to call the legislature into an extraordinary

reapportionment session. After fifty-three days of desultory

meetings a bill was finally passed which the governor would

accept. Its effect was quite minimal.

From 1955 to 1967--In the period from 1955 until the

judicial resolution of the question in 1967, reapportionment

was a major political issue in the state of Florida. The

issue was pressed early in 1955 by Governor LeRoy Collins.






7

He acknowledged the significance of the reapportionment issue

by appointing a study commission during the period when he

was governor-elect to make recommendations to him on the

question. When the legislature met in regular session in

1955 it turned to the constitutional amendment route to deal

with the reapportionment question. The proposal they drew

up was in keeping with the existing constitutional provisions

for the House. The proposal for the Senate, however, included

the extreme notion of applying the federal principle rather

than the population principle. This amendment was rejected

by the voters in November, 1956.

In the meantime, Governor Collins refused to let the

legislature off so lightly. He called them into an extra-

ordinary session, which, by constitutional provision, could

deal only with the question of reapportionment and was bound

to continue until reapportionment was achieved. After

executive vetoes of two legislative proposals, it was clear

that nothing could be accomplished. With the governor and

the legislature at a stalemate, the session formally lasted

until the expiration of the legislators' terms in 1956.

(Actually the legislators recessed themselves, perhaps

unconstitutionally, after forty-five days.3)

The next attempt to move toward reapportionment came

in a special session in 1957 called by the governor to









consider constitutional revision. The legislature wrote its

own proposal for apportionment, ignoring the prior work of

the constitutional revision commission. Like earlier amend-

ments, this proposal represented only tinkering with the

system--no real reform. The usual constraining clauses of

the constitution preventing a population-based apportionment

remained intact. This amendment was never placed on the ballot

because the Florida Supreme Court found the procedure by which

it was to be presented unconstitutional.4

In 1959 the legislature proposed a new constitutional

amendment for reapportionment. One of the few distinguishing

characteristics of this proposal was that it would have

eliminated the equal population requirement from the consti-

tutional provisions regarding the senate. It indicated

clearly that economic and geographic considerations could

be included with population in apportioning seats in the

senate. Once again the proposal went down to defeat in the

referendum vote.

It its 1961 session the legislature tried again to

reapportion itself--still using the constitutional amendment

technique. The compromise proposal which they submitted to

the public for approval went a little further than previous

efforts, but was still inadequate. This amendment never was

presented to the people. The "reapportionment revolution"






9

began in earnest with the United States Supreme Court's

now famous a.er v. Carl decision in 1962. Since the court

had ruled that apportionment was a justifiable issue, an

immediate challenge to the pending constitutional amendment

in Florida was filed. In July, 1962, the federal district

court ruled that the apportionment provisions of the Florida

constitution were in violation of the federal constitution.

Thus, the 1961 proposed amendment was not presented to the
1 6
voters as scheduled.

The narrative from here on involves the plodding efforts

of a legislature unable to reapportion itself through

repeated efforts, even when the judicial gun was pointed at

its head. The legislature passed an inadequate amendment in

August, 1962 which the people rejected in November, 1962.

The governor recalled the legislature in November, 1962. It

deadlocked and adjourned after twenty days. The governor

called another special session in January, 1963. It came up

with a proposal which it enacted by statute and submitted to

the people in the form of a constitutional amendment. The

result was an anomalous situation in which the people rejected

the amendment at the polls in November, 1964 while the United

States district court had accepted the identical statutory

provisions as constitutional.7






10

Attempts.to invalidate the 1965 session of the legis-

lature failed, but the three-judge federal court ruled in

January, 1965 that the legislature had only until July 1,

1965, to reapportion itself in accordance with constitutional

standards. In a. special session in the summer of 1965 the

legislature presented a plan. The three-judge court accepted

the proposal, but the United States Supreme Court rejected

it. Once again the legislature was called into special

session. In March, 1966 it worked for a. week and came up

with a new plan which was closer to a genuine population-

based apportionment. In January, 1967, the United States

Supreme Court ruled that the plan did not seem, on the

surface,to meet standards of legality. It instructed the

lower courts to gather more evidence in its behalf or reject

it.

The Supreme Court's rejection of the 1966 proposal

found the legislature already in a special session called

to consider constitutional revision. When called upon by

the three-judge federal court to offer a new plan for

apportionment, the legislature made three responses. It

presented briefs defending the 1966 plan. New apportion-

ment plans were presented. A plan for weighted voting was

put forward. The court rejected all of these alternatives

and accepted the plan offered in an amicus curiae brief by






I1

Dr, Manning J. Dauer of the University of Florida's political

science department. The court ordered this plan into effect,

and after new elections Florida's first fairly apportioned

legislature met in 1967.

The struggle to this point--where legislative districts

are as nearly equal in population as practicable--has been

long and arduous. As our brief historical sketch has indi-

cated, there has never been a period in Florida history when

malapportionment was not a problem. Particularly in the

period since 1955 the reapportionment problem has been a

continuing, nagging issue occupying a central place in

Florida's politics. The frequent attention to this kind of

issue which clearly sets urban against rural interests may

have had its impact on the style and substance of Florida

politics. This type of concern relates to a style of policy

analysis important in political science today.

Overview of Policy Analysis and Reapportionment--To

place our study in the context of this ongoing research

enterprise, we need to review some of the research and find-

ings of the policy analysis work. We begin by noting that

the belief that the characteristics of a political system

and its socio-economic environment bear some systematic

relationship to the manner in which the system operates is

as old as Aristotle's Politics. Thus, the current interest






12

in the determinants of public policy is not new in terms of

the questions asked. Instead the contribution of current

policy research stems from the methods used. Much of the

traditional concern for such matters rested upon processes

of speculative inference and casual observation with quite

limited reference to real data. By contrast contemporary

state policy analysis shows great concern for careful

measurement of relevant variables and systematic analysis

of their interrelationships. Indeed, one indication of the

greater concern for measurement and testing is the level of

methodological bickering which has accompanied this current
10
research effort. The distinctive contribution of the

current state policy analysis, then, is an empirical base

and careful testing of an old line of hypotheses.

Much of the interest in comparative policy analysis

seems to have grown from seeds planted in V. 0. Key's
11
Southern Politics. As Key was summing up the factors at

work in Southern politics, he put forward what has come to

be called "the venerable Key hypothesis."12 Key suggested

that the unprogressive tone of Southern state policies

resulted from the lack of competitive party structures in

that region. He speculated that vigorous party competition

tended to orient politics around the struggle between the

economic haves and have-nots. A by-product of party






13

competition would be a more progressive (i.e., have-not

benefitting) set of policies. The key hypothesis provided

the basis for a line of research and testing which has not
13
yet culminated with the definitive test.3

Interest in party competition and public policy was

eventually to merge with an interest in malapportionment

public policy to form a line of analysis with the goal of

evaluating the independent effect of political variables on

public policy. The growing concern for malapportionment in

the decade of the 1950's, although primarily analyzed in

legal and ethical terms, fit into the policy analysis approach

because reformers frequently made allusions to presumed

policy consequences of malapportionment. These implicit

hypotheses in the lore of the reapportionment reformers

joined with the Key hypothesis to provide grist for the policy

analysts' research mills. Both inter-party competition and

malapportionment have thus become standard political vari-

ables to be included in studies of the determinants of state

policy.

We begin our review of the research with an early attempt

to assess the policy implications of malapportionment carried

out by Charles Shull.14 This piece blends aspects of the

reformer literature with an attempt to get at the consequences

of malapportionment. Shull sent out a brief questionnaire






14

to potential informants in the states. These informants

were made up of "legislators, public officials, political

scientists, and personal friends believed to have judgment

and the ability to exercise it."5 The set of consequences

of malapportionment reported by Shull were based on the

beliefs of these presumed expert judges.

Before the study of malapportionment and public policy

could move on to any quantitative and statistical approaches,

some measurement of degree of malapportionment had to be

devised which could order the states from worst to best

apportioned. One early attempt at such a measurement pro-

duced what is now commonly called the Dauer-Kelsay index.6

This measure, which is also called the minimal-majority

index, adds together legislative districts for each house

beginning with the least populous and moving upward until a

majority of the seats are accounted for. The index is that

percentage of the state population which resides in that

combination of districts which would make up a majority of

the house in question but which would minimize population.
17
The David-Eisenberg measure,7 is an attempt to assign a

score indicating the degree to which the urban vote is

devalued in a state as a. result of reapportionment. The

statistically most elegant measure to be devised is known

as the Schubert-Press index.8 However, this index performs






15

the dubious feat of combining scores for each house of a.

two chamber legislature. The index measures deviations from

certain statistical norms which would be expected to exist

under conditions of perfect apportionment. These efforts at

measuring malapportionment paved the way for later attempts

to relate apportionment systems to public policy.

As the last of the articles dealing with the appropriate

means to measure malapportionment was being published, one

of the first to include a statistical analysis of the rela-

tionship of malapportionment to public policy appeared. In

this article, Jacobl9 devised three of his own measures of'

malapportionment. The measure which he devised for the

legislative house based on population seems identical to the

Dauer-Kelsay index. Using rank-order correlation techniques,

Jacob tested for relationships between his measures of

malapportionment and a limited number of expenditure measures.

The results were negative. That is, he did not find a clear

relationship between a state's ranking on the malapportion-

ment scales and its ranking on the expenditure measures.

Jacob also included a simple test of the relationship of

malapportionment to such nonexpenditure policies as home

rule legislation and right-to-work laws. Again, he found

no strong evidence of a relationship to the degree of mal-

apportionment.








The next effort to evaluate the policy consequences

of malapportionment was published by Dye.20 All three of

the malapportionment indices which we have discussed were

used by Dye. This provided an opportunity not only to dis-

cuss malapportionment, but also a chance to comment on the

utility of the indices. To measure public policy Dye used

thirty variables--twenty-two education and welfare expendi-

ture measures and eight measures of taxation. He also

included socio-economic measures such as urbanization,

industrialization, income, and education. Dye found that

when the effects of the socio-economic measures were

partialled out that moderate correlations remained between

four of his education measures and the David-Eisenberg Index.

But, all other relationships disappeared.

The year following Dye's article Hofferbert published

a study in which he focused on two specific types of conse-

quences which had been associated with malapportionment in

reform literature: welfare orientation and fiscal aid to

cities.21 Using rank order correlation techniques Hofferbert

related the rankings of the states on the Schubert-Press

index to their rankings on a composite measure of welfare

orientation and on a measure of state aid to cities. He

found no significant relationship.







17

Brady and Edmonds, in a paper prepared the same year

as Hofferbert's article, tried to put to a systematic test

some of the relationships between malapportionment and policy
22
reported in the Shull article. Using the Schubert-Press

and David-Eisenberg indices of malapportionment, they found

quite weak correlations with their selection of taxing and

spending measures. They found that they could explain a much

larger percentage of the variance in the taxing and spending

measures by a multiple correlation using three socio-economic

measures. Like Hofferbert they investigated the distribution

of state funds to local governments. They used counties

rather than cities, however. They did not find any signif-

icant relationship between the way the distribution was

carried out and their malapportionment scores. In an effort

to explore relationships beyond expenditure measures of

policy, Brady and Edmonds carried out a contingency table

analysis relating malapportionment to the presence or absence

of such things as right-to-work laws, state income tax laws,

Kerr-Mills adoptions, and liquor monopolies. They rejected

the hypothesis that malapportionment was related to these

policies.

In later work which dealt with total levels of spending

as well as spending for specific programs, Sharkansky tried

to sort out the relationships between political and







18

socio-economic variables and expenditure patterns.23

Sharkans]ky's work has two distinguishing characteristics.

First, his expenditure variables were restricted to state

expenditure rather than the state-plus-local measures char-

acteristic of the work of Dye and others. Second, Sharkansky

gave attention to change in expenditure patterns. All three

of the standard indices of malapportionment were in his

analysis. He found no significant relationships between

malapportionment and his state expenditure measure.

A protest against the growing list of publications with

negative findings concerning malapportionment was published

in 1968 by Pulsipher and Weatherby.24 After reviewing the

literature questioning the policy significance of reapportion-

ment, these authors declared that their paper would be an

" attempt to furnish a grain of scholarly salt for
,125
assertions of this type.25 They used regression analysis

on a variety of expenditure, political, and socio-economic

variables. Unfortunately they threw away most of the informa-

tion yielded by the regression analysis. As Dye and Dyson

pointed out, "Nowhere in the article do the authors assess

the independent impact of their political variables on public

spending or compare the impact of political and socioeconomic

variables."26 Pulsipher and Weatherby chose instead to

perform significance tests upon the regression coefficient






19

showing the contribution of malapportionment to the equation.

This kind of approach mistakes statistical significance for

political significance. It yields only the information that

the malapportionment term in the equation is statistically

reliable. While our confidence in the reliability of the

regression term representing malapportionment may be increased,

we might suspect (on the basis of other research) that the

contribution of the malapportionment term to the regression

equation is quite small. Since the authors report only

unstandardized regression coefficients, direct comparison of

the terms in the equation is not possible.

Two years later Fry and Winters published a study which
27
focused on a different aspect of state policy output. They

took as their ". .. dependent variable the net redistributive

impact of revenues and expenditures as represented by the

ratio of expenditure benefits to revenue burdens for the

three lowest income classes in each state."28 This study

included the Schubert-Press index of malapportionment in its

battery of political variables. Both zero-order and partial

correlation coefficients indicated no relationship between

malapportionment and the amount of redistribution accomplished

by state expenditures.

One study with results at variance with the others

reported so far was Jack Walker's study of innovation among






20
29
the states.2 Walker's basic data were nonexpenditure

policies which had eventually been adopted by all of the

states. He calculated a score for each state based on the

amount of elapsed time between the first adoption of a

particular policy and that state's adoption. An overall

innovation score was calculated for each state by averaging

the scores for each policy included. Walker included the

Schubert-Press and David-Eisenberg malapportionment indices

among the political variables in his analysis. One of the

striking findings of his analysis was that the David-

Eisenberg index had a strong correlation with the innovation

score which did not disappear when socio-economic variables

were partialled out.

Hofstetter's is the last approach to assessing the

consequences of malapportionment which we want to mention.30

The basic data were roll call votes taken in the Indiana

House of Representatives between 1923 and 1963. The method

was to weight votes in such a way as to correct for malappor-

tionment. The effect of malapportionment could be estimated

by comparing these "corrected" votes with the actual votes.

Hofstetter concluded ". that numerically equitable

apportionment might not have had a large effect on the out-

come of roll-call votes and the passage of legislation."31






21

Thus, Hofstetter's results were congruent with the expenditure

analysis approach.

This review of some of the policy analysis literature

indicates that notions about the relationship of malapportion-

ment to public policy have been subjected to some considerable

amount of research scrutiny. The finding that there is

little or no relationship between measures of malapportion-

ment and public expenditure measures has virtually become a

commonplace of this type of literature. The Walker study,

however, forces us to leave open the possibility that some

policy significance can be attached to apportionment systems.

It encourages us to further investigation of the apportion-

ment problem.

Justification for this study--With the array of negative

findings concerning the policy relevance of apportionment in

the literature reviewed above, one might legitimately ques-

tion the need for further studies. We recognize, therefore,

that the burden is upon us to show why a study of the effects

of reapportionment is worthwhile.

The beginning point for a justification for our study

is found in the conclusions to some of the studies we have

discussed. In summing up what he accomplished with his

research Dye remarked that:







22

All that has been shown is that reapportionment
is not likely to have a. direct impact on party
competition or on certain policy outcomes. This
is not to say that reapportionment will have no
effect on state political systems or processes.
Quantification necessitates a simplification of
what may be a very complex question. The
consequences of reapportionment may be so subtle
and diverse that they defy quantitative measure-
ment. Perhaps the consequences in each state
will vary so much that direct interstate com-
parisons are inappropriate. Certainly we need
more refined analyses of the impact of appor-
tionment systems on state political processes
and policy outcomes; we especially need more
"before and after" studies of reapportionment.

Thus Dye does not hold to a blanket assertion that malappor-

tionment is of no consequence to politics and policy.

Sharkansky makes a remarkably similar comment. He says that:

The lack of significant correlation coefficients
between state government spending and the measures
of voter turnout, party strength, inter-party
competition, and the nature of the legislature
does not signify a lack of political influence
over.state budget-makers. Findings only show
that certain readily measured political char-
acteristics do not vary from state to state in
the same manner as expenditures.33

Thus, the authors of two of the more systematic studies of

state policy reject the notion that the study of the policy

significance of political structures such as apportionment

system has been exhausted by the correlation and regression

analyses completed. They leave room for further refinements

of our knowledge in this area.






23

Another opening for further research is provided by a

remark in the conclusions of Jacob's article. He notes that

it may be that the consequences of reapportionment appear
34
only under certain circumstances. This is in itself a

hint for a line of further research on this question--examina-

tion of conditions which might cause apportionment to be of

more or less policy significance.

If we gather together the hints from these researchers,

we pretty much have in hand a justification for the study of

reapportionment consequences in Florida. Dye and Sharkansky

both allude to the possibility of significant consequences

of political characteristics such as malapportionment which

might elude comparative statistical study. Moreover, Dye

suggests that one way of probing into these kinds of conse-

quences is through the use of before and after studies of

reapportionment. One such study of the Georgia legislature

reported by Brett Hawkins attests to the value of this

approach. It also suggests that such case studies may not

always confirm the comparative studies.35 Our study of

Florida is in the same vein as the Georgia study and has the

same intent. It is designed to look for those lines of

influence from apportionment to policy which have not been

picked up in existing comparative policy studies.






24

Another important justification for our study derives

from Jacob's comment concerning the possibility of the impact

of apportionment varying with some set of factors not

included in existing studies. We pursue that line of think-

ing by including an analysis of the basic lines of cleavage

underlying legislative voting in Florida both before and

after reapportionment. We hypothesize that malapportionment

is not likely to have important political consequences unless

political alignments are significantly influenced by urban-

rural differences. By including this analysis we expect

that our study may be able to make a worthwhile contribution

to our knowledge of the determinants of public policy.

The reasons why we believe this study worth doing can

be summarized as follows. First, we use a before and after

design, as suggested by Dye, to trace out the chain of events

surrounding the fact of legislative reapportionment. Thus

where previous studies have looked for covariation of mea-

sures of malapportionment and expenditure, we shall look at

a legislature undergo reapportionment and attempt to assess

the consequences of that fact. Second, pursuing Jacob's

speculation, we shall examine what we believe might be an

important prerequisite for apportionment systems to have

policy significance--the existence of real political align-

ments along an urban-rural basis. Several studies have made







25

it clear that assuming the existence of such urban-rural

36
alignments in all states would be quite misleading. We

believe that some additional light may be shed on the

subject looking first to see if urban-rural political align-

ments exist before looking at the consequences of reapportion-

ment. The third basis for justifying this study is that we

plan to go beyond expenditure measures to explore other types

of policy consequences. The Walker findings seem to make

this effort important. We believe that these three grounds

provide sufficient justification for this study.

Methodological Overview--With the history of reapportion-

ment issue in Florida sketched out, reviewed, and the

justifications for the study set out, we move now to the task

of sketching out the methods to be used in carrying out the

study. We might note that since several types of data will

be required and a variety of analysis techniques employed

we intend only to give a preview at this point rather than

a detailed methodological exposition. The detailed explana-

tions will accompany the relevant portions of the report.

We have noted above our hunch that whether or not

reapportionment has real political consequences depends on

the nature of the political alignments which structure

political issues and decision-making. This leads us rather

naturally to an examination of roll-call voting in the







26

legislature. If one of the important lines of political

cleavage in the state of Florida sets urban interests against

rural, then we should be able to find evidence of that-in

the way that legislators cast their votes.

A variety of means of analyzing roll call votes have

been used by political scientists. Two which have been

widely used are factor analysis and Guttman scaling. Since

our interest is in the broad cleavages underlying voting we

utilized factor analysis for a preliminary search. Once

factors were identified roll calls which seemed to tap an

underlying cleavage particularly well could be used to form

Guttman scales. These scales, in turn, allowed us to

classify individual legislators with regard to the underlying

dimension.

The data for our roll call analysis were collected from

four sessions of the Florida legislature--two preceding

reapportionment (1963 and 1965) and two following (1967 and

1969). For each of those sessions we included all roll calls

in which at least 20 per cent of those voting voted for

the losing side in our initial set for analysis. For tech-

nical reasons, which we shall explain in the report of the

analysis, it was necessary in some cases to cut down the

size of that set further. We tried always to keep a broadly

representative set of roll calls from each house for each

session.






27

Another legislative arena in which political alignments

can manifest themselves is in the committee system. We

looked at the committee system to see if it worked to the

advantage of any identifiable group and to see what implica-

tions reapportionment would have on the workings of the

system. We believed that the committee system was significant

enough to warrant the search for political alignments there

as well as in the roll call votes.

Our analysis of the policy consequences of reapportion-

ment will be divided into two parts. Part one deals with

expenditure variables and their relationship to reapportion-

ment. We believe, however, that the significant policy

actions taken by a legislative body are not exhausted by its

fiscal decisions. We therefore examine nonexpenditure

policies in addition to the expenditure matters.

To study change in expenditure patterns we collected

data on a variety of taxing and spending measures covering

the time period of our legislative analysis. We collected

these data for all fifty states. By regressing an expenditure

variable for one year upon the values of that variable for

the preceding year we can get a base line which we take to

represent normal change. It is then possible to examine

Florida expenditures on that variable over time to see if

they spurt ahead or lag behind this base line computed from






28

changes for all fifty states, We can then try to relate

those spurts to the fact of reapportionment, Additionally

we use correlation techniques to relate these expenditure

variables for all the states to certain crude measures of

political variables.

Our data on nonexpenditure policies are primarily

perceptions gathered through questionnaires. We asked the

producers of policy themselves--the legislators--to indicate

significant policies made and to evaluate th6 effect of

reapportionment in the passage of those bills. The other

group whose perceptions we were interested in was composed

of city administrators. Since state action vis-a-vis

municipalities has been generally regarded as being affected

by malapportionment, we thought it important to ask city

officials if they thought reapportionment had resulted in

policies more favorable to them.

The chapters following will be allocated to reporting

these research operations just described. Chapter two

reports our explorations into the nature of political

cleavage in Florida. That is, it covers the roll call

analysis and the committee system analysis. Chapter three

begins the policy output analysis by reporting our treatment

of expenditure policies in relationship to reapportionment.

Chapter four continues the emphasis on policy output and






29

reapportionment but shifts to nonexpenditure policy areas.

Finally, chapter five will provide a summary and conclusions.






30

Footnotes to Chapter I

"-Outputs, Structure, and Power: An Assessment of
Change in the Study of State and Local Politics," Journal
Lof Politics, XXX (May, 1968), 510-53. The quotation is from
pp. 551-12.

2William C. Havard and Loren P. Beth, Representative
Government and. eapportionment: A Case Study of Florida.
(Gainesville: Public Administration Clearing Service of the
University of Florida, 1960). Studies in Public Administra-
tion No. 20. The quote is from p. 23. This source forms the
primary basis for our account of the reapportionment con-
troversy through the 1950's. To avoid many repetitive and
relatively uninformative footnotes, we will restrict ourselves
to footnotes giving page citations for quotations or acknow-
ledging that we are drawing upon another source for this
secticn. Unless some alternative source is cited, the Havard
and Beth is our authority.

3Daisy Parker and H. Odell Waldby, "Twenty Years of
Struggle for Legislative Apportionment in Florida," Florida
iState UniversitJ Governmental Research Bulletin, II (May,
1935), 1-5. The point made here is on p. 2.

4The procedure involves the use of a, so-called "daisy
chain" clause which tied the success of the amendments
together so that if one failed all failed. The Florida
Supreme Court refused to accept this procedure.

5369 U. S. 186 (1962).

OParker and Waldby, pp. 2-3.

7Ilid., p. 3.

8Allen Morris, The Florida Handbook (Tallahassee:
Peninsular Publishing Co., 1969), p. 121.

Manning J. Dauer, "Florida Reapportionment," Business
and Economic Dimensions, III (March, 1967), 8-14. The
points covered here are on pp. 11-12.






31

10
SFor a sampling of the kind of methodolog ical critique
engaged in see Charles F. Cnudde and Donald J, McCrone, "Party
Competition and Welfare Policies in the American States,"
Anmerican Political Science Revliew, LXIII (June, 1969), 528-
529; see also John H. Fenton and Donald W, Chamberlayne,
"The Literature Dealing with the Relationship Between
Political Processes, Socio-Economic Conditions, and Public
Policies in the American States: A Bibliographic Essay,"
Polity, I (Spring, 1969), 388-394.

ll(New York: Alfred A. Knopf, 1949).

12Brian R. Fry and Richard Winters, "The Politics of
Redistribution," Arerican Political Science Review, LXIV
(June, 1970), 508-522. The quoted phrase appears at p. 521.

13Early extensions of the hypothesis to other regions
are John H. Fenton, Midwest Politics (New York: Holt,
Rinehart, & Winston, 1966) and Duane Lockard, New England
State Politics (Chicago: Henry Regnery, 1959). One of the
first studies comparing all states and casting doubt on the
validity of the Key hypothesis was Richard Dawson and James
Robinson, "Inter-Party Competition, Economic Variables, and
Welfare Policies in American States," Journal of Politics,
XXV (May, 1963), 265-289. Dye also concluded that party
competition had no independent effect on state and local
expenditure patterns. See Politics, Economics, and the Public
(Chicago: Rand McNally, 1966). Sharkansky reports a similar
finding using state spending only in Spendjing in the American
States (Chicago: Rand McNally, 1968). However, Cnudde and
McCrone, op. cit., argue that there is some relationship for
welfare policies.

14"Political and Partisan Implications of State Legis-
lative Apportionment," Law and Contemporary Problems, XVII
(Spring, 1952), 417-439.

15Ibid., p. 430.

16iVanning J. Dauer and Robert G. Kelsay, "Unrepresentative
States," National Municipal Review, XLIV (December, 1955),
571-575.

17Paul T. David and Ralph Eisenberg, Devaluation of the
Urban and Suburban Vote (Bureau of Public Administration,
University of Virginia, 1961).






32

18Glendon Schubert and Charles Press, "Measuring
Malapportionment," American Politiical Science Review," LVIII
(June, 1964), 302-327.

19Herbert Jacob, "The Consequences of Malapportionment:
A Note of Caution," Social Forces, XLIII (December, 1964),
256-261.
-0
20Thomas R. Dye, "Malapportionment and Public Policy in
the States," Journal of Politics, XXVII (August, 1965), 586-
601.

21
Richard I. Hofferbert, "The Relation Between Public
Policy and Some Structural and Environmental Variables in
the American States," American Political Science Review, LX
(March, 1966), 73-82.

22David Brady and Douglas Edmonds, The Effect of Malappor-
tionment on Policy Output in the American States (Iowa City:
University of Iowa Laboratory for Political Research, 1966).
23
23Ira Sharkansky, Spending in the American States
(Chicago: Rand McNally, 1968).

24Allan G. Pulsipher and James L. Weatherby, Jr.,
"Malapportionment Party Competition, and the Functional
Distribution of Governmental Expenditures," American Political
Science Review, LXII (December, 1968), 1207-1219.

25Ibid., p. 1208.

26Thomas R. Dye and James W. Dyson, "Communication to
the Editor," American Political Science Review, LXIII (June,
1969), 528-529.

27Brian R. Fry and Richard F. Winters, "Politics of
Redistribution," American Political Science Review, LXIV
(June, 1970), 508-522.

28bid., p. 508.

29Jack L. Walker, "The Diffusion of Innovations Among
the American States," American Political Science Review,
LXIII (September, 1969), 880-899.







33

30C. Richard Hofstetter, "Malapportionment and Roll-Call
Voting in Indiana," Journal of Politics, XXXIII (February,
1971, 92-111.

31Ibid., p. 109.

32Dye, "Malapportionment and Public Policy in the States,"
p. 600.

33Sharkansky, Spending in the American States, p. 64.

34Jacob, "The Consequences of Malapportionment," p. 260.

35Brett W. Hawkins, "Consequences of Reapportionment in
Georgia," in Richard I. Hofferbert and Ira Sharkansky (eds.),
State and Urban Politics: Readings in Comparative Public
Policy (Boston: Little, Brown, 1971), pp. 273-298.

36For articles denying the existence of meaningful urban-
rural political alignments see David Derge, "Metropolitan and
Outstate Alignments in Illinois and Missouri Legislative
Delegations," American Political Science Review, LII (December,
1958), 1051-1065; Thomas A. Flinn, "The Outline of Ohio
Politics," Western Political Quarterly, XIII (September, 1960),
702-721; and John G. Grumm, "The Means of Measuring Conflict
and Cohesion in the Legislature," Southwestern Social Sciences
Quarterly, XLIV (March, 1964), 377-388.













Chapter II


SOCIAL CLEAVAGE AND LEGISLATIVE
DECISION-MAKING


Much of the political significance of apportionment is

derived from its relationship to the problem of representa-

tion. In other words, it is necessary to assume that

malapportionment distorts the representation of social

interests to endow the problem with political--as opposed

to legal or ethical--significance. This notion was implied

by reapportionment reformers. They assumed that the con-

cepts "urban" and "rural" referred to two partitions of

society with important differences in policy preferences.

Or, as Robert Friedman puts it, "The assumption made .

in using urban and rural distinctions .is that the

terms provide a useful tool for distinguishing socio-

economic group interests which tend to congregate in the two

different types of environments."l

The assumption that rural people, who are overrepresent-

ed by malapportionment, are different from urban people is

important only if these differences affect the outcome of

legislative issues. Reformers assumed, therefore, that the

34







35

urban-rural cleavage underlay many of the significant

decisions taken by any legislative body. In turn, they

predicted that reapportionment would cause changes in legis-

lative behavior. They expected that the cleavage would

continue to have force in legislatures and that reapportion-

ment would shift the balance towards the urban side.

The assumed importance of the urban-rural cleavage as

a variable influencing legislative politics is the primary

focus for consideration in this chapter. We take two

approaches to uncovering the importance of this cleavage to

legislative decision-making in Florida. First, we under-

take a method of roll call analysis which has the capability

of uncovering sets of related roll calls, i.e., factor analy-

sis. If the urban-rural cleavage is important in the outcome

of the subset of roll calls, it is reasonable to assume that

factor analysis can help discover that subset. Second,

recognizing that the roll call vote is not the only signif-

icant act of decision-making in a. legislative body, we

examined the committee system of the legislature. We looked

for possible relationships between the allocation of commit-

tee assignments and the urban-rural cleavage. We recognize

that a definitive analysis of power in committees would

involve analysis of issues and the effect of an urban or

rural bloc on their disposition in committees. Since such






36

an extensive analysis was beyond our resources, we settled

for a positional analysis as a. cheap surrogate.2 The over-

representation of urban or rural legislators in influential

committees would seem especially convincing if accompanied

by a finding of bloc voting in the roll call analysis.

Roll Call Analysis

Methodoloqg--We turn now to the task of showing how the

techniques of roll call analysis can be put to service in

examining the place of urban-rural cleavage in legislative

behavior in Florida. We begin from the assumption that if

the same variable influences the outcome of a subset of roll

call votes, this should be expressed by a basic similarity

in the way the legislators distribute themselves between

the "yeas" and "nays" on the subset of roll calls. Other

roll calls, influenced by different variables, would show a

different distribution between the yeas and nays. Grouping

roll calls by the similarity of response to them is a

beginning point for uncovering the set of cleavages--or

dimensions--underlying a legislature's roll call behavior.

To obtain this kind of grouping we turned to the factor

analysis technique.

Factor analysis is a generic term which includes

several specific statistical procedures. The thing which

all of them have in common is, as Duncan MacRae has put it,






37

that they are concerned with ". simplifying the data

matrix and expression important aspects of it in terms of

a smaller number of underlying variables."3 Harman describes

over ten types of factor solutions in his well known work on

factor analysis.4 These various factor solutions can be

classified into two broad groupings on the basis of whether

the objective is to extract the maximum of variance within

the matrix with each factor or to maximally reproduce the

correlation matrix from which the factors are extracted. One

of the best known solutions which maximizes the variance for

each factor is principle components analysis.5 Harman also

distinguishes principle components analysis from principle

factor analysis. In the case of the former, unity is the

value assigned to the cells falling along the major diagonal

of the correlation matrix. Principle factor analysis uses

the communalities (the amount of variance in a measure

accounted for by the factor analysis) along the main diag-

onal. MacRae notes, however, that with large data matrices

such as we shall be using here there is little difference in

results between principle factor and principle components
7
methods. The principle components solution is the one which

we have used here. Orthogonal rotation of factors was used

to produce a set of uncorrelated factors, or summary

varriables, derived from the matrix of interrcorrelations

among roll calls.







38

Once such factors have been produced, the next step

is to search out their meaning. Factors derived from roll

call data are generally interpreted in terms of issues

(attitudes, values) or factions (parties, blocs, delegations,

etc.). As MacRae has indicated, "a purely empirical

approach cannot suffice to tell us the conditions under

which our inferences from roll calls are to be interpreted

8
in terms of issues or factions." In other words, a technique

such as principle components analysis provides a, classifica-

tion of roll calls by similarity of response patterns. It

does not tell us why these similarities occur.

There are several approaches to discovering the meaning

of factors. Perhaps the most common approach is to examine

the questions raised by the bills which occasioned the roll

call votes with high loadings on a given factor. Grumm

indicates that

the substantive content of a. subset of bills
each loaded highly on a given factor may
provide direct clues for factor definition.
What one looks for is a common element of
the subset that is not found in bills with
zero and near zero loadings. If this is
discovered, the factors may be defined in
terms of this element.9

This approach, however, seems most likely to be useful when

the underlying dimension is an issue or attitudinal con-

tinuum. It would probably not be very helpful if a group







39

allegiance were the operative force. We will, nevertheless,

look at the issue content of roll calls for the possibility

that an underlying urban or rural interest might be revealed.

Although this method may not lead to unambiguous identifica-

tion of factor meaning, if other methods are clearer in

identifying urban-rural factors we will then be interested

in looking back at the issue content of bills loading on

that factor.

Another method of finding the meaning of factors--

suggested by both MacRae and Grumm--is to examine the

correlates of the roll call votes loading on a given factor.0

We can take a set of roll calls which load highly on a

factor and examine the way the legislators responded (yea or

nay) in relationship to such things as urban-rural or

regional background. If we found that for all of the roll

call loading on a factor the yeas and nays varied system-

atically with urban-rural background (e.g., all urban

legislators consistently voted on one side and all rural

legislators consistently voted on the other) then we would

have strong grounds for inferring that the factor basically

tapped an urban-rural cleavage. To prepare for this kind

of analysis we calculated coefficients of association for

each of the roll calls in our analysis relating legislators'

votes to such background characteristics as urbanism-ruralism,







40
11 12
region,1 and party.12 Once factor loadings were available

from the factor analysis they could be related to the

coefficients of association. The factor loadings and

coefficients of association become data for correlation and

partial correlation analysis. Thus, we are able to express

in terms of a numerical value (i.e., a correlation coefficient)

the degree to which the roll calls which load highly on a

given factor have voting patterns closely associated with

urban-rural, regional, or party differences. If correlation

analysis indicates a high degree of covariation between

loadings on a given factor and coefficients of association

between votes and one of our background variables, then we

can infer that the factor analysis has grouped the roll calls

together on the basis of their relationship to that back-

ground variable. In other words, if we found a correlation

of .90 between the loadings of a set of roll calls on a

given factor and the coefficients of association between the

votes cast on those roll calls and urban-rural differences

we could be relatively sure that the factor was an urban-

rural factor. It is also possible to clarify some ambiguous

relationships by calculating partial correlation coefficients

which express the strength of the relationship between one

set of coefficients of association (e.g., between votes and

urban-rural) and factor loadings while partialling out the







41

effects of the relationslip between roll call votes and other

variables (e.g., partialling out correlations between party

associations and factor loadings).

A third method of uncovering the meaning of factors is

to assign factor scores or scale scores to individual respon-

dents--in this case legislators. MacRae indicates that "when

such a distribution separates two known groups such as

legislative parties or when it has intervals of lower density,

we [can] infer in either case that meaningful social groups

are being separated."13 The approach we take here is to take

the highest loading roll calls from the factor analysis and

submit them to Guttman scaling so as to derive the kind of

distribution MacRae described. We chose Guttman scaling

over factor scores for this purpose in the belief that the

important thing is to use those roll calls which are the

purest indicators of an underlying dimension to scale

legislators. This eliminates extraneous elements and error

variance which are likely to contaminate factor analysis

unless it is cleaned up.4 Because cleaning up a. factor

analysis (dropping out measures which contribute little to

the analysis and re-running) is expensive, we chose to

select good roll call indicators from the factor analysis

and Guttman scale them.






42

Guttman scaling, although originally developed for use

with attitude data, has become widely used in the area of

legislative roll call analysis. Two methods have commonly

been used in selecting roll calls to submit to scaling. One

method involves using the researcher's judgment to select a

preliminary set of roll calls which bear some apparent

relationship to one another. These roll calls are then

scaled, and perhaps some will be dropped if they do not fit

the scale pattern. Another method involves the use of

statistical procedures to select sets of scalable roll calls.

Most commonly this involves calculating some coefficient of

association (frequently Yule's Q) for all pairs of roll calls.

Closely associated roll calls can then be selected from a

matrix of such coefficients.5 Our approach is closer to

the second of these two approaches. We use factor analysis

to select sets of scalable sets of roll calls. These scales,

in turn, provide an ordering of legislators in relationship

to the underlying dimension.

Our analysis of group cleavages covers four sessions of

the Florida legislature. Two of them, 1963 and 1965,

occurred before reapportionment and two, 1967 and 1969, came

after. Since the Florida legislature met biennially until

the 1969 session, these are the four regular sessions which

bracket the historical fact of reapportionment. To avoid






43

biasing the sample of roll calls in our factor analysis, our

preliminary set included all roll calls taken in a session.

Prom this set we eliminated some roll calls in each session.

First, we eliminated roll calls with extreme marginals (i.e.,

those in which less than 20 per cent of those voting cast

ballots for the losing side). For some sessions this was all

of the elimination required (House and Senate, 1965 and House

and Senate, 1967). In other cases the limitations of the

available computer programs which performed factor analysis

served as an additional constraint on the number of roll calls

used. Two programs were used. For all except the House 1969

session the Bio-Med M03 program was used.6 However, the

large number of roll calls meeting our marginals criterion

in the 1969 House session (320) made a program with larger

capacity necessary. Using component subroutines from the

IBM Scientific Subroutines Packagel7 we were able to put

together a program which testing convinced us produced the

same results as the M03. This program had a maximum vari-

able capacity of 100. We thus adopted several additional

criteria in reducing the number of roll calls from 320 to

100. First, we further restricted the marginals requirement

so as to include only those bills which showed at least 35

per cent of those voting casting ballots on the losing side.

This, however, did not eliminate enough roll calls. Noting






44

the fact that this session had produced several examples of

multiple roll calls on a single bill, we decided to limit

to three the nuiber of roll calls included from a. single bill.

Within the three roll call rule, we chose to vary the

marginals by picking the roll calls with the highest, lowest,

and mean percentage voting for the losing side. Not only

did this procedure help to eliminate roll calls, but it also

provided an additional protection for the meaning of the

factor analysis. Cattell points out that loading a factor

analysis up with closely related measurements tends to

produce factors which are simply artifacts of the presence of
18
those measurements. The three roll call rule prevents the

production of factors which only demonstrate the fact that

multiple roll calls on a bill are likely to be related.

1963 SessionI--With these methodological guidelines

sketched out we can proceed now to an examination of the roll

call analysis. In both houses of the legislature in 1963 we

found it necessary to go beyond our beginning 20 per cent

dissent rule in order to reduce the number of roll calls to

a set which could be accomodated by the computer program

which we were using.19 By gradually increasing the percent-

age of dissent required for a roll call to stay in the

analysis, we reduced a set of 90 roll calls in the House and

100 in the Senate to 80 each which we submitted to the factor








analysis. In both cases the analysis produced a larger numbier

of factors with eigenvalues in excess of 1.00.20 We chose

(arbitrarily) in both cases to rotate only the first seven

factors with the expectation that this would exhaust the set

of significant factors. This proved to be a valid assumption.

In the house the first seven factors accounted for about 42

per cent of the variance. However, the first two factors

accounted for about 26 per cent of the variance (14 and 12

per cent respectively) with each of the succeeding factors

adding only about 4 per cent each. In the case of the Senate

these first seven factors accounted for 56 per cent of the

variance with the factors accounting successively for 17, 10,

9, 6, 5, 5, and 4 per cent of the variance.

The first step in beginning to sort out these factors

was to examine the structure of the loadings of the roll calls

on them. We selected .60 as the minimum loading which we

would regard as significant on any given factor. Although

this minimum was initially chosen arbitrarily to bring the

analysis into clearer relief, it was observed that for most of

our analyses the communality for an individual roll call

tended to drop below .50 at about the loading point of .60.

Since the communality measures the proportion of the variance

in a variable accounted for by the factors computed, this

cut off point meant that we examined initially only those







46

variables that were at least 50 per cent accounted for by

the factor analysis. By examining the roll calls loading

above the .60 we were able to see more clearly the pattern

of association between them and the factor structure. The

larger the number of such loadings, the more roll calls the

factor could account for alone.

In the House in 1963 only the first two factors appeared

to be important on the basis of the number of high loadings.

Respectively the factors had 12, 6, 0, 0, 3, 0, 0 roll calls

loading at the level of .60 or better. In the Senate the

pattern was somewhat different. The number of high loadings

per factor was 3, 5, 4, 1, 8, 2, 0. These results led us to

focus our attention on factors I and II in the House and

factors II and V in the Senate.

Having selected on the basis of loading structure two

factors from each house for further analysis, we moved next

to the examination of the content of the bills producing the

roll calls. Table II-1 presents a. listing of the roll calls,

in order of loading on Factor I, along with a brief descrip-

tion of the substantive content of each bill involved.

Although a varied set of subject matters seem to be involved,

there are some clues that this might be a factor involving

urban-rural differences. For example, the highest loading

bill deals with the subject of apportionment itself which is









Table 11-1
SUMMTARY OF SUBJECT MATTER ISSUES INVOLVED IN ROLL CALLS
WITH HIGH LOADINGS ON FACTOR I, HOUSE 1963


Roll Call Roll Call
Number Loading Description

36 .84778 Constitutional Amendment for apportion-
ment; referred to legislative apportion-
ment as well as constitutional amendment
committee (y=58, n=47)

70 -.78805 Amendment to allow the Milk Commission
to fix prices at the producer level
(y=54, n=59)

33 -.77879 Sets boundaries of Hendry, Martin,
Okeechobee, Palm Beach, and Glades
counties (y=54, n=55)

69 -.72957 Amendment to take away Milk Commission's
power to fix prices (y=44, n=70)

47 .70768 Constitutional amendment providing for
second gas tax (y=67, n=54)

53 -.69036 Appointment of administrator of Milk
Commission by Commissioner rather than
Governor (y=52, n=55)

60 -.65441 Repeal budgetary procedure under Board
of Control for education (y=58, n=61)

76 .63425 Excise taxes, severance of natural
resources: motion to refer to commit-
tee (y=47, n=53)

72 -.60874 Requirements and fees for registration
of barber shops (y=46, n=56)

35 .60611 Provides state and county retirement
for municipal service (y=44, n=73)

42 .60247 Constitutional amendment--non-exemption
of conservation laws (y=42, n=61)
37 .60135 Non-taxation of non-bearing fruit trees
and nursery stock until maturity (y=57,
n=47)






48

clearly an urban-rural issue. A se.ond clue is found in

the fact that variable five is concerned with the second

gas tax. As will be discussed below, this has been an issue

closely related to urban-rural differences. Primarily the

urban-rural relationship to the second gas tax derives from

the fact that the revenues are partially returned to the

counties on the basis of a formula which has been taken to

be biased in favor of the more rural counties. The 6

high-loading roll calls on Factor II, listed with subject

matter summaries in Table 11-2, seem on the basis of subject

matter alone to be urban type issues. However, since we

performed an orthogonal rotation of factors we would not

expect closely related factors to emerge. We leave the

identification of this factor to other steps in our analysis.

Of the factors from the Senate 1963 analysis only

Factors II and V had enough high loading roll calls to

provide a basis for subsequent steps of our analysis, so

we restricted our attention to them. The summary of bill

contents for Factor II is presented in Table 11-3. The

most striking fact is that 4 of the 5 roll calls are

on the same bill dealing with taxation of alcoholic beverages.

This factor may be an artifact of the type we mentioned above

which emerges because too many measures of the same thing

have been included. For this Senate session we eliminated






49

Table 11-2
SU MMARY OF SUBJECT M ATTER ISSUES ITVOLVED IN ROLL CALLS
WITH HIGH LOADINGS ON FACTOR II, HOUSE 1963

Roll Call Roll Call
Number Loading Description

13 .84296 Extends eminent domain and allows Key
West to secure federal aid for slum
clearance (y=69, n=35)

50 .80024 Slum clearance for Palatka (y=34,
n=47)

1 .79245 Urban renewal power for Sarasota
(y=69, n=28)

6 .78410 Slum clearance for Ormand Beach
(y=65, n=32)

65 .68683 Establish revolving fund for administra-
tion of federal urban planning (y=70,
n=35)

45 .65290 Memorial to U. S. Congress proposing
amendment to abolish income tax and
set ceiling of 25% (y=52, n=60)


roll calls only on the basis of percentage of dissent and

thus perhaps did not guard adequately against this possi-

bility. Since the factor seemed primarily concerned with

attitudes toward a single bill and preliminary analysis

indicated no significant associations between the votes on

the five high loading roll calls and variables with which

we were concerned, such as urban-rural or region, we dropped

this factor from further analysis. The remaining factor

from the 1963 Senate analysis--Factor V which is presented







50

Table II-3
SUMMAI lRY OF SUBJECT MATTER ISSUES INVOLVED IN ROLL CALLS
WITH HIGH LOADINGS ON FACTOR II, SENATE 1963

Roll Call Roll Call
Number Loading Description

58 .84553 Related to beverage law administration.
Restricts excise tax exemptions to
citrus products rather than agricul-
tural products (y=28, n=15)

48 .81942 Relates to Beverage law administration
of citrus products; provides for addi-
tional tax on certain beverages (y=29,
n=14)

49 .80567 Relates to beverage law administration
of Florida products; provides for
additional tax on alcoholic beverages
containing 14% or more alcohol (y=28,
n=15)

56 .65463 Provides for across the board increases
in taxes on beer, wine, liquor, etc.
Vote on amendment which lowers the tax
on alcoholic beverages over 48% that
are made from Florida citrus (y=21,
n=20)

59 .64717 Relates to Florida Agricultural Experi-
ment station; provides an appropriation
for research into an earthworm malady
(y=29, n=13)


with its high loading roll calls in Table II-4--presents a

more interesting potential. The subject matter is more

varied yet there are some roll calls which seem related to

the urban-rural cleavage. The second gas tax bill appears

here as it did in the House. This suggests urban-rural

voting patterns. Other approaches were required in order to

assure identification of the factor, however.







51

Table 11-4
SUMMARY OF SUBJECT iMATTER ISSUES INVOLVED IN ROLL CALLS
WITH HIGH LOADINGS ON FACTOR V, SENATE 1963


Roll Call Roll Call
Number Loading Description

1 -.82821 Redefines Boundaries of Hendry, Martin,
Okeechobee, Palm Beach, and Glades
counties (y=25, n=17)

41 -.77962 Constitutional amendment--increases
from 50 to 75 years the period during
which proceeds from the second gas
tax shall be placed in the State Road
Department Distribution Trust Fund
(y=30, n=15)

61 -.76815 Relates to beverage law administration
of citrus products--additional tax on
certain beverages (y=29, n=14)

30 .67960 Constitutional amendment--provides for
a county legislative council to pass
local laws. Motion to refer the bill
to Committee on Constitutional Amend-
ments only instead of both it and
Committee on Judiciary (y=19, n=21)

31 .65537 Changes Florida. statutes relating to
weapons and firearms (y=24, n=20)

65 .65537 Repeals penalty for carrying a pistol
or repeating rifles without license;
repeals procedure for obtaining a
license (y=24, n=20)

33 -.63755 Relates to use of lumber for construc-
tion--certain uses made unlawful (y=26,
n=15)

17 -.63653 Creates a committee to investigate
infiltration of homosexuals in state
agencies (y=30, n=14)






52

The next step in our path to factor identification

involved correlation of all roll call votes in our analysis

with independent variables. For this session we correlated

the votes on each of the roll calls with the urban-rural and

regional characteristics21 of the home county of the legis-

lators.22 We found these variables to have significant

relationships to many of the roll calls in the set. Of the

80 roll calls in the analysis, 21 showed a definite rela-

tionship to the urban-rural variable (i.e., produced chi

square values significant at the .05 level or better). The

regional variable showed a significant association with

17 of the roll calls. This kind of analysis in the 1963

House revealed that 40 of the 80 roll calls in our set

were significantly related to urban-rural differences and

29 were significantly related to regional background. Thus,

these variables certainly played a part in the outcome of

roll call votes in both the House and Senate in 1963.

Having demonstrated that urban-rural differences and

regional differences were related to roll call voting in

1963, we can proceed to see what might be added by relating

the pattern of these associations to the factor loadings

produced by our factor analysis. The coefficients of associa-

tion already calculated (showing degree of association between






53

roll call votes and urban-rural and regional differences)

were correlated with the loadings of the roll calls on the

factors we were examining. These correlations are shown in

matrix form in Table 11-5. The table shows that in the 1963

Senate all three sets of values for the roll calls are

correlated. Significantly, the loadings on Factor V show a

correlation of .78 with measures of association between roll

call votes and urban-rural backgrounds. A problem arises,

however, from the fact that these loadings also correlate at

.55 with measures of association between roll calls and

region. We were thus faced with trying to assess independently

the role of regional and urban-rural influences. We deal with

that problem below.

The roll calls in the House 1963 analysis share some of

the same relationships to our independent variables as those

demonstrated by the roll calls in the Senate analysis. The

loadings on Factor I show high correlations with roll call

associations with both urban-rural (.76) and regional (.74)

differences. The only thing which is made clear about Factor

II is that it is not explained by any of the variables

included in the table. The negative correlations, it should

be remembered, indicate that as loadings on Factor II

increase the association of individual roll calls with urbanism-

ruralism and region decreases. Whatever the nature of this







54

Table YTI-5
MATRIX OF INTERCORRELATIOr S AMONG MEASURES OF ROLL, CALL
ASSOCIATIONSa AND LOADI:NGS OF ROLL CALLS ON SELECTED
FACTORS, SENATE .AND HOUSE, 1963


Senate 1963
I II III
Roll Call Roll Call Roll Call
Association Association Loading on
with Urban- with Region Factor V
Rural

Ic 1.00 0.55 0.78

II 0.55 1.00 0.55

III 0.78 0.55 1.00


House 1963

I II IV III
Roll Call Roll Call Roll Call Roll Call
Association Association Loading on Loading on
with Urban- with Region Factor I Factor II
Rural

Ic 1.00 0.60 0.76 -0.47

II 0.60 1.00 0.74 -0.14

III 0.76 0.74 1.00 -0.39

IV 0.47 0.14 -0.39 1.00

aCoefficients of Association relate the "yeas" and "nays"
of each roll call vote to the urban-rural or regional breakdown.

Cell entries report Pearson correlation coefficients.

CRoman numerals on the vertical axis of the table refer
to the variables associated with the same Roman numerals on
the horizontal axis. The table thus shows all possible
correlations among the variables listed across the horizontal
axis.






55

factor, it cuts across the effects of regional and urban-

rural differences.

To try to improve our understanding of the relationship

between Factor I in the 1963 House and Factor V in the Senate

and association of roll calls to urban-rural and region, we

turned next to partial correlation analysis. In the House,

the original correlation between association with urban-

rural and loading on Factor I was .76. Controlling for

association of the roll calls with region reduced the coefficient

only to .58. By contrast controlling for association with

urban-rural reduced the correlation between regional associa-

tion and loading on Factor I from .74 to .21. Thus, the

partial correlation analysis points to identification of

Factor I as tapping primarily the effects of urban-rural

voting. A very similar pattern of partial correlations

appears in the analysis of Factor V in the 1963 Senate.

Controlling for roll call associations with regionalism

reduces the correlation between factor loadings and associa-

tions with urban-rural from .78 to .68. By contrast,

partialling out the effects of roll call associations with

urban-rural reduces the correlation between associations

with region and loadings on Factor V from .55 to .24. In

both cases, then, the pattern of partial correlations

suggests a. developmental sequence leading from regional







56

differences to urban-rural differences to roll call voting

patterns.23

While the relationships suggested by the partial correla-

tion analysis are intuitively persuasive, we pursued the

analysis further by the use of Guttman scaling. As indicated

above, one of the means of identifying the common factor tying

a set of measures together is by scaling those measures and

examining the distribution of respondents over the scale

scores. If respondents "lump" together in certain scale scores,

then we can examine the characteristics of those respondents

for underlying similarities. In the House, we initially

attempted to scale the 10 roll calls loading on Factor I at

.60 or better which also had a communality of at least .50.

These fell short of producing a coefficient of reproducibility

of .90 and we finally narrowed the number of roll calls down

to 8--eliminating those roll calls which produced excessive

scale errors. This produced a scale with reproducibility of

.90.

The House 1963 scale produces marginals which conform to

the pattern MacRae suggests should obtain when the scale is

actually tapping some group cleavage. Table 11-6 shows a

clustering of respondents around scale types I (41 respondents

or 34 per cent) and IX (53 cases, or 43 per cent). To

identify these "lumps" of respondents, we used a contingency







57

Tabl.e I1-6
FREQUENCY DISTRIBUTION OF GSITTMAN SCAIE SCORES
PRODUCED BY SCALE OF? ROLL CALLS LOADING HIGH
ON FACTOR I, HOUSE 1963


Score Frequency Per Cent


I 41 34

II 9 7

III 1 1

IV 3 2

V 2 2

VI 4 3

VII 8 7

VIII 1 1

IX 53 43
_______122*2____ 1 00.___
*The number of respondents totals 122 rather than 125
because 3 respondents were dropped for not responding to over
half the items in the scale.

table analysis crosstabulating scale scores by region and

urban-rural. These findings, summarized in Table 11-7,

indicate some interesting features of the interaction of

region urban-rural differences in Florida politics. Although

the pattern emerges from reading either side of the table, it

is perhaps clearest when the relationship of the scale scores

to urban-rural is examined (both alone and in control for

regional influences). Reading down the right side of the








Table 11-7
RELATIONSHIP OF HOUSE 1963 SCALE SCORES DERIVED FROM FACTOR I
TO URBAN-RURAL AND REGIONAL DIFFERENCESa

Scale Region T Urban-Rual__
Score North Central South Metro Urban Rural
I 6.7 52.9 48.8 71.1 17.8 4.7
Vb 22.2 20.6 27.9 17.8 36.4 20.9
IX 71.1 26.5 23.3 11.1 42.4 74.4
(N) (45) (34) (43) (45) (33) (43)

Controlling for Urban Rural
Scale Metro Urban Rural
Score North Central South North Central South North Central South
I 12.5 93.3 77.3 0.0 28.6 27.3 3.6 0.0 10,0
Vb 25.0 6.7 22.7 37.5 35.7 36.4 17.9 20.0 30.0
IX 62.5 0.0 0.0 62.5 35.7 36.4 78.6 80.0 60.0
(N) (8) (15) (22) (8) (14) (11) (28) (5) (10)

Controlling for Region
Scale North Central South
Score Metro Urban Rural Metro Urban Rural Metro Urban Rural
I 12.5 0.0 3.6 93.3 28.6 0.0 77.3 27.3 10.0
Vb 25.0 37.5 17.9 6.7 35.7 20.0 22.7 36.4 30.0
IX 62.5 62.5 78.6 0.0 35.7 80.0 0.0 36.4 60.0
(N) (8) (8) (28) (15) (14) (5) (22) (11) (10)


aAll cell entries are column percentages. The "N" from which the percentages were
calculated is in parentheses beneath each column.

bAll respondents having scale scores from II to VIII are included here.
Ul
CO







59

table indicates that metropolitan legislators tend toward

scale type I (over 70 per cent), rural legislators towards

type IX (over 70 per cent), and urban legislators distribute

themselves fairly evenly over the three categories with some

greater tendency towards type IX. Reading the next row of

the table indicates that this pattern holds up except under

control for Northern region----where all urban-rural distinc-

tions disappear. Reading the third row of the table indicates

a regional pattern which holds except under control for rural

home county, where all regional differences disappear.

The pattern uncovered in this analysis suggests that

whatever the cluster of interests associated with the rural

end of the urban-rural dimension may be, these interests

characterize all areas of North Florida whether metropolitan

or rural. In other words, the rural type interests dominate

all of the North Florida region regardless of the level of

urbanization. This finding is very similar to one made by

Havard and Beth on the basis of analysis of electoral patterns.

In trying to explain why certain urbanized North Florida

counties (e.g., Escambia and Duval) vote more like rural

areas, they suggest that 'the 'Old South' historical and

social background is still strong enough to prevent the

dominance of clearly urban interests, partly because the new

settlers are mostly from rural Georgia and Alabama."24






60

In the 1963 Senate, those 8 roll calls which loaded

at .60 or better on Factor V formed a Guttman scale which

had a coefficient of reproducibility of .92. Again, the

pattern of the scale type marginals (see Table II--8) suggests

a group cleavage. There is a grouping of respondents around

scale type I (33 per cent)and type VI (42 percent). To examine

the variables underlying this distribution, we used the same

kind of contingency table analysis as used for the House


Table II-8
FREQUENCY DISTRIBUTION OF GUTTMAN SCALE SCORES
PRODUCED BY SCALE OF ROLL CALLS LOADING HIGH
ON FACTOR V, SENATE 1963


Score Frequency Per Cent


I 15 33

II 1 2

III 3 7

IV 1 2

V 6 13

VI 19 42

45 99



(see Table 11-9). This analysis reveals that a relationship

between the scale scores and urban-rural differences exists

and that it exists in all three regions of the state (with

South Florida legislators from urban and rural counties









Table 11-9
RELATIONSHIP OF SENATE 1963 SCALE SCORES DERIVED FROM FACTOR V
TO URBAN-RURAL AND REGIONAL DIFFERENCESa

Scale Region _Urban-Rural
Score North Central South Metro Urban Rural
I 15.0 46.2 50.0 100.0 33.3 4.8
IIIb 20.0 23.1 33.3 0.0 33.3 28.6
VI 65,0 30.8 16.7, 0.0 33.3 66.7
(N) (20) (13) (12) (9) (15) (21)

Controlling for Urban-Rural
Scale Metro Urban I Rural
Score North Central South North Central South North Central South
1 100.0 100.0 100.0 0.0 42.9 66.7 7.7 0.0 0.0
IIIb 0.0 0.0 0.0 60.0 28.6 0.0 7.7 33.3 80.0
VI 0.0 0.0 0.0 40.0 28.6 33.3 84.6 66.7 20.0
(N) (2) (3) (4) (5) (7) (3) (13) (3) (5)

Controlling for Region
Scale North Central South
Score Metro Urban Rural Metro Urban Rural Metro Urban Rural
I 100.0 0.0 7.7 100.0 42.9 0.0 100.0 66.7 0.0
IIIb 0.0 60.0 7.7 0.0 28.6 33.3 0.0 0.0 80.0
VI 0.0 40.0 84.6 0.0 28.6 66.7 0.0 33.3 20.0
(N) (2) (5) (13) (3) (7) (3) (4) (3) (5)


aAll cell entries are column percentages. The "N" from which the percentages were
calculated is in parentheses beneath each column.

bAll respondents having scale scores from II to V are included here.

1'






62

being drawn upward towards the metro end of the scale).

Regional differences appear much weaker and tend either to

disappear or be softened by urban-rural control. Unfortu-

nately, we can only note the difference between the House

and Senate in the way that the scale scores are related to

regional and urban-rural differences. We have no explanation.

Our analysis of the 1963 session of the legislature

identified one significant line of cleavage in each House.

Both the content of the bills involved and the correlation

analysis indicated that urban-rural differences were impor-

tant in explaining this cleavage. Guttman scale analysis

clarified the relationships by indicating that urban-rural

differences predominated in the set of roll calls examined

in the Senate while the House voting showed an urban-rural

pattern modified by regional differences.

1965 Session--In our analysis of the 1965 session of

the legislature we submitted a set of 74 roll calls in the

House and 41 roll calls in the Senate to a factor analysis.

As in the 1963 session, we rotated the first seven factors.

For the House the number of roll calls loading at .60 or

better for each of the seven factors was 6, 2, 2, 0, 0, 1, 1.

For the Senate the pattern of high loadings was 4, 2, 3, 5,

2, 2, 1 for the seven factors. Since we were interested in

those factors with enough high loading roll calls to form







63

Guttma.n scales (a inimum of five) we limited our attention

to Factor I in the House and Factor IV in the Senate.

Examination of the subject matter of these 6 roll calls

with high loadings on Factor I in the House leads to a

suspicion that urban-rural cleavage is the underlying variable

(see Table II-10). Roll call number 34 was on a memorial to

the United States Congress urging an amendment to the con-

stitution authorizing the use of non-population factors in

the apportionment of bicameral legislatures. Certainly this

matter strikes at the heart of urban-rural political conflicts.

The subject matter of the high loading roll calls on Factor

IV in the Senate (see Table II-ll) show no clearly discern-

able thread tying these roll calls together.

In attempting to discover the identity of these factors

in the 1965 analysis we turned next to the relationships

between associations of the roll call votes in the analysis

to urban-rural and regional differences and the pattern of

loadings on the factors produced. The analysis demonstrated

that these variables (urban-rural and regional differences)

had a lessened impact on roll call voting in 1965 than in

1963. In the 1963 House, urban-rural differences had a

significant relationship to voting in 50 per cent of the

set of 80 roll calls in our analysis. Regional differences

had a significant relationship to the votes in about 36






64

Table 11-10
SUMM IARY OF SUBJECT .MATTER ISSUES INVOLVED IN ROLL CALLS
WITH HIGH LOADINGS ON FACTOR I, HOUSE 1965

Roll Call Roll Call
Number Loading Description
41 -.72 Creates Redlands County. Vote on motion
to lay bill on table (y=69, n=36)

40 -.70 Creates county of Miami Beach. Vote on
motion to lay on table. (y=61, n-40)

48 .68 Deals with acquisition of lands for state
buildings and facilities for the capitol
center; authorizes issuance of revenue
certificates. (y=56, n=45)

34 .66 Memorial to U. S. Congress urging an
amendment authorizing a state with a
bicameral legislature to utilize factors
other than population in apportioning
one house if approved by the electorate.
(y=77, n=33)

11 .61 Memorial to U. S. Congress proposing an
amendment to Article X of the U. S.
Constitution specifying that exclusive
jurisdiction of schools shall be reserved
to the states (y=62, n=48)

56 -.60 Appropriations bill. Vote on amendment
which provides that no district board
surplus funds shall be used to acquire
a land right-of-way for the Cross
Florida Barge Canal (y=56, n=50)



25
per cent of those roll calls.25 In the 1965 House the votes

on about 10 per cent fewer roll calls (39 per cent of 74)

showed significant relationships to urban-rural differences.

Regional differences are related to the votes on about the

same per cent of the roll calls as in 1963 (34 per cent). The








Table II-11
SUMMIARY OF SUBJECT MATTER ISSUES INVOLVED IN ROLL CALLS
WITH HIGH LOADINGS ON FACTOR IV, SENATE 1965

Roll Call| Roll Call
Number Loading Description

10 -.79 Provides for regulation of lie detector
examiners. (y=18, n=21)

11 .70 Amendment to state constitution relating
to the District Court of Appeals. Pro-
vides for the number of appellate
districts, number of judges in each, and
composition of court. (y=21, n=21)

40 .67 Allows the sale of Florida citrus or
goods promoting the state along the
Florida Turnpike; allows advertising
promoting Florida attractions. (y=21,
n=15)

37 .66 Extends dates for racing and jai alai;
provides that part of the proceeds are
for charitable purposes. (y=25, n=9)

36 .64 Provides that attorney's fees in eminent
domain proceedings shall be set by jury.
(y=28, n=9)



reduction in the impact of the independent variables is even

more dramatic in the 1965 Senate. Whereas votes on about

26 per cent of the roll calls in the 1963 Senate analysis

were significantly related to urban-rural differences, the

votes of only 10 per cent of the set of 41 roll calls in

the 1965 Senate analysis demonstrated this kind of relation-

ship. Similarly, while regional differences were significantly

related to about 2i per cent of the roll calls in the 1963






66

Senate analysis, only 12 per cent of the roll calls in

the 1965 Senate analysis were so related. The reduction in

the impact of urban-rural differences in the House and both

urban-rural and regional differences in the Senate in 1965

suggest an explanation for the much weaker structure of the

factor analysis for the two houses. The factor analysis

shows less because the bases of voting which underlay the

1963 factor analysis have a lessened impact on the 1965

legislature.

A preliminary analysis of the relationship of the Senate

1965 factor analysis to the association between the roll call

votes and urban-rural and regional differences gave no clues

for factor identification. That is, high loading roll calls

demonstrated a very erratic relationship to the urban-rural

and regional variables. For example, roll calls ten and

eleven load on Factor IV at -.79 and .70 respectively. But,

roll call ten has a coefficient of association of .19 the

urban-rural variable while roll call eleven shows a coeffi-

cient of .38 for such a relationship. We thus abandoned

our analysis of Senate Factor IV. We did, however, pursue

the identity of House Factor I by zero-order and partial

correlation analysis relating roll call loadings to the

coefficients of association between the roll call votes and

urban-rural and regional background. Table II-12 shows that






67

Table II-1.2
MATRIX OF INTERCORRELATIONS 'AMNG MEASTURES OF ROLL
CALL ASSOCIATIONSa AND LOADINGS OF ROLL CALLS ON
FACTlOR I, HOUSE 1.965b

I II III
Roll Call Roll Call Roll Call
Association Association Loading on
with Urban-Rural with Region Factor I


Ic 1.00 0.43 0.58

II 0.43 1.00 0.42

III 0.58 0.42 1.00

aCoefficients of association relate the "yeas" and "nays"
of each roll call vote to the urban-rural or regional break-
down.

bCell entries report Pearson correlation coefficients.

cRoman numerals on the vertical axis of the table refer
to the variables associated with the same Roman numerals on
the horizontal axis. The table thus shows all possible
correlations among the variables listed across the horizontal
axis.


the roll calls loading highly on the factor also tend to show

an association with urban-rural and regional differences.

The covariation of roll calls loadings and roll call associa-

tions with urban-rural differences persists even under control

for relationship of the roll calls to regional differences

(the zero-order correlation is .58 and the partial is .48).

On the other hand the relationship between loadings and

association with regional differences disappears when







68

association with urban-rural differences is partialled out

(the zero-order correlation is .42 while the partial is .23).

This part of the analysis suggests that Factor I in the 1963

House taps the influence of urban-rural differences in the

roll call voting of this session.

The third phase of our attempt to uncover the pattern

of voting in the 1965 House was to Guttman scale the six

roll calls with high loadings on Factor I. These produced

a scale with a coefficient of reproducibility of .90. The

distribution of legislators among the scale types (see Table

11-13) is generally similar to that we obtained for scales


Table II-13
FREQUENCY DISTRIBUTION OF GUTTMAN SCALE SCORES
PRODUCED BY SCALE OF ROLL CALLS LOADING
HIGH ON FACTOR I, HOUSE 1965


Score Frequency Per Cent

I 39 35

II 8 7

III 7 6

IV 11 10

V 14 13

VI 33 29
112 100







69

in the 1963 session. This pattern suggests some sort of

bloc voting, but the greater representation of respondents

in the middle scale types suggests that the impact may be

less than in 1963.

In order to examine the nature of these scale types

more fully, we used a contingency table analysis of the scale

types by urban-rural and regional differences. In the initial

tables, with no control variables, a clear relationship be-

tween both urban-rural and regional differences and the scale

types appears (see Table 11-14). However, controlling one

variable for the effects of the other demonstrates that there

is an interaction of regional and urban-rural differences in

relationship to the scale types. Within North Florida, for

example, metropolitan and urban legislators are closer to

the rural end of the scale while in South Florida rural

legislators are closer to the urban end of the scale. This

suggests the presence of a kind of regional subculture which

mediates and softens differences between urban and rural

legislators, within the region. The urban-rural variable

seems to be the more basic, but its impact is modified by

regional differences. This modification is not, however,

as great as in the 1963 House where the effect of urban-

rural differences disappears altogether within North Florida.










Table 11-14
RELATIONSHIP OF HOUSE 1965 SCALE SCORES DERIVED FROM FACTOR I
TO URBAN-RURAL AND REGIONAL DIFFERENCESa

Scale Region Urban-Rural
Score North Central South Metro Urban Rural
I 58.1 29.0 13.2 8.2 33.3 64.4
IIIb 34.9 38.7 34.2 34.7 44,4 33.3
VI 7.0 32.3 52.6 57.1 22.2 2.2
(N) (34) (31) (38) (49) (18) (45)

Controlling for Urban-Rural

Scale Metro Urban Rural
Score North Central South North Central South North Central South
I 30.0 6.3 0.0 50.0 30.0 25.0 69.0 100.0 36.4
bb
III 50.0 43.8 21.7 50.0 50.0 25.0 27.6 0.0 63.6
VI 20.0 50.0 78.3 0.0 20.0 50.0 50.0 0.0 0.0
(N) (10) (16) (23) (4) (10) (4) (4) (5) (11)


Controlling for Region

Scale North Central South
Score Metro Urban Rural Metro Urban Rural Metro Urban Rural
I 30.0 50.0 69.0 6.3 30.0 100.0 0.0 25.0 36.4
IIIb 50.0 50.0 27.6 43.8 50.0 0.0 21.7 25.0 63.6
VI 20.0 0.0 3.4 50.0 20.0 0.0 78.3 50.0 0.0
(N) (10) (4) (29) (16) (10) (5) (23) (4) (11)


aAll cell entries are column percentages. The "N" from which the percentages were
calculated is in parentheses beneath each column.
bAll respondents having scale scores from II to V are included here.








71

The 1965 session provides us, then, with only one

dimension of voting which can be clearly identified, and

that for only one house. The urban-rural cleavage appeared

as a force in House voting, albeit to a lesser extent than

earlier. In the Senate, we are not able by methods we used

to identify any social bases of voting at all in 1965.

1967 Session--In the 1967 session of the legislature

we analyzed 61 roll calls in the House of Representatives

and 39 in the Senate. As for other sessions, we rotated

the first seven factors produced by the factor analysis.

The number of loadings of .60 or better was 21, 1, 0, 1,

1, 0, 0 for each of the seven rotated factors in the House.

For the Senate, the respective numbers of high loadings

were 13, 3, 2, 2, 2, 1, 3. This structure of loading

points to the dominance of a single influence in the set

of roll calls we analyzed. We therefore devoted our

attention to the identification of Factor I for both House

and Senate in 1967.

Examining the subject matter of the roll calls involved

in these factors did not provide much help in factor identi-

fication. The subject matters ranged over a number of issues

(see Tables II-15 and 11-16). This suggested that the common

bond underlying these factors was not a single attitudinal

continuum, but was probably some kind of group loyalty.








72

Table 1I-15
SU.MMARY OF SUBJECT ATTER ISSUE'S INVOLVED IN ROLL CALLS
WITH HIGH LOADINGS ON FACTOR I, HOUSE 1967

Roll Call Factor
Number Loading Subject Matter

9 .89 Provides appropriation for the junior
college Minimum Foundation Program (y=78,
n=34)

33 .89 Appropriations for fixed capital outlay
buildings and improvements. (y=75, n=42)

13 .87 Provides for group health insurance for
state employees (y=86, n=29)

12 .86 Provides appropriation for State Depart-
ment of Public Welfare (y=63, n=48)

15 .85 Creates a state department to provide
medical assistance to the needy. Pro-
vides an appropriation. (y=68, n=43)

4 .85 Determines amount and procedures for
expending funds for school lunches (y=81,
n=32)

31 .84 General appropriations bill for various
state agencies. (y=70, n=46)

57 .83 Creates the Florida Bureau of Law
Enforcement. Florida Sheriff's Bureau
to be transferred to it. (y=80, n=35)

27 .83 Provides an appropriation to the Florida
Board of Forestry. (y=64, n=47)

49 .82 Provides for an appropriation to the
Florida. Public Service Commission (y=63,
n=44)

42 .80 Provides that polling places should open
at 6:00 a.m. and remain open until 8:00,
thus extending the closing hour. (y=86,
n=28)







73

Table II-15 Continued

Roll Call Factor
Number Loading Subject Matter

3 .78 Authorizes exercise of urban renewal
power for counties with population of
390,000 to 450,000 and over 900,000.
(y=64, n=36)

10 .73 Provides an appropriation for the station
and laboratory of the Institute of Food
and Agricultural Products of the Univer-
sity of Florida. (y=68, n=27)

6 .71 Comprehensive public education bill. Vote
is on whether to table a motion to recon-
sider the vote by which the bill was
defeated (y=46, n=56)

19 .70 Relates to the Department of Public
Safety's Driver's Handbook. Provides an
appropriation. (y=58, n=52)

5 .70 Allows county school boards to build school
facilities on land leased from a government
agency. (y=85, n=22)

35 .67 Increases sales and privileges taxes to
provide funds for the reduction of ad
valorem taxes levied for school district
purposes. (y=64, n=53)

39 .65 Provides that school district tax monies
may be used for the provision of recrea-
tional facilities. (y=60, n=48)

53 .65 Provides for control of junkyard regula-
tion. Junkyards may not operate within
1000 feet of the highway right of way
unless screened from view (y-66, n=43)

8 .63 Provides that Florida will enter the
national compact for education and
establish an educational council. Pro-
vides an appropriation. (y-82, n=21)

2 .62 Deals with city planning. Provides for
regulation and codes. (y=55, n=59)







74

Table 11-16
SUMMARY OF SUBJECT MATTER ISSUES INVOLVED IN ROLL CALLS
WITH HIGH LOADINGS ON FACTOR :, SENATE 1967

Roll Ca.ll Factor
Number Loading Subject Matter

2 .97 Provides for the apportionment of funds
to county boards for the transportation
of school children. Vote is on motion
to waive the rules and read for the third
time in full. (y=26, n=20)

20 .96 General appropriations bill for state
agencies--for salaries and capital out-
lay. (y=28, n=20)

1 .96 Relates to junior colleges. Limits the
required local share of the junior
college Minimum Foundation Program. Vote
on motion to waive rule and read third
time. (y=28, n=19)

21 .96 Provides appropriation for fixed capital
outlay. (y-24, n=18)

5 .95 Increases the amount for instructional
salaries in junior colleges. Provides
an appropriation. (y=27, n=19)

4 .95 Increases the amount for instructional
salaries in public schools; provides an
appropriation. (y=28, n=18)

22 .94 Authorizes expenditures for fixed capital
outlay projects at junior colleges,
vocational-technical centers, and institu-
tions under the Board of Regents. (y=29,
n=17)

6 .93 (same as 22 above; y=30, n=18)

23 .93 General appropriations bill to pay salaries
and capital outlay for state agencies.
(y=28, n=20)









Table 11-16 Continued

Roll Call Factor
Number Loading Subject Matter

24 .93 Provides an appropriation to the Florida
Board of Forestry. (y=28, n=20)

3 .92 Changes formula for determining capital
outlay funds for education. Limits
county matching funds to two-thirds the
amount of state funds. (y=29, n=17)

31 .86 Extends the closing hour of polls from
6:00 p.m. to 7:00 p.m. (y=28, n=17)

7 -.86 Resolution approving the amount of
tuition to be collected from students
during 1967-69. Vote on amendment
which would have increased the registra-
tion fee for state students. (y=17, n=30)



To explore the basis of these factors further we looked

at the pattern of association of the roll calls to urban-

rural, regional, and party differences. Urban-rural

differences show a reduced influence in this session compared

with earlier ones. In the House only 16 out of the 61

(or some 26 per cent) of the roll calls produced a. chi square

value significant at .05 level or better when related to

urban-rural differences. This should be compared to the

39 per cent in 1965 to the 50 per cent in 1963. In the

Senate,significant associations between roll calls and urban-

rural difference dropped from 26 per cent in 1963 to just

10 per cent in 1965 and 1967. In the House, regional








76

differences show a much more stable pattern of associations

to roll calls from session to session. Regional differences

produced significant relationships to 34 per cent of the

roll calls in 1963, 36 per cent in 1965, and 33 per cent in

1967. In the Senate regional differences were significantly

related to 21 per cent of the roll calls in.the 1963 analysis,

12 per cent in 1965, and 13 per cent in 1967. Thus,it seems

to show less fluctuation than urban-rural differences as was

the case with the House. Party differences were added to

the analysis in the 1967 session because this was the first

time the Republicans had had significant membership in either

house of the legislature. Party differences were significantly

related to 41 per cent of the roll calls in the 1967 House

analysis and 46 per cent of the roll calls in the Senate

analysis. To try to sort out this pattern of associations

we related them to the pattern of loadings on the factor

analyses. The correlation coefficients between the factor

loadings and the coefficients of association between the

roll calls and urban-rural, regional, and party differences

is shown in Table 1I-17. These correlation matrices indicate

that, both in the House and the Senate, the strongest pattern

of covariation is between the factor loadings and degree of

association of the roll calls with party. In other words,








77

Table 11-17
MATRIX OF INTERCORRELATIONS AMONG MEASURES OF ROLL CALL
ASSOCIATIONSa AND LOADINGS OF ROLL CALLS ON FACTOR
I, HOUSE AND SENATE, 1967b
House 1967
I II III IV
Roll Call Roll Call Roll Call Roll Call
Association Association Association Loading on
with with with Factor
Urban-Rural Region Party I

Ic 1.00 0.44 -0.14 -0.33
II 0.44 1.00 0.14 -0.04
II -0.14 0.14 1.00 0.93
IV -0.33 -0.04 0.93 1.00


Senate 1967

I II III IV
Roll Call Roll Call Roll Call Roll Call
Association Association Association Loading on
with with with Factor
Urban-Rural Region Party I

Ic 1.00 -0.25 -0.34 -0.49
II -0.25 1.00 0.33 0.35
III -0.34 0.33 1.00 0.92
IV -0.49 0.35 0.92 1.00

aCoefficients of association relate the "yeas" and "nays"
of each roll call vote to the urban-rural, regional, or party
breakdown.

Cell entries report Pearson correlation coefficients.

cRoman numerals on the vertical axis of the table refer
to the variables associated with the same Roman numerals on
the horizontal axis. The table thus shows all possible
correlations among the variables listed across the horizontal
axis.


for both House and Senate high loadings on Factor I and high

correlation of the roll call votes to party differences appear

together. Moreover, it should be noted that as the degree of








78

association between roll calls and party increases, the

degree of association with urban-rural differences decreases.

Party differences displaced urban-rural differences as the

primary line of voting cleavage. Party differences were not

simply an extension of urban-rural differences. Instead,

when party loyalty was invoked by a roll call, that loyalty

cut across urban-rural differences.

The relationships among these variables, along with the

identity of the underlying dimension itself, are further

illuminated by Guttman scales of the roll calls with high

loadings on the factors we considered. In the House the

set of 21 roll calls with loadings of .60 or better on

Factor I scaled with a reproducibility of .92. As seen in

Table 11-18, the group opposition pattern appears in the

frequency distribution although there is a large number of

cases scattered between the two polar groups of the scale.

The pattern of group opposition is even more marked in the

scale for the Senate. The set of 13 roll calls with loadings

of .60 or better produced a scale with a reproducibility of

.99. Table 11-19 makes the pattern of polarized group

opposition clear. Although these 13 roll calls have a

potential for producing fourteen scale types, only six

actually appear. Moreover, a full 87 per cent of the

legislators are accounted for by the two extreme scale types.








79

Table 11-18
FREQUENCY DISTRIBUTION OF GUTTMAN SCALE SCORES
PRODUCED BY SCALE OF ROLL CALLS LOADING
HIGH ON FACTOR I, HOUSE 1967


Score Frequency Per Cent


I 59 50
II 2 2
III 6 5
IV 1 1
V 1 1
VI 1 1
VII 3 3
VIII 1 1
IX 2 2
X 5 4
XI 2 2
XII 1 1
XIII 3 3
XIV 3 3
XV 1 1
XVI 7 6
XVII 2 2
XVIII 3 3
XIX 2 2
XX 14 12
119 100


Table II-19
FREQUENCY DISTRIBUTION OF GUTTMAN SCALE SCORES
PRODUCED BY SCALE OF ROLL CALLS LOADING
HIGH ON FACTOR I, SENATE 1967

Score Frequency Per Cent

I 27 56
II 1 2
III 1 2
IV 1 2
V 3 6
VI 15 2
_48 I._I._0i







80

Examination of the House scale types in terms of urban-

rural, regional, and party differences produced some

interesting relationships. As Table II-20 indicates, only

urban-rural differences show no relationship to scale posi-

tion. The clearest relationship is between party differences

and scale position. Moreover, the table shows that the

correlation analysis was accurately displaying the ability

of party to cut across urban-rural differences. The table

demonstrates that regionalism bears a relationship to

Republican party voting but not to Democratic party voting

on these roll calls. Central Florida Republicans are pre-

dominately spread through the eighteen intermediate scale

types while South Florida Republicans are mainly in type XX.

Nevertheless, the primary dimension tapped by this scale is

clearly that of party.

The Senate scale shows a somewhat different set of

relationships to these variables. As with the House, only

urban-rural differences appear unrelated to scale position

in the first-order (without controls) tables (see Table II-

21). The parts of the table which examine scale type in

relationship to party under control for regional and urban-

rural differences validate the primacy of the party variable

as suggested by the correlation analysis. Neither of these

control variables disrupts or alters the form of the









Table 11-20
RELATIONSHIP OF HOUSE 1967 SCALE SCORES DERIVED FROM FACTOR I
TO REGIONAL, URBAN-RURAL, AND PARTY DIFFERENCESa

Scale Party Urban-Rural Region
Z2_e Dem. Re. Metro Urban Rural North Central South
I 72.5 2.6 48.1 50.0 55.6 59.4 41.5 50.0
Xb 27.5 61.5 37.7 37.5 44.4 40.6 56.1 21.7
XX 0.0 35.9 14.3 12.5 0.0 0.0 2.4 28.3
(N) (80) (39) (77) (24) (18) (32) (41) (46)


Controlling for Region

Scale North Central South
YmNe Dem. Rep. Dem. Rep. Dem. Rep.
I 61.3 0.0 72.7 5.3 85.2 0.0
Xb 38.7 100.0 27.3 89.5 14.8 31.6
XX 0.0 0.0 0.0 5.3 0.0 68.4
(N) (31) (1) (22) (19) (27) (19)

Controlling for Urban-Rural

Scale Metro Urban Rural
TLve- Dem. Rep. Dem. Re_. Dem. Re
I 76.6 3.3 66.7 0.0 66.7 0,0
Xb 23.4 60.0 33.3 50.0 33.3 100.0
XX 0.0 36.7 0.0 50.0 0.0 0.0
(N) (47) (30) (18) (6) (15) (3)


aAll cell entries are column percentages. The "N" from which the percentages
were calculated is in parentheses beneath each column.
S
bAll respondents having scores from II to XIX are included here.









Table II-21
RELATIONSHIP OF SENATE 1967 SCALE SCORES DERIVED FROM FACTOR I
TO REGIONAL, URBAN-RURAL, AND PARTY DIFFERENCESa

Scale Party Urban-Rural Region
Type Dem. Rep. Metro Urban Rural North Centr al South
I 82.8 15.8 54.8 57.1 66.7 75.0 38.9 61.1
IIIb 6.9 21.1 12.9 14.3 0.0 8.3 5.6 22.2
VI 10.3 63.2 32.3 28.6 33.3 16.7 55.6 16.7
(N) (29) (19) (31) (14) (3) (12) (18) (18)



Scale North Central South
Type Dem. Rep. Dem. Re. Dem. Re.
I 90.0 0.0 75.0 10.0 81.8 28.6
IIb 0.0 50.0 0.0 10.0 18.2 28.6
VI 10.0 50.0 25.0 80.0 0.0 42.9
(N) (10) (2) (8) (10) (11) (7)



Scale Metro Urban Rural
Type Dem. Re_p. Dem. Rep. Dem. Rez.
I 87.5 20.0 72.7 0.0 100.0 0.0
IIIb 0.0 26.7 18.2 0.0 0.0 0.0
VI 12.5 53.3 9.1 100.0 0.0 100.0
(N) (16) (15) (11) (3) (2) (1)


aAll cell entries are column percentages. The "N" from which the percentages
were calculated is in parentheses beneath each column.

bAll respondents with scale scores from II to V are included here.







83

relationship between party and scale prosition. It seems

clear that this dimension is party cleavage.

In the 1967 session, then, we were able to identify

one factor for each house, and that factor was demonstrated

to be party cleavage. The point which might bear re-

emphasizing is the fact that urban-rural differences and

party loyalties were not part of a developmental sequence

here, but rather they tended to be mutually exclusive.26

Although urban representation had been increased, urban-

rural differences were displaced by party differences in

influencing roll call votes in this session.

1969 Session--In this last session included in our roll

call analysis, we were faced with an upsurge in the number

of roll calls. In the House 320 roll calls met our ititial

criterion that at least 20 per cent vote for the losing side

of the roll call. In the Senate 132 roll calls met that

criterion. By increasing the required dissent to 35 per

cent and allowing only the 3 roll calls on a single bill to

be included we narrowed the number down to 100 roll calls in

the House. Increasing the required dissent allowed us to

reduce the number to 80 in the Senate.

Because of the tremendous expense of attempting to

factor analyze 100 roll calls we were forced to reduce the

number of rotated factors from seven to four. Since the







84

first few factors had been of greatest importance in

previous sessions we took the risk of rotating fewver factors

with the expectation that our analysis would not be severely

hampered. As will be seen shortly, nothing of great impor-

tance seems to have been lost. The number of roll calls

loading at .60 or better was 12, 4, 2, 5,respectively, for

each of the first four factors. In the Senate the pattern

was 8, 6, 7, 5, 2, 3, 7 suggesting the operation of more

significant variables in influencing roll call voting in

that house. Using the techniques and variables used in

analysis of previous sessions, we were able to make sense

out of only Factors I, II, and IV in the House and Factors

I and II in the Senate. Because these were the only

factors amenable to subsequent treatment, and because of the

need to conserve space, we present summaries of subject

matter issues for these factors only. (See Tables II-22

through 11-26.) These subject matter summaries indicated

only the need to probe deeper through alternative techniques.

They did not suggest any clear identity for the factors.

As in earlier sessions, we went back to the roll calls

and examined their associations with urban-rural, regional,

and party differences. As with the 1967 session, party

differences seemed to play a significant role in roll call

voting. In the 1969 House, 53 per cent of the roll




Full Text

PAGE 1

Reapportionment emd Legislatrlvs Outputs 'A Florida Case Study By JOHM RIGHAIID TODD A DISSERTATION PRESENTED TO TILS GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA Deo eirub e r 1971 ii r i I • : .ss-*^ .->-,•*-?-

PAGE 2

ACKNOWLSDGMEFSiTS • : Many people have contributad to my inteilactual davelopmant and analytical abilities as a political scientist which were required to complete this piece of research. The individual who looims largest in my intellectual bac>.ground is the late Charles D. Farris. My ovm feeble capabilities were greatly sharpened by association v/ith this outstanding political scientist. I regret that he died before this could be coinpleted. In this particular research effort several people deserve mention for the constructive roles they played. Dr. I^^nning J. Dauer, chairman of my supervisory committee, first interested ma in this particular study. Moreover, he allowed jT.e to d-rav/ upon his years of accumulated expertise in Florida politics in the task of making political sense out of the statistical analysis. Dr. Alfred B Clubok assisted by providing continuing advice on methodology while the project V7as being executed and by giving a careful evaluation of the finished product. The remaining members of my supervisory Gormriittee, Dr. David Conradt, Dr. Thomas Henderson, and Dr. Marvin Shaw, assisted by a careful reading and by asking 11

PAGE 3

good questions which clarified and improved the final product. I must, of course, hasten to absolve any of thesegentleinen froiTi complicity in ivhatever errors exist in this final version. My v/ife Pa.t deserves commendation for her patience throughout the v/hole ordeal and for her encouraging words to keep me going. Ill

PAGE 4

. Page ACKIMOWLPJDGMENTS ii LIST OF T^3LES „ v 7-iBSTRACT xiii Chapter I. INTRODUCTION: REAPPORTIONMENT AND POLICY X II. SOGIAJ". CLEAVAGE AIMD LEGISLATIVE DECISION-MAKING 34 III. POLICY OUTPUT ANALYSIS---I 132 IV. POLICY OUTPUT MIKLYSIS — II 225 V. SUFiMARY AND CONCLUSIONS 252 BIBLIOGRAPHY 263 .IV

PAGE 5

LIST OF TABLES Table Page II-l Summary of su?oject matter issues involved in roll calls with high loadings on Factor I, House, 1963 47 11-2 Summary of subject matter issues involved in roll calls with high loadings on Factor II, House, 1963 49 II-3 Surrjnary of subject matter issues iiivolved in roll calls with high loadings on Factor II, Senate, 1953 50 II-4 Summary of subject matter issues involved in roll calls with high loadings on Factor V, Senate, 1963 51 II~5 Matrix of intercorrelations among measures of roll call associations and loadings of roll calls on selected factors. Senate and House, 1963 54 II--6 Frequency Distribution of Guttman scale scores produced by scale of roll calls loading high on Factor I, House, 1963 57 11-7 Relationship of House 1953 scale scores derived from Factor I to urban-rural and regional differences 58 II~8 Frequency distribution of Guttman scale scores produced by scale of roll calls loading high on Factor V, Senate, 1963 60 II-9 Relationship of Senate 1953 scale scores derived from Factor V to urban-rural and regional differences 61

PAGE 6

LIST OF TABLES (contj Table Page 11-10 SurruTiary of subject matter issues involved in roll calls with high loadings on Factor I, House, 1965 54 11-11 Summary of subject matter issues involved in roll calls with high loadings on Factor IV, Senate, 1965 65 11-12 Matrix of intercorrelations among measures of roll call associations and loadings of roll calls on Factor I, House, 1965 57 11-13 Frequence distribution of Guttman scale scores produced by scale of roll calls loading high on Factor I, House, 1965 58 11-14 Relationship of House 1965 scale scores derived from Factor I to urban-rural and regional differences 70 11-15 Summ.ary of subject matter issues involved in roll calls with high loadings on Factor I, House, 1957 ..... 72 11-16 Summary of subject matter issues involved in roll calls with high loadings on Factor I, Senate, 1967 74 II~17 Matrix of intercorrelations among measures of roll call associations and loadings of roll calls on Factor I, House and Senate, 1957 77 11-18 Frequency distribution of Guttman scale scores produced by scale of roll calls loading high on Factor I, House, 1957 79 11-19 Frequency distribution of Guttman scale scores produced by scale of roll calls loading high on Factor I, Senate, 1967 79 VI.

PAGE 7

LIST OF TABLES (cont.) Table Page 11-20 Relationship of Housa 1967 scale scores derived from Factor I to regional, urban-rural, and party differences 81 11-21 Relationship of Senate 1967 scale scores derived from Factor 1 to regional, urban-rural, and party differences, 82 11-22 Suinmary of subject matter issues involved in roll calls with high loadings on Factor I, House, 1959 85 II--23 Summary of subject matter issues involved in roll calls with high loadings on Factor II, House, 1969 85 11-24 Summary of subject matter issues involved in roll calls with high loadings on Factor IV, House, 1969 87 11-25 Summary of subject matter issues involved in roll calls with high loadings on Factor I, Senate, 1969 88 11-25 SumiTiary of subject m.atter issues involved in roll calls with high loadings on Factor II, Senate, 1969 89 11-27 Matrix of intercorrelations among measures of roll call associations and loadings of roll calls on selected factors. House and Senate, 1959 92 11-28 Frequency distributions of Guttm.an scale scores produced on scales of roll calls loading high on selected factors. House and Senate, 1959 95 11-29 Relationship of House 1959 scale scores derived from Factor I and Factor IV to regional, urban-rural, and party differences 97 vii

PAGE 8

LIST OF TABLES (cont.) Table Page II~30 Relationship of Senate 1969 scale scores derived from Factor I and Factor II to regional, urban-rural, and party differences 99 II~31 House committee memberships as related to region and urbanism-ruralisra, 1963-1959 107 11-32 Senate committee memberships a.s related to region and urbanism-ruralism, 1953-1959 107 11-33 CoiTimittee assigmnents per person, by influence score and urban-rural. House, 1963-1969 110 11-34 Committee assignments per person, by influence score and region; House, 1963-1959 Ill 11-35 Committee assignments per person, by influence score and urban-rural. Senate, 1963-1969 ... 112 11-35 Committee assignments per person, by influence score and region. Senate, 1953-1969 113 11-37 Distribution of over/under-representation scores for committee assignments by influence score and by urban-rural, House, 1963-1969 114 11-38 Distribution of over/under-representation scores for committee assignments by influence score and by region. House, 1963-1959 ...... 115 11-39 Distribution of over/under-representation scores for committee assignments by influence score and by urban-rural. Senate, 1953-1969 ........ 117 viii

PAGE 9

LIST OF TABLES (cont.) Table Page II~40 Distribution of over/under-representation scores for coiranittee assignments by influence score and by region. Senate, 1953-1969 118 II~41 Percent of committee assignments, total and by influence score, for urban-rural groupings. House, 1953-1959 120 11-42 Percent of committee assignments, total and by influence score, for regional groupings. House, 1963-1969 121 11-43 Percent of committee assignments, total and by influence score, for urban-rural groupings. Senate, 1953-1969 123 11-44 Per Cent of comiiiittee assignments, total and by influence score, for regional groupings. Senate, 1963-1969 124 III-l Political Variables entered into Statistical analysis of expenditures 140 III-2 Correlation and slope coefficients for adjacent pairs of years of tax and debt measures 144 III-3 Correlations between political variables and tax and debt measures, 1963-196 8 146 III-4 Partial correlations between political variables and tax and debt measures, controlling for previous year's expenditures 150 III-5 Changes in levels of debt and taxation in Florida as related to trends of change for all states 153 III~6 Correlation and slope coefficients for adjacent pairs of years of education measures 155 ix

PAGE 10

LIST OF TABLES (cont.) T-'ble Page III-7 Correlations between political variables and education measures, 1963-19S8 ....... 157 II.I-8 Partial correlations between political variables and education measures, controlling for previous year's expenditures 163 III~9 Cha,nges in levels of education expenditure in Florida as related to trends of change for all states 167 III-IO Correlation and slope coefficients for adjacent pairs of years on welfare measures 169 III-ll Correlations between political variables and welfare measures, 1963-1968 ........ 171 III-12 Partial correlations between political variables and welfare measures, controlling for previous year's expenditures 175 III-13 Changes in levels of welfare expenditures in Florida as related to trends of change for all states 181 III--14 Correlation and slope coefficients for adjacent pairs of years of health and hospital measures 184 III-15 Correlations between political variables and expenditures for health and hospitals, 1963-1968 185 III-16 Partial correlations between political variables and health and hospitals measures, controlling for previous year's expenditures 188

PAGE 11

LIST OF TABLES (cont.) Table Pag e III--17 Changes in levels of health and hospital expenditures in Florida as related to 191 trends of change for all states III-18 Correlation and slope coefficients for adjacent pairs of years of highway ineasures 192 III-19 Correlations between political variables a,nd expenditures for highways, 1953-1968 193 T.II-20 Partial correlations between political variables and highway Measures, controlling for previous year's expenditures 195 III-21 Changes in levels of highway expenditures in Florida as related to trends of change for all states 199 III-22 Relationship between gasoline tax collections and distributions in the ten most and least populous counties, 1963 208 III-23 Relationship between gasoline tax collections and distributions in the ten most and least populous counties, 1969-1970 211 III-24 Relationship between state aid to education and county population 1963 and 1968 215 IV-1 Degree and types of change in legislative operations perceived by Florida legislators as results of reapportionment 2 29 IV--2 Legislative perception of factors influencing passage of constitutional revision and executive reorganization 235 IV-3 Legislative perception of post-reapportionment response to urban problems and factors affecting response ..... 237 xi

PAGE 12

LIST OF TABLES (cont.) Table Page IV~4 City Administrator perception of post-reapportionrrient response to urban problems and factors affecting response • 238 IV-5 Summary of Florida league of municipalities legislative activities 1953-1969 ........ 241 xi;

PAGE 13

Abstract of Dissertation Presented to the Graduate Cov-ncii of the University of Florida in Partial Fulfillmenl of the Requirements for tha Degree of Doctor of Philosophy REAPPCRTIONJyiENT AND LEGISLATI^^E OUTPUTS : A FLORIDA C/\SE STUDY By O'ohn Richard Todd December, 1971 Chairraan: Manning J. Dauer Major Dapartment: Political Science This study atterapted to push further into the analysis of policy output and its deterrainants. The studies based on correlational analysis of public expenditures cast doubt on the importance of reapportionment as a political variable having influence on policy decisions. At the same tirae, many of the analysts responsible for this type of conclusion have suggested that there Viras more to be learned about this question with studies of differing designs. The use of before and after dasigns to relate cma,nge3. in public policy ^•c changes, in political variables has been suggested by some. This was the approach taken in this study. The Florida legislature v;as examined before and after the fact of reapportionment with the goal of locating changes in legislative behavior v/hich might be associated with this change in political structure. X i i i

PAGE 14

As a prerecraisite to j:ela.t5_ng reapportionnient to pol5_cy in Florida, a rol?u call analysis was carried out on four sessions of that state's legislature (two sessions before and two sessions after reapportionment) The purpose of the roll call analysis was to see if cleavages underlyiiig roll call voting were such as to malce it probable that policy would be adjusted after reapportionment. Factor analysis and Guttman scaling revealed that urban-rural cleavage was important in three of the four sessions studied. The exception was the first session after reapportionment, which v/as dominated by party voting. Thus, a type of cleavage v/as important during the study period v;hich would make apportionment a potentially significant political variable. Policy output was analyzed in a two-pronged approach. One approach involved analysis of taxing and spending variables over the time period covered by the legislative analysis. These taxing and spending variables were related to political variables through correlation and regression techniques. The analysis of levels of spending and taxation in Florida showed no significant changes which could unequivocally be related to the fact of reapportionment. However, analysis of the distribution of state funds to local areas did reveal that the formula for the distribution of xiv

PAGE 15

road funds to counties had clianged in a manner which was beneficial to the more larbi^n, populous counties. The other part of the output analysis involved an effort to assess the impact of reapportionment in nonexpenditure policies. These nonexpenditure policies xvere studied through a mail questionnaire administered to legislators and city officials. These data indicated that although the legislators had made some real changes in the amount of authority extended to cities (especially in the second session after reapportionment) that the legislature got a negative evaluation from cities. This negative evaluation resulted from the city administrator's preoccupation with the failures of the legislature to provide adequate fiscal assistance. The questionnaire also revealed that the legislators believed reapportionment to be closely related to the passage of what they regarded as their outstanding legislative achievements in the post-reapportionment sessions. The study suggests that the political significance of apportionment may have been excessively downplayed by studies limited to a correlation or regression analysis of expenditure variables. It emphasizes the need to expand measures of policy output beyond easily accessible expenditure data. An additional point made by the study is that more explicit attention needs to be given to the time dimension. That is, XV

PAGE 16

the study argues that variability across units cannot always be s-ubstituted for variability over time. The need for further analysis of the nature of political cleavages and their relationship to policy output is highlighted by the analysis of urban-rural cleavage in this study. Comparative studies using some indicators of the nature of political cleavages seem to ba an important further step in the analysis of politics and policy. XVI

PAGE 17

, Chapter I INTRODUCTION: REAPFOHTIONjMENT AND POLICY In a bibliographic essay surveying recent developments in the study of state and local politics, Herbert Jacob and Michael Lipsky conclude that "the most marked innovation [in the field of state governraent] has been the investigation of the relationship of policy outputs to social and economic variables. ."-'This systematic and comparative analysis of policy outputs has had the kind of revivifying effect in the field of state government that power structure studies had in the field of local politics. It continues to attract time and talents to its tasks. The study to be reported here is an attempt to add to this line of research centering on state politics and policy. More specif iciilly it is a study of the effects of reapportionment on tlie behavior and output of the Florida legislature. To set the framework and background for the report of the analysis, several preliminary steps need to be carried out in this chapter. First, to underscore the importance of reapportionment to Florida, politics we will outline the history of the reapportionment controversy. Second, we will 1

PAGE 18

briefly sketch the background of policy analysis and its relationship to the reapportic-nraent issue. Third, a preliminary assessTT.erit will bo made of the likely contribution of a study of the kind undertaken here. Finally, an overview will be given of the various kinds of data needed and the analysis techniques to be used in this study. We turn now to the first of these stage-setting steps. Th e Re 3.pporti onirient Controversy in Florida : An Histor ical .Sketx'li — The point to be emphasized in this historical overview is that reapportionment has been a persistent and intense controversy in Florida politics. Ifelapportionment has not been just a matter of the unnoticed silent gerrymanderInstead, it has been a highly visible issue, with hotly contested debate. The politics of reapportionment has ooiranoaly drawn most of the political actors of the state into its vortex. This fact may affect the relationship between policy and apportionment in Florida. ~ l Txvo scholars who have given probably the most attention to understanding the malapportionment problem in Florida, suggest that the allocation of seats in the legislative body has alv/ays been a problem in Florida, They note that "even in the period before the state v/as admitted to the union difficulties were involved in the apportionment of the territorial council."'* In order to examine that notion in

PAGE 19

coiT\prehensible bites we shall divide that long historical period into three par ts~-fron^ statehood to 1900, from 1900 to 1955, and from 1955 to 1957. We believe that these divisioris will give both the scope v/e desire and, at the same time, focxis our attention uipon those events of the reapportionment controversy raost likely to be relevant to our study period in the 1960 's. E£2R statehood to 19 OO— Constitutional provisions have always been intimately involved with the malapportionment problem in Florida. The consitutions of the state have tended to virtually guarantee that malapportionment would be the outcome of any attempt to allocate legislative seats. In addition, reapportionment has been accom.plished most commonly through changing constitutional provisions (either by changing constitutions or amiending existing ones) rather than by statutory authority, V^e will, therefore, devote some attention at the outset to the constituLional provisions relating to apportionment. The precedent for restrictive apportionment provisions v^hich prevent a genuine population-based apportionment was set from the very ?oeginning with the constitution of 1838. This constitution, which served as the constitution v;hen B'lorida becaroe a state in 1845, set out guidelines for apportionment which ultim.ately would make it impossible to

PAGE 20

4 allocat-G enough luembers to ths most populous counties to give their citizens adeaviate repr(?santa.tion, Tlie docnment (1) limited the total niimber of rr.ambers of the House of Representatives, (2) guaranteed at least one reprasentative to each county, (3) stij-valated that senatorial districts should be single-rnen^'oGr districts, and (4) prohibited the subdividing of counties for senatorial districts. The guarantee of oae representative per county and the prohibition against subdividing counties for single-inertiber senatorial districts were sufficient to raalce populationbased apportionment irapossible. A.lthough Florida had several constitutions in the remainder of the nineteenth century — with new constitutions in 1861, 1865, and 1885 — the basic constraining principles of apportionment letid dov/n in the constitution of 1838 (or 1345) ware not overturned. The constitutions of 1868 and 1885 substituted a limit on the maximuin number of house seats to be allocated to one county for a limit on the total number of seats. Nevertheless, if the most populous counties can, at most, have three or four members of the house and all counties are giiaranteed one member, rural overrepresentation is virtually bound to occur. The provision for single-maiober senatorial seats which must encompass at least one whole county persisted throughout the nineteenth

PAGE 21

century „ This guaranteed a malapportionment problera for the senate. The senata provisions concerning apportionmentare even more strii
PAGE 22

the 1915 proposal— atteiTiTited to overturn the pop^ulation basis for the senata and substitute a more restrictive provision based on the federal principle, i.e... one senator per county. Of all these proposals only that of 1923 was accepted by the voters at the polls. Its appeal seems to have derived from the fact that it slightly increased urban representation in the house, although it came nowhere near an equal population basis. During the 1900 to 1955 period the legislature enacted reapportionment statutes twice. In 1925 it simply applied the formula of the 1923 amendment to the House and modified the Senate slightly. The Senate apportionment fell far short of an equal population apportionment since the largest district had over 130,000 people and the smallest had only 4700. The other case of statutory reapportionment in this period cam.e in 1945 when the governor used his constitutional authority to call the legislature into an extraordinary reapportionment session. After fifty-three days of desultory meetings a bill was finally passed which the governor would accept. Its effect was quite minimal. -LVSiHl 19 55 tg_ 1_9..67--In the period from 1955 until the judicial resolution of the question in 1967, reapportionment was a major political issue in the state of Florida. The issue was pressed early in 1955 by Governor LeRoy Collins.

PAGE 23

7 Ks acknowledged the significance of the reapportionment issue by appointing a study coramission during the period v/hen he was governor-elect to ma'ke recoiranendations to him on the question. When the legislature met in regular session in 1955 it turned to the constitutional amendment route to deal v/ith the reapportionment question. The proposal they drew up was in keeping with the existing constitutional provisions for the House. The proposal for the Senate, however, included the extreme notion of applying the federal principle rather than the population principle. This amendment was rejected by the voters in November, 1956. In the meantime. Governor Collins refused to let the legislature off so lightly. He called them into an extraordinary session, v^hich, by constitutional provision, could deal only with the question of reapportionment and v/as bound to continue until reapportionment was achieved. After executive vetoes of two legislative proposals, it was clear that nothing could be accomplished. With the governor and the legislature at a stalemate, the session formally lasted until the expiration of the legislators' terms in 1955. (Actually the legislators recessed themselves, perhaps unconstitutionally, after forty-five days.^) The next attempt to move toward reapportionment came in a special session in 1957 called by the governor to

PAGE 24

8 consider constitutional revision. The legislature wrote its ov;n proposal for apportionment, ignoring the prior v/ork of the constitutional revision commission. Like earlier amendments, this proposal represented only tinkering v/ith the system — no real reform. The usual constraining clauses of the constitution preventing a population-based apportionment remained intact. This amendment was never placed on the ballot because the Florida Supreme Court found the procedure by v/hich 4 it v;as to be presented unconstitutional. In 1959 the legislature proposed a new constitutional amendment for reapportionmentOne of the few distinguishing characteristics of this proposal was that it would have eliminated the equal population requirement from the constitutional provisions regarding the senate. It indicated clearly that economic and geographic considerations could be included with population in apportioning seats in the senate. Once again the proposal v;ent dowm to defeat in the re f erendum vote It its 1961 session the legislature tried again to reapportion itself — still using the constitutional amendment technique. The compromise proposal which they submitted to the public for approval went a little further than previous efforts, but was still inadequate. This amendment never was oresented to the people. The "reapportionment revolution"

PAGE 25

began in earnest v;ith the United States Supreme Court's now ftirno'ds .%;fiSsr v, Cajn^f decision, in 1962,, Since the court had ruled that apportionraent was a justifiable issue, an immediate challenge to the panding constitutional amendinent in Florida v/as filed. In July, 1952, the federal district court ruled that the apportionment provisions of the Florida constitution ware in violation of the federal constitution. Thus, the 1961 proposed amendment was not presented to the voters as scheduled. The narrative from here on involves the plodding efforts of a legislature unable to reapportion itself through repeated efforts, even v/hen the judicial gun was pointed at its hea.d. The legislature passed an inadequate amendment in August, 1962 v/hich the people rejected in November, 1952. The governor recalled the legislature in November, 1962. It deadlocked and adjourned after twenty days. The governor called another special session in January, 1963. It came up with a proposal v/hich it enacted by statute and submitted to the people in the form of a constitutional amendmient. The result was an anomalous situation in which the people rejected the amendment at the polls in November, 1954 v^hile the United States district court had accepted the identical statutory pirovisi.ons as cons-ci-cutronal

PAGE 26

10 Attempts to invaliciivfce the 1955 session of the legislature failed, but the three-judge federal court ruled in January, 1965 that the legislature had only until July 1, 1965, to reappor ir.ion itself in accordance with constitutional standards. In a special session in the sumrfLer of 1955 the legislature presented a plan. The threejudge court accepted the proposal, but the United States Supreme Court rejected it. Once again the legislature was called into special session. In March, 1966 it worked for a, week and came up v/ith a new plan which -//as closer to a genuine populationbased apportionment. In January, 1967, the United States Supreme Court ruled that the plan did not seem, on the surface, to meet standards of legality. It instructed the lov;ar courts to gather more evidence in its behalf or reject 8 it. The Supreme Court's rejection of the 1955 proposal found the legislature already in a special session called to consider constitutional revision. When called upon by the three-judge federal court to offer a new plan for apportiorrment, the legislature made three responses. It presented briefs defending the 1966 plan. New apportionment plans were presented. A plan for weighted voting was put forward. The court rejected all of these alternatives and accepted the plan offered in an aia lcus curi.ae brief by

PAGE 27

.• 11 Dr. Manning J. Dauer of the University of Florida's poixtical 9 scieiice departnient. The court ordered this pla.n rnto effect, and after new elections Florida's first fairly apportioned legislature met in 1957. The struggle to this poiRt--v/hera legislative districts are as nearly equal in population as practicable — has been long and arduous. As our brief historical sketch has indicated, there has never been a period in B'lorida history when malapportionment was not a problem. Particularly in the period since 1955 the reapportionment problem has been a continuing, nagging issue occupying a central place in Florida's politics. The frequent attention to this kind of issue which clearly sets urban against rural interests may have had its impact on the style and substance of Florida politics. This type of concern relates to a style of policy analysis important in political science today. Overv iew of Po_liG_y Analysis a_nd Reapportionment--To place our study in the context of this ongoing research enterprise, we need to review some of the research and findings of the policy analysis v/ork. We begin by noting that the belief that the characteristics of a political system and its socio -econom.ic environment bear some systematic relationship to the m.anner in which the system operates is as old as Aristotle's Polit ics Thus, the current interest

PAGE 28

12 in the determinants of public policy is not new in terms of the questions asked. Instead the contribution of current policy research stems from the methods x:sed. Much of the traditional concern for such matters rested upon processes of speculative inference and casual observation with quite limited reference to real data. By contrast contemporary state policy analysis shows great concern for careful measurement of relevant variables and systematic analysis of their interrelationships. Indeed, one indication of the greater concern for measurement and testing is the level of methodological bickering which has accompanied this current research effort. The distinctive contribution of the current state policy analysis, then, is an empirical base and careful testing of an old line of hypotheses. Much of the interest in comparative policy analysis seems to have grown from seeds planted in V. O. Key's 11 Southern Politics As Key was summing up the factors at work in Southern politics, he put forward what has come to be called "the venerable Key hypothesis.""^ Key suggested that the unprogressive tone of Southern state policies resulted from the lack of competitive party stritctures in that region. He speculated that vigorous party competition tended to orient politics aroiind the struggle bet^veen the economic haves and have-nots. A by-product of party

PAGE 29

13 :o: nvpetition would be a more px:ogressive (i.e., have-not benefitting) set of policies. The key hypothesis provided the basis for a line of research and testing which has not 13 yet culminated with the definitive test. Interest in party competition and public policy was eventually to merge with an interest in malapportionment public policy to form a line of analysis with the goal of evaluating the independent effect of political variables on public policy. The growing concern for malapportionment in the decade of the 19 50 's, although primarily analyzed in legal and ethical terms, fit into the policy analysis approach because reformers frequently made allusions to presumed policy consequences of malapportionment. These implicit hypotheses in the lore of the reapportionment reformers joined with the Key hypothesis to provide grist for the policy analysts' research mills. Both inter-party competition and malapportionm.ent have thus become standard political variables to be included in studies of the determinants of state policy. We begin our review of the research with an early attempt to assess the policy implications of malapportionment carried out by Charles Shull. This piece blends aspects of the reformer literature with an attempt to get at the consequences of malapportionment. Shull sent out a brief questionnaire

PAGE 30

14 to potential informants in the states. These inforraants ware niacle up of "legislators, public officials, political scientists, and personal friends believed to have iudgment a 15 and the ability to exercise it," The set of consequences of malapportionment reported by Shull were based on the beliefs of these presumed expert judges. Before the study of malapportionment and public policy could move on to any quantitative and statistical approaches, some measurement of degree of malapportionment had to be | devised which could order the states from worst to best apportioned. One early attempt at such a measurement pro-' 16 i duced what is now commonly called the Dauer-Kelsay index. f This measure, which is also called the minimal-majority index, adds together legislative districts for each house f beginning with the least populous and moving upward until a j m.ajority of the seats are accounted for. The index is that percentage of the state population which resides in that combination of districts which would make up a m.ajority of the house in question but which would minimize population. 17 The Dav3.d~Eisenberg measure, is an attempt to assign a score indicating the degree to which the urban vote is devalued in a state as a, result of reapportionment. The statistically most elegant measure to be devised is known 18 as the Schubert-Press index. However, this index performs

PAGE 31

15 the dubious feat of combining scores for each house of a two chamber legislature. The index measures deviations from certain statistical norms which would be expected to exist under conditions of perfect apportionment. These efforts at measuring malapportionment paved the way for later attempts to relate apportionment systems to public policy. As the last of the articles dealing with the appropriate means to measure malapportionment v/as being published, one of the first to include a statistical analysis of the relationship of malapportionment to public policy appeared. In 19 this article, Jacob devised three of his own measures of raalapportionment. The measure which he devised for the legislative house based on population seems identical to the Dauer-Kelsay index. Using rank-order correlation techniques, Jacob tested for relationships between his measures of malapportionment and a limited number of expenditure measures. The results were negative. That is, he did not find a clear relationship between a state's ranking on the malapportionment scales and its ranking on the expenditure measures. Jacob also included a simple test of the relationship of malapportionment to such nonexpenditure policies as home rule legislation and right-to-work laws. Again, he found no strong evidence of a relationship to the degree of malapportionment.

PAGE 32

16 The next effort to evaluate the policy eonseqij.ences of malappoi-tionment was published by Dye." All three of the mala.pportioninent indices v/hich we have discussed were used by Dye. This provided an opportunity not only to discuss raalapportionrnent, but also a chance to corranent on the utility of the indices. To measure public policy Dye used thirty variables — twenty-two education and v/elfare expenditure measures and eight measures of taxation. He also included socio-economic measures such as urbanization, industrialization, income, and edijcation. Dye found that when the effects of the socio-economic measures were partialled out that moderate correlations remained betv/een four of his education measures and the David-Eisenberg Index. But, all other relationships disappeared. The year following Dye's article Hofferbert published a study in which he focused on two specific types of consequences which had been associated with malapportionment in reform literature: v;elfa.re orientation and fiscal aid to cities. Using rank order correlation techniques Hofferbert related the rankings of the states on the Schubert-Press index to their rankings on a, composite measure of welfare orientation and on a measure of state aid to cities. He found no significa.nt relationship.

PAGE 33

Brady and Edmonds, in a paper prepared the same year as Hofferbert's article, tried to put to a, systematic test some of the relationships between malapportionment and policy 22 reported in the Shull article. Using the Schubert-Press and David-Eisenberg indices of malapportionment, they found quite weak correlations with their selection of taxing and spending measures. They found that they could explain a much larger percentage of the variance in the taxing and spending measures by a multiple correlation using three socio-economic measures. Like Hofferbert they investigated the distribution of state funds to local governments. They used counties rather than cities, however. They did not find any significant relationship between the way the distribution was carried out and their malapportionment scores. In an effort to explore relationships beyond expenditure measures of policy, Brady and Edmonds carried out a contingency table analysis relating malapportionment to the presence or absence of such things as right-to-work laws, state income tax laws, Kerr-Mills adoptions, and liquor monopolies. They rejected the hypothesis that malapportionment was related to these policies In later work which dealt with total levels of spending as well as spending for specific programs, Sharkansky tried to sort out the rela.tionships between political and

PAGE 34

18 socio-ecoRornic var5.ables and expenditure patterns.^ Sharkansky's work lias two distinguishing characteristics. First, his expenditure variables were restricted to state expenditure rather than the state-plus-local measures characteristic of the work of Dye and others. Second, Sharkansky gave attention to change in expenditure patterns. All three of the standard indices of malapportionment were in his analysis. He found no significant relationships between malapportionment and his state expenditure measure. A protest against the growing list of publications with negative findings concerning malapportionment was published in 1963 by Pulsipher and Weatherby. After reviewing the literature questioning the policy significance of reapportionment, these authors declared that their paper would be an "... attempt to furnish a grain of scholarly salt for 25 assertions of this type." They used regression analysis on a variety of expenditure, political, and socio-economic variables. Unfortunately they threw away most of the information yielded by the regression analysis. As Dye and Dyson pointed out, "Nowhere in the article do the authors assess the independent impact of their political variables on public spending or compare the impact of political and socioeconomic 26 variables." Pulsipher and Weatherby chose instead to perform significance tests upon the regression coefficient

PAGE 35

sltowing the contributi-on of malappor tiomxient to the equation. This kind of approach mistakes statistical significance for political significance. It yields only the information that the malapportionment term in the equation is statistically reliable. While our confidence in the reliability of the regression terra representing malapportionment may be increased, we might suspect (on the basis of other research) that the contribution of the malapportionment term to the regression equation is quite small. Since the authors report only unstandardized regression coefficients, direct comparison of the terms in the equation is not possible. Two years later Fry and Winters published a study which 27 tocused on a different aspect of state policy output. They took as their "... dependent variable t?ie net redistributive impact of revenues and expenditures as represented by the ratio of expenditure benefits to revenue burdens for the 28 three lowest income classes in each state." This study included the Schubert-Press index of malapportionment in its battery of political variables. Both zero-order and partial correlation coefficients indicated no relationship between m.alapportionraent and the amount of redistribution accomplished by state expenditures. One study v/ith results at variance with the others reported so far was Jack Walker's study of innovation among

PAGE 36

20 29 the states. Wa.ll<:er's basic data were nonexpenditurs policies v/hich had eventually been adopted by all of the states. He ca.lculated a score for each state based on the amount of elapsed time between the first adoption of a particular policy and that state's adoption. An overall innovation score was calculated for each state by averaging the scores for each policy included. Walker included the Schubert-Press and David-Eisenberg malapportionment indices among the political variables in his analysis. One of the striking findings of his analysis was that the DavidEisenberg index had a strong correlation with the innovation score V7hich did not disappear when socio-economic variables were partialled out. Hof stetter s is the last approach to assessing the consequences of malapportionment which we want to miention. The basic data were roll call votes taken in the Indiana House of Representatives between 1923 and 1953. The method was to weight votes in such a way as to correct for malapportionm.ent. The effect of malapportionment could be estimated by comiparing these "corrected" votes with the actual votes. Hof stetter concluded "... that numerically equitable apportionment might not have had a. large effect on the out!I 3 1 come of roll-call votes and the passage of legislation. ~

PAGE 37

21 Thus, Kofstetter's results were congruent with the expenditure analysis approach. This review of some of the policy analysis literature indicates that notions about the relationship of malapportionment to public policy have been subjected to some considerable amount of research scrutiny. The finding that there is little or no relationship between measures of malapportionment and public expenditure measures has virtually become a commonplace of this type of literature. The Walker study, however, forces us to leave open the possibility that some policy significance can be attached to apportionment systems. It encourages us to further investigation of the apportionm.e n t pr ob 1 em Ju stification for. this study — ^VJith the array of negative findings concerning the policy relevance of apportionment in the literature reviewed above, one might legitimately question the need for further studies. We recognize, therefore, that the biirden is upon us to show why a study of the effects of reapportionment is worthwhile. The beginning point for a justification for our study is found in the conclusions to some of the studies we have discussed. In summing up what he accomplished with his research Dye remarked that:

PAGE 38

22 All that has been shown is that reapportionment is not likely to have a direct impact on party competition or on certain policy outcomes. This is not to say that reapportionment will have no effect on state political systems or processes. Quantification necessitates a simplification of what may be a very complex question. The consequences of reapportionment may be so subtle and diverse that they defy quantitative measurement. Perhaps the consequences in each state will vary so much that direct interstate comparisons are inappropriate. Certainly v/e need more refined analyses of the im.pact of apportionment systems on state political processes and policy outcomes; we especially need more "before and after" studies of reapportionment. Thus Dye does not hold to a blanTcet assertion that malapportionment is of no consequence to politics and policy. Shar]<:ansky makes a remarkably similar comiTient. He says that: The lack of significant correlation coefficients between state government spending and the measures of voter turnout, party strength, inter-party competition, and the nature of the legislature does not signify a lack of political influence over state budget-makers. Findings only show that certain readily mteasured political characteristics do not vary from state to state in the same manner as expenditures.-^-^ Thus, the authors of two of the more systematic studies of state policy reject the notion that the study of the policy significance of political structures such as apportionment system has been exhausted by the correlation and regression analyses completed. They leave room for further refinements of our knowledge in this area.

PAGE 39

23 Another opening for further research is provided by a remark in the conclusions of Jacob's article. He notes that it may be that the consequences of reapportionm.ent appear only under certain circumstances. This is m itself a hint for a. line of further research on this question — examination of conditions v^hich might cause apportionment to be of m.ore or less policy significance. If we gather together the hints from these researchers, we pretty m.uch have in hand a justification for the study of reapportionment consequences in Florida. Dye and Sharkansky both allude to the possibility of significant consequences of political characteristics such as malapportionment which might elude comparative statistical study. Moreover, Dye suggests that one way of probing into these kinds of consequences is through the use of before and a.fter studies of reapportionment. One such study of the Georgia legislature reported by Brett Hawkins attests to the value of this approach. It also suggests that such case studies may not 35 always confirm the compa.rative studies. Our study of Florida is in the same vein as the Georgia study and has the samie intent. It is designed to look for those lines of influence from apportionment to policy which have not been picked up in existing coiaparative policy studies. t

PAGE 40

24 Another important justification for our study derives from Jacob's coi-pi-aent concerning the possibility of the impact of apportionraent varying with soma set of factors not included in existing studies. We pursue that line of thinking by including an analysis of the basic lines of cleavage underlying legislative voting in Florida both before and after reapportionment. We hypothesize that malapportionment is not lilcely to have important political consequences unless political alignments are significantly influenced by urbanrural differences. By including this analysis we expect that our study may be able to make a worthwhile contribution to our knowledge of the determiinants of public policy. The reasons v/hy we believe this study worth doing can be summarized as follows. First, we use a before and after design, as suggested by Dye, to trace out the chain of events surrounding the fact of legislative reapportionment. Thus v;here previous studies have looked for covariation of measures of malapportionmient and expenditure, v;e shall look at a legislature undergo reapportionment and attempt to assess the consequences of that fact. Second, pursuing Jacob's speculation, we shall examine what we believe might be an important prerequisite for apportionment systems to have policy significance — the existence of real political alignments along an urban-rural basis. Several studies have made

PAGE 41

it clear that assuming the existence of such urban-rural 36 alignments m all states would be quite misleading. We believe that some additional light may be shed on the subject looking first to see if urban-rural political alignments exist before looking at the consequences of reapportionment. The third basis for justifying this study is that we plan to go beyond expenditure m.easures to explore other types of policy consequences. The Walker findings seem to make this effort important. We believe that these three grounds provide sufficient justification for this study. Me thodologica l Over view -With the history of reapportionment issue in Florida sketched out, reviewed, and the justifications for the study set out, v/e move now to the task of sketching out the methods to be used in carrying out the study. We m.ight note that since several types of data will be required and a. variety of analysis techniques employed we intend only to give a preview at this point rather than a detailed methodological exposition. The detailed explanations will accompany the relevant portions of the report. We have noted above our hunch that whether or not reapportionment has real political consequences depends on the nature of the political alignments which structure political issues and decision-making. This leads us rather naturally to an examination of roll-call voting in the

PAGE 42

legislature. If one o£ the important lines of political cleavage in the state of Florida sets urban interests against rura.l, then v/e should be able to find evidence of that in the way that legislators cast their votes. A variety of means of analyzing roll call votes have been used by political scientists, Tv-zo which have been widely used are factor analysis and Guttma.n scaling. Since our interest is in the broad cleavages underlying voting we utilized factor analysis for a preliminary search. Once factors were identified roll calls which seemed to tap an underlying cleavage particularly v^ell could be used to form Guttman scales. These scales, in turn, allowed us to classify individual legislators with regard to the underlying dimension. The data for our roll call analysis were collected from four sessions of the Florida legisla ture~--two preceding reapportionment (1963 and 1965) and two following (1967 and 1969) For each of those sessions we included all roll calls in which at least 20 per cent of those voting voted for the losing side in our initial set for analysis. For technical reasons, which we shall explain in the report of the analysis, it was necessary in some cases to cut down the size of that set further. We tried always to Iceep a. broadly representative set of roll calls from each house for each session.

PAGE 43

Another legislative arena in v/hich political alignmants can manifest themselves is in the coiniaittGe systera. We looked at the committee system to see if it v/orked to the advantage of any identifiable group and to see what implications reapportionment would have on the workings of the system. We believed that the committee system was significant enough to v;a.rrant the search for political alignments there as well as in the roll call votes. Our analysis of the policy consequences of reapportionment will be divided into two parts. Part one deals with expenditure variables and their relationship to rea.pportionment. We believe, however, that the significant policy actions taken by a, legislative body are not exhausted by its fiscal decisions. We therefore examine nonexpenditure policies in addition to the expenditure matters. To study change in expenditure patterns we collected data on a variety of taxing and spending measures covering the time period of our legislative analysis. We collected these data for all fifty states. By regressing an expenditure variable for one year upon the values of that variable for the preceding year we can get a base line which we take to represent normal change. It is then possible to examine Florida expenditures on that variable over time to see if they spurt ahead or lag behind this base line computed from

PAGE 44

28 changes for all fifty states. We can then try to relate those spurts to the fact of reapportionraent Additionally we use correlation techniques to relate these expenditure variables for all the states to certain crude measures of political variables. Our data on nonexpenditure policies are primarily perceptions gathered through questionnaires. We asked the producers of policy themselves — the legislators--to indicate significant policies made and to evaluate th4 effect of reapportionment in the passage of those bills. The other group whose perceptions we v;ere interested in was composed of city Jidministrators Since state action vis-a-vis municipalities has been generally regarded as being affected by malapportionm.ent, we thought it important to ask city officials if they thought reapportionment had resulted in policies more favorable to them. The chapters following will be allocated to reporting these research operations just described. Chapter two reports our explorations into the nature of political cleavage in Florida. That is, it covers the roll call analysis and the committee system analysis. Chapter three begins the policy output analysis by reporting our treatment of expenditure policies in relationship to reapportionment. Chapter four continues the emphasis on policy output and

PAGE 45

29 reapportionment but shifts to nonexpenditura policy areas. Finally, chapter five v/ill provide a summary and conclusions,

PAGE 46

30 Footnotes to Chapter I ^ "'""Outputs,Structure, and. Power: An Assessment of C'hanga in the Study of State a,nd Local Politics," Journa l of Politics, XXX (May, 1968), 510-53, The quotation is from PP 551-12. 9 '"Williain. C. Eavard and Loren P, Beth, Pvapre senta tive Government and Reappor tio nment : A C_a^se Study of Flori da. (Gainesville: Public Administration Clearing Service of the University of Florida, 1960) Studies in Public Administration Ko, 20, The quote is from p. 23. This source forms the primary basis for our account of the reapportionment controversy through the 1950 's. To avoid many repetitive and relatively uninformative footnotes, we will restrict ourselves to footnotes giv.ing page citations for quotations or acknowledging that we are drawing upon another source for this section. Unless some alternative source is cited, the Havard and Beth is our authority. "^Daisy Parker and H. Odell Waldby, "T\^'enty Years of Struggle for Legislative Apportionment in Florida," Fl orida fiX-:?-^S. nili:'isr_si_ty Governmental, Resear ch Bulletin, II (May, 1955), 1-5. The point made here is on p. 2, 'The procedure involves the use of a. so-called "daisy chain" cLausa which tied the success of the amendments together so chat if one failed all failed. The Florida Supreme Court refused to accept this procedure. ''369 U, S. 185 (1962) *^Pa.,rksr and Waldby, pp. 2-3. Sli.4' p. 3 A 1 1 e n Mo r .r i s T he Flori da Handbook (Tallahassee: Peninsular Publishing Co., 1969), p. 121. Manning J. Dauer, "Florida Reapportionment," Business and Economic Dimensions, III (March, 1967), 3-14. The points covered here are on pp. 11-12.

PAGE 47

1 x X". 31 10 For a sampling of the kind of raethodologicai critique angaged in see Charles ?. Cnudde and Donald J, McCrone, "Party Competition and Welfaire Policies in the Ameri.can States, Arrterican. Pol-itical Scierice. Review, LXI3:i (June, 1969) 528-529; see also John H, Fenton and Donald W. Chambarlayne, "The Literature Dealing with the Relationship Between Politica,! Processes, Socio-Economic Conditions, and Public Policies in the Araerican States: A Bibliographic Essay," Poli ty, I (Spring, 1959), 388-394. '^^(New York: Alfred A. Knopf, 1949). ^^Brian R. Fry and Richard Winters, "The Politics of Redistribution, /Vrn.erica.n. Po l it i cal Science Rev_iew, LXIV (June, 1970), 508-522. The quoted phrase appears at p. 521. ^^Early extensions of the hypothesis to other regions are John H. Fenton, Midwest Politics (New York: Holt, Rinehart, & Winston, 1966) and Duane Lockard, New England Stat e P olitics (Chicago: Henry Regnery, 1959). One of the first studies comparing all states and casting doubt on the validity of the Key hypothesis was Richard Dawson and James Robinson, "Inter-Party Competition, Economic Variables, and V7elfare Policies in American States, Jou rnal c^f P olitic s, XXV (May, 1953), 255-289. Dye also concluded that party competition had no independent effect on state and local expenditure patterns. See Polit ics Ec onomic s, and the Public (Chicago: Rand McNally, 1955) Sharkansky reports a similar finding using state spending only in Spending in the America n Sta tes. (Chicago: Rand McNally, 1968). However, Cnudde and McCrone, ojd. oit:. argue that there is some relationship for welfare policies. -'-^"Political and Partisan Implications of State Legislative Apportionment, Law and Contempora ry Problems XVII (Srjring, 1852) 417-439 -^^Ikid. p. 430. -"L^Manning J. Dauer and Robert G. Kelsay, "Unrepresentative States," Nati onal Municipal Review XLIV (December, 1955), :^ 571-575. -^"^Paul T. David and Ralph Eisenberg, Devaluation of the CJrban a_n_d Suburban Vote. (Bureau of Public Administration, University of Virginia, 1961)

PAGE 48

32 ISGlendcn Schubert and Charles Press, "-Measuring Fialapportionroent, American. Political Science Rev iev/ LVII3: (June, 1954), 302-327. 1 9 Herbert Jacob, "The Consequences of Malapportionment: A Note of Caution," Socia l Forces, XLIII (Decerrber, 19S4) 256-261. Thomas R. Dye, "Malapportionment and Public Policy in the States, J ourn al _qf Po litics XXVII (August, 1955) 586601. 21 Richard I. Hofferbert, "The Relation Between Public Policy and Some Structural and Environmental Variables in the American States," American Politi cal Science Revi ew, LX (March, 1966), 73-82. 2 '^-David Brady and Douglas Edmonds, The Sffe_ct of Mal apportionmen t on Policy Output in. the American St ates (Iowa City: University of Iowa Laboratory for Political Research, 1956) 23 Ira Sharkansky, S pen ding in the Americ an Sta tes (Chicago: Rand McNally, 1968) 94 Allan G. Pulsipher and James L, Weatherby, Jr., "Malapportionment Party Competition, and the Functional Distribution of Governmental Expenditures," A.merica n Political Sci ence Review, LXII (December, 1958), 1207-1219. 25 ^Ibi d. p. 12 08. ""^Thomas R. Dye and James W, Dyson, "Communication to the Editor, Am erican Political Sci ence Revi ew, LXIII (June, 1969), 528-529. ^ 'Brian R. Fry and Richard F. Winters, "Politics of Redistribution, American. .Po.lit_ical. Science. Review, LXIV (June, 1970), 508-522. ^^Ikilw p. 508. "-^Jack L. Walker, "The Diffusion of Innovations Among the American States," Amer ican Poli tical Sci ence Review LXIII (September, 1969), 880-899.

PAGE 49

^^C. Richard Hofstetter, "Malapportionment and Roll-Call Voting in Indiana, '^ Jo-arnal of Politics, XXXTII (February, 1371, 92-111. ^-"-IbjiJ.., p. 109. ^^Dye, "l;4a.i apportionment and Public Policy in the States," p. 600. ^^Sharkansky, Spend ing. i,n the A meric an States, p. 64. ^''^JaGob, "The Consequences of Malapportionment," p. 260. -^^Brett W. Hawkins, "Consequences of Reapportionment in Georgia," in Richard I, Hofferbert and Ira Sharkansky (eds.), £tate and Urban Pol itic s: Readings in Comparative Public Policy (Boston: Little, Brown, 1971), pp. 273-298. For articles denying the existence of meaningful urbanrural political alignments see David Derge, "Metropolitan and Outstate Alignments in Illinois and Missouri Legislative Delegations," Aigerican Political Scien ce Review, LII (December, 1958), 1051-1055; Thomas A. Flinn, "The Outline of Ohio Politics," V'7es tern Pol itical Quarterly, XIII (September, 1960), 702-721; and John G. Grumi-n, "The Means of Measuring Conflict and Cohesion in the Legislature, Southw ester n Socia l Sciences Quarterly, XLIV (March, 1964), 377-388.

PAGE 50

Chapter II SOCIAL CLEAVAGE AND LEGISLATIVE DEC I S ION-MJi,KING Much of the political significance of apportionment is derived froni its relationship to the problem of representation. In other vrords, it is necessary to assunie that malapportionment distorts the representation of social interests to endovv the problem with political — as opposed to legal or ethical — significance. This notion was implied by reapportionment reformers. They assumed that the concepts "urban" and "rural" referred to two partitions of society with important differences in policy preferences. Or, as Robert Friedman puts it, "The assum.ption made in using urban and rural distinctions ... is that tha f terms provide a useful tool for distinguishing socioeconomic group interests which tend to congregate in the two different types of environments." The assum^ption that rural people, w'no are overrepresented by malapportionment, are different from urban people is important only if these differences affect the outcome of legislative issues. Reformers assumed, therefore, that the 34

PAGE 51

35 urban-rural cleavage underlay many of the significant decisions taken by any legisla.tive body. In turn, they predicted thcit reapportionment would cause changes in legislative behavior. They expected that the cleavage would continue to have force in legislatures and that reapportionment would shift the balance towards the urban side. The assum.ed importa/nce of the urban-rural cleavage as a variable influencing legislative politics is the primary focus for consideration in this chapter. We take two approaches to uncovering the importance of this cleavage to legislative decision-making in Florida. First, we undertake a method of roll call analysis which has the capability of uncovering sets of rela,ted roll calls, i.e., factor ana.lysis. If the urban-rural cleavage is important in the outcome of the subset of roll calls, it is reasonable to assume that factor analysis can help discover that subset. Second, recognizing that the roll call vote is not the only significant act of decision-making in a legislative body, v/e examined the committee system of the legislature. We looked for possible relationships between the allocation of committee assignmients and the urban-rural cleavage. We recognize that a. definitive analysis of power in committees would involve ana.lysis of issues and the effect of an urban or rural bloc on their disposition in committees. Since such

PAGE 52

36 arx extensive analysis was beyond our resources, we sattled for a positional analysis as a cheap surrogate.''The overrepresentation of urban or rural legislators in influ.ential committees would seem especially convincing if accompanied by a finding of bloc voting in the roll call analysis. Roll Call Analysis Methodolocfv— We turn now to the task of shov-zing how the techniques of roll call analysis can be put to service in examining the place of urban-rural cleavage in legislative behavior in Florida. We begin from the assumption that if the same variable influences the outcome of a subset of roll call votes, this should be expressed by a basic similarity in the way the legislators distribute themselves between the "yeas" and "nays" on the subset of roll calls. Other roll calls, influenced by different variables, would show a different distribution between the yeas and nays. Grouping roll calls by the similarity of response to them is a beginning point for uncovering the set of cleavages — or dimensions — underlying a legislature's roll call behavior. To obtain this kind of grouping we turned to the factor analysis technique. Factor analysis is a generic term which includes several specific statistical procedures. The thing which all of them have in common is, as Duncan MacRae has put it.

PAGE 53

37 that they are concerned. v;it.h ", simplifying the data matrix and expression important aspects of it in terms of 3 a smaller number of underlying variables,"' Herman describes over ten types of factor solutions in his well knov/n work on factor analysis. These va.rious factor solutions can be classified into two broad groupings on the basis of whether the objective is to extract the maximum of variance within the matrix with each factor or to maximally reproduce the correlation matrix from which the factors are extracted. One of the best known solutions which maximizes the variance for each factor is principle components analysis.^ Harman also distinguishes principle components analysis from principle factor analysis. In the case of the former, unity is the value assigned to the cells falling along the major diagonal of the correlation matrix. Principle factor analysis uses the communal i ties (the amount of variance in a measure accounted for by the factor analysis) along the main diagonal. MacRae notes, however, that with large data matrices such as v/e shall be using here there is little difference in results between principle factor and principle components methods. The principle components solution is the one which we ha.ve used here. Orthogonal rotation of factors was used to produce a set of uncorrelated factors, or summary varriables, derived from the matrix of interrcorrelations among roll calls.

PAGE 54

38 Once such factors have been produced, the next step is to search out their rrieaning. Factors derived from roll call data are generally interpreted in terms of issues (attitudes, valiaes) or factions (parties, blocs, delegations, etc.). As MacRae has indicated, "a purely empirical approach cannot suffice to tell us the conditions under which our inferences from roll calls are to be interpreted m terms of issues or factions. In other words, a technique such as principle components analysis provides a classification of roll calls by similarity of response patterns. It does not tell us why these similarities occur. There are several approaches to discovering the meaning of factors. Perhaps the most common approa.Gh is to examine the questions raised by the bills v/hich occasioned the roll call votes with high loadings on a given factor. Grumm indicates that the substantive content of a subset of bills each loaded highly on a given factor may provide direct clues for factor definition. What one looks for is a common element of the subset that is not found in bills with zero and near zero loadings. If this is discovered, the factors may be defined in terms of this element. This approach, however, seems most likely to be useful when the underlying dimension is an issue or attitudinal continuum. It woiild probably not be very helpful if a. group

PAGE 55

allegiance v;ere the ope^rative force. We will, nevertheless, look at the issue content of roll calls for the possibilitythat an underlying urban or rural interest might be revealed. Although this method may not lead to unambiguous identification of factor meaning, if other methods are clearer in identifying urban-rural factors we will then be interested in looking back at the issue content of bills loading on that factor. Another miethod of finding the meaning of factors-suggested by both MacRae and Grumm — is to examine the 10 correlates of the roil call votes loading on a given factor. We can take a set of roll calls which load highly on a factor and examine the way the legislators responded (yea or nay) in relationship to such things as urban-rural or regional background. If we found that for all of the roll call loading on a factor the yeas and nays varied systematically with urban-rural background (e.g., all urban legislators consistently voted on one side and all rural legislators consistently voted on the other) then we v/ould have strong grounds for inferring that the factor basically tapped an urban-rural cleavage. To prepare for this kind of analysis we calculated coefficients of association for each of the roll calls in our analysis relating legislators' votes to such background characteristics as urbanism-ruralism.

PAGE 56

40 region,"^ and party. Once factor loadings ^were available from the factor analysis they could be relared to the coefficients of association. The factor loadings and coefficients of association become data for correlation and partial correlation analysis. Thus, we are able to express in terms of a numerical value (i.e., a. correlation coefficient) the degree to which the roll calls which load highly on a given factor have voting patterns closely associated with urban-rural, regional, or party differences. If correlation analysis indicates a high degree of covariation between loadings on a given factor and coefficients of association between votes and one of our bacT<:ground variables, then we can infer that the factor analysis has grouped the roll calls together on the basis of their relationship to that backgrotind variable. In other words, if we found a correlation of .90 between the loadings of a set of roll calls on a given factor and the coefficients of association between the votes cast on those roll calls and urban-rural differences we could be relatively sure that the factor was an urbanrural factor. It is also possible to clarify some ambiguous relationships by calculating partial correlation coefficients which express the strength of the relationship between one set of coefficients of association (e.g., betv/een votes and urban-rural) and factor loadings while partialling out the

PAGE 57

41 effects of the relationship betv/een roll call votes and other variables (e.g., partialling out correlations between party associations and factor loadings) A third method of uncovering the meaning of factors is to assign factor scores or scale scores to individual respondents-'-in this case legislators. MacRae indicates that "when such a distribution separates two knov/n groups such as legislative parties or when it has intervals of lower density, we [can] infer in either case that meaningful social groups are being separated." The approach we take here is to take the highest loading roll calls from the factor analysis and submit them to Guttman scaling so as to derive the kind of distribution MacRae described. We chose Guttm.an scaling over factor scores for this purpose in the belief that the important thing is to use those roll calls which are the purest indicators of an underlying dimension to scale legislators. This eliminates extraneous elements and error variance which are likely to contaminate factor analysis 14 unless it is cleaned up. Because cleaning up a factor analysis (dropping out measures which contribute little to the analysis and re~-running) is expensive, we chose to select good roll call indicators from the factor analysis and Guttman scale them.

PAGE 58

Guttman scaling, a.ltliough originally developed for use V7it:h attitude data, has become v/idely used in the area of legislative roll call analysis. Two methods have coiT-monly bev-n used in selecting roll calls to submit to scaling. One method involves using the researcher's judgment to select a preliminary set of roll calls which bear some apparent relationship to one another. These roll calls are then scaled, and perhaps some will be dropped if they do not fit the scale pattern. Another method involves the use of statistical procedures to select sets of scalable roll calls. Most coiraTi.only this involves calculating soma coefficient of association (frequently Yule's Q) for all pairs of roll calls, Closely associated roll calls can then be selected from a matrix of such coefficients.-*-^ Our approach is closer to the second of these two approaches. We use factor analysis to select sets of scalable sets of roll calls. These scales, in turn, provide an ordering of legislators in relationship to the underlying dimension. Our analysis of group cleavages covers four sessions of the Florida legislature. Two of them, 1953 and 1965, occurred before reapportionment and two, 1967 and 1969, came after. Since the Florida legislature met biennially until the 1969 session, these are the four regular sessions which bracket the historical fact of reapportionment. To avoid

PAGE 59

biasing the saiiiple of roll calls in our factor ane^lysis, our p-reliminary set included all ]:oll calls taken in a session^ From this set we eliminated some roll calls in each session. First, we eliminiited roll calls with extrerae marginals (i.e., those in which less than 20 per cent of those voting cast ballots for the losing side) For some sessions this was all of the elimination required (House and Senate, 1965 and House and Senate, 1957) In other cases the limitations of the available comiputer programs which performed factor analysis served as an additional constraint on the number of roll calls used. Two programs were used. For all except the House 1969 session the Bio-Med M03 program v;as used. Hov/ever, the large number of roll ce^lis meeting our miarginals criterion in the 1969 House session (320) made a program with larger capacity necessary. Using component subroutines from the 17 IBM Scientific Subroutines Package we were able to put together a program which testing convinced us produced the same results as the M03 This program had a maximum variable Ga.pa.Gity of 100. We thus adopted several additiona.l criteria in reducing the number of roll calls from 320 to 100. First, we further restricted the marginals requirement so as to include only those bills which showed at least 35 per c;ent of those voLing casting ballots on the losing side. This, however, did not eliminate enough roll calls. Noting

PAGE 60

44 the fact that this session had produced several exaicp.ples of m'altiple roll calls on a single bill, we decided to limit to three tha nuruber of roll calls incladed from a single bill Within the three roll call rule, we chose to vary the marginals by picking the roll calls with the highest, lowest, and mean percentage voting for the losing side. Not only did this procedure help to eliminate roll calls, but it also provided an additional protection for the meaning of the factor analysis. Cattail points out that loading a factor analysis up with closely related m.easurements tends to produce factors v/hich are simply artifacts of the presence of IS those measurements, The three roll call rule prevents the production of factors which only demonstrate the fact that multiple roll calls on a bill are likely to be related. ll?:^A Sj^iSsipjx — With these methodological guidelines sketched out w-e C3n proceed now to an examination of the roll call analysis. In both houses of the legislature in 1963 we found it necessary to go beyond our beginning 20 per cent dissent rule in order to reduce the nuidjer of roll calls to a set v/hich could be accomodated by the com.puter program .. • 1 9 wnich we were using." By gradually increasing the percentage of dissent required for a roll call to stay in the analysis, we reduced a set of 90 roll calls in the House and 100 in the Senate to 80 each which we submitted to the faccor

PAGE 61

45 analysis. In both cases the analysis produced a larger rmirtjer on oi: ractors with eigenvalues in excess of 1.00. Ws chose (arbitrarily) in both cases to rotate only the first seven factors with the expectation that this v/ould exhaust the set of significant factors. This proved to be a valid assumption. In the house the first seven factors accounted for about 42 per cent of the variance. However, the first two factors accounted for about 2 6 per cent of the variance (14 and 12 per cent respectively) with each of the succeeding factors adding only about 4 per cent each. In the case of the Senate these first seven factors accounted for 56 per cent of the variance v;ith the factors accounting successively for 17, 10, 9, 6,, 5, 5, and 4 per cent of the variance. The first step in beginning to sort out these factors was to examine the structure of the loadings of the roll calls on them. We selected .60 as the minimum loading which we would regard as significant on any given factor. Although this minimum was initially chosen arbitrarily to bring the analysis into clearer relief, it was observed that for most of our analyses the comraunality for an individual roll call tended to drop below .50 at about the loading point of .60. Since the cornmunality measures the proportion of the variance in a variable accounted for by the factors computed, this cut off point meant that we examined initially only those

PAGE 62

-i-D variables t'nat were at least 50 per cent accounted for by the factor analysis. By examining the roll calls loading above the ,60 we were able to see more clearly the pattern of association bet\veen them and the factor structure. The larger the number of such loadings, the more roll calls the factor could account for alone. In the House in 1963 only the first two factors appeared to be important on the basis of the number of high loadings. Respectively the factors had 12, 6, 0, 0, 3, 0, roll calls I loading at the level of .60 or better. In the Senate the pattern was somewhat different. The number of high loadings per factor was 3, 5, 4, 1, 8, 2, 0. These results led us to i focus our attention on factors I and II in the House and factors II and V in the Senate. Having selected on the basis of loading structure two factors from each house for further analysis, we moved next I to the examination of the content of the bills producing the roll calls. Table II-l presents a listing of the roll calls, in order of loading on Factor I, along with a brief description of the substantive content of each bill involved. Although a varied set of subject matters seem to be involved, there are some clues that this might be a factor involving ; urban-rural differences. For example, the highest loading bill deals with the subject of apportionment itself which is

PAGE 63

47 Table II~1 SUMMARY OF SQBJECT J/ATTER ISSUES IIWOLVED IN ROLL CALLS WITH HIGH LOADINGS ON E'ACTOR I, HOUSE 19S3 Roll Ca.ll Muipber 70 33 69 47 53 60 76 72 42 37 Loading -.78805 -.77879 -.72957 .70768 -.69036 -.65441 ,63425 -.60874 .60611 .60247 .50135 Roll Call Description Constitutional Amendment for apportionment; referred to legislative apportionment as well as constitutional arnendnient coiTuTiittee (y--=58, n==47) Amendment to allow the Milk Commission to fix prices at the producer level (y=54, n-=59) Sets boundaries of Hendry, Martin, Okeechobee, Palm Beach, and Glades counties (y=54, n=55) Amendment to take away Milk Commission's power to fix prices (y=44, n=70) Constitutional amendment providing for second gas tax (y--6 7, n=54) Appointment of administrator of Milk Commission by Commissioner rather than Governor (y=52, n-55) Repeal budgetary procedure under Board of Control for education (y=58, n=61) Excise taxes, severance of natural resources: motion to refer to coinmittse {y=47, n=53) Requirements and fees for registration of barber shops (y==45, n=56) Provides state and county retirement for municipal service (y=44, n=73) Constitutional amendm^ent — non -exemption of conservation laws (y=42, n=6l) Non-taxation of non-bearing fruit trees and nursery stock until maturity (y-57, n=47)

PAGE 64

4.R clearly an urban-rural issue. a se'jond ciuo is fcund in the fact that variable five is concerned with the second gas tax. As will be discussed below, this has been an issue closely related to urban-rural differences. Primarily the urban-rural relationship to the second gas tax derives from the fact that the revenues are partially returned to the counties on the basis of a formula which has been taken to be biased in favor of the more rural counties. The 6 high-loading roll calls on Factor II, listed with subject matter summaries in Table II-2, seem on the basis of subject matter alone to be urban type issues. However, since we perform.ed an orthogonal rotation of factors we would not expect closely related factors to emerge. We leave the identification of this factor to other steps in our analysis. Of the factors from the Senate 1963 analysis only Factors II and V had enough high loading roll calls to provide a basis for subsequent steps of our analysis, so we restricted our attention to them. The sunimary of bill contents for Factor II is presented in Table II-3 The m.ost striking fact is that 4 of the 5 roll calls are on the same bill dealing with taxation of alcoholic beverages, This factor may be an artifact of the type Vv^e mentioned above which emerges because too many m.easures of the same thing have been included. For this Senate session we eliminated

PAGE 65

49 Table II~2 SOMMAKY OF SUBJECT MATTER ISSUES liTVOLVED IN ROLL CAI.LS WITH HIGH LOADINGS ON FACTOR II, ROUSE 1963 Roll Call Kumher 13 50 45 Loadiric Roll Call Description h84296 80024 79245 78410 ,63683 Extends eminent domain and allovjs Key West to secure federal aid for slum clearance (y=69, n=3 5) Slum clearance for Palatka (y=34, n-47) Urban renewal power for Sarasota (y=69, n-=28) Slum claaranca for Ormand Beach (y-6 5, n-3 2) Establish revolving fund for administration of federal urban planning (y=70, n=3 5) 65290 j Memorial to U S, Congress proposing amendment to abolish income tax and set ceiling of 25% (y=52, n=50) roll calls only on the basis of percentage of dissent and thus perhaps did not guard adequately against this possibility. Since the factor seemed primarily concerned with attitudes toward a single bill and preliminary analysis indicated no significant associations between the votes on the five high loading roll calls and variables with which we were concerned, such as urban-rural or region, v/e dropped this factor from further analysis. The remaining factor from the 1963 Senate ana lysis-Factor V which is presented

PAGE 66

50 Table II-3 summhry of subject matter issues in^/olved in roll calls with high loadings on factor ii, senate 1963 Roll Call Roll Call Nnnibar Loading Description 58 .84553 Related to beverage law administration. Restricts excise tax exemptions to citrus products rather than agricultural products (y=28, n=15) 48 .81942 Relates to Beverage law administration of citrus products; provides for additional tax on certain beverages (y--=29, n=14) 49 .80567 Relates to beverage law administration of Florida products; provides for additional tax on alcoholic beverages containing 14% or more alcohol (y=28, n=15) 56 .55453 Provides for across the board increases in taxes on beer, wine, liquor, etc. Vote on amendment which lowers the tax on alcoholic beverages over 48% that are made from Florida citrus (y=21, n=20) 59 .64717 Relates to Florida Agricultural Experiment station; provides an appropriation for research into an earthworm malady (y=29, n=13) with its high loading roll calls in Table II-4 — presents a more interesting potential. The subject matter is more varied yet there are some roll calls which seem related to the urban-rural cleavage. The second gas tax bill appears hare as it did in the House. This suggests urban-rural voting patterns. Other approaches were required in order to assure J dentif ication of the factor, however „

PAGE 67

51 Table I I -4 SUM4ARY OF SUBJECT I>m.TTER ISSUE liT^/OLVED IN ROLL C/iLLS WITH HIGH LOADINGS ON FACTOR V, SENATE 1963 Roll Call Nuiriber 41 61 30 31 65 I^oading -.82821 -.77962 -.76815 67960 17 ,65537 ,55537 -.63755 -.63653 Roll Call Description Redefines Boundaries of Hendry, Martin, Okeechobee, Palm Beach, and Glades counties (y-2 5, n=17) Constitutiona 1 amendment — increases from 50 to 75 years the period during which proceeds from the second gas tax shall be placed in the State Road Department Distribution Trust Fund (y=30, n=15) Relates to beverage law administration of citrus products — additional tax on certain beverages (y=29, n=14) Constitutional amendment — provides for a county legislative council to pass local laws. Motion to refer the bill to Committee on Constitutional Amendments only instead of both it and Committee on Judiciary (y=19, n=21) Changes Florida statutes relating to weapons and firearms {y=24, n=20) Repeals penalty for carrying a pistol or repeating rifles without license; repeals procedure for obtaining a license (y=24, n=^20) Relates to use of lumier for construction — certain uses made unlawful (y=26, n=15) Creates a coiTimittee to investigate infiltration of homosexuals in state agencies (y=30, n=^14)

PAGE 68

The next step in ou.,r path to factor identification involved correlation of all roll call votes in cur analysis v/ith independent variables. Fo.r this session we correlated the votes on each of the roll calls with the urban-rural and regional characteristics^-'of the home county of the legislators.'^'^ We found these variribles to have significant relationships to many of the roll calls in the set. Of the 80 roll calls in the analysis, 21 showed a definite relationship to the urban-rural variable (i.e., produced chi square values significant at the .05 level or better) The regional variable shov/ed a significant association with 17 of the roll calls. This kind of analysis in the 1963 House revealed that 40 of the 80 roll calls in our set were significantly related to urban-rural differences and 29 were significantly related to regional background. Thus, these variables certainly played a part in the outcome of roll call votes in both the House and Senate in 1963. Having demonstrated that urban-rural differences and regional differences were related to roll call voting in 1953, v/e can proceed to see what might be added by relating the pattern of these associations to the factor loadings produced by our factor analysis. The coefficients of association already calculated (shov/ing degree of association between

PAGE 69

53 roil ca,ll votes a.nd urba.n-rural and regional differences) were correlated with the loadings of the roll calls on the factojrs we were examining. These correlations are shown in matrix form in Tab3.e II ~5, The table shows that in the 1953 Senate all three sets of values for the roil calls are correlated. Significantly, the loadings on Factor V show a correlation of .78 v/ith measures of association between roll call votes and urban-rural backgrounds. A problem arises, however, from the fact that these loadings also correlate at ,55 with measures of association between roll calls and region. We were thus faced with trying to assess independently r the role of regional and urban-rural influences. We deal v/ith that problem below. The roll calls in the House 1953 analysis share some of the sarae relationships to our independent variables as those demonstrated by the roll calls in the Senate analysis. The loadings on Factor I show high correlations with roll call associations v/ith both urban-rural (.76) and regional (.74) differences. The only thing which is made clear about Factor II is that it is not explained by any of the variables included in the table. The negative correlations, it shcald be remembered, indicate that as loadings on Factor II increase the association of individual roll calls with urbanismriaralism and region decreases. Whatever the nature of this

PAGE 70

54 Table 11-5 MATE.IX 0? INTERCORRELATIOKS AMONG M:eASURJ3S OF ROLIi CALL ASSOCIATIONS^ aIjD LO/iDINGS OP ROLL CALLS ON SELECTED FACTORS,^ SENATE MID HOUSE, 1963 Senate 19 53 I II III Roll Call Roll Call Roll Call Association Association Loading on v/ith Urbanwith Region Factor V Rural I^' 1.00 0.55 -0.78 II 0.55 1.00 0.55 III 0.78 0.55 1,00 House 1963 I II IV III Roll Call Roll Call Roll Call Roll Call Association Association Loading on Loading on vvith Urbanwiuh Region Factor I Factor II Rural I^ 1.00 0.60 0.76 -0.47 II 0.60 1.00 0.74 -0.14 III 0.76 0.74 1.00 -0.39 IV 0.47 0.14 -0.39 1.00 "Coetf icients of Association relate the "yeas" and "nays" of each roll call vote to the urban-rural or regional breakdown. ^Cell entries report Pearson correla.tion coefficients. "Rornein nr.nierals on the vertical axis of the table refer to the variables associated with the same Roman numerals on the horizontal axis. The table thus shows all possible correlations araong the variables listed across the horizontal axis.

PAGE 71

S5 factor, it cuts across the effects of regional and urbanrural differences. To try to improve our understanding of the relationship between Factor I in the 1953 House and Factor V in the Senate and association of roll calls to urban-rural and region, we turned next to partial correlation analysis. In the House, the original correlation between association with urbanrural and loading on Factor I was .76. Controlling for association of the roll calls with region reduced the coefficient, only to .58. By contrast controlling for association with 1 .: urban-rural reduced the correlation between regional associai tion and loading on Factor I from .74 to .21. Thus, the partial correlation analysis points to identification of Factor I as tapping primarily the effects of urban-rural voting. A very similar pattern of partial correlations appears in the analysis of Factor V in the 1953 Senate. Controlling for roll call associations with regionalism reduces the correlation between factor loadings and associations with urban-rural from .78 to .58. By contra st* partialling cut the effects of roll call associations with urban-rural reduces the correlation between associations with region and loadings on Factor V from .55 to .24. In both cases, then, the pattern of partial correlations suggests a. developmenta.l sequence leading from regional

PAGE 72

56 differences to urban-ruxal differences to roll call voting 23 patterns. While the relationships suggested by the partial correlation analysis are intuitively persuasive, wa pursued the analysis further by the use of Guttman scaling. As indicated above, one of the means of identifying the coiranon factor tying a set of measures together is by scaling those measures and examining the distribution of respondents over the scale scores. If respondents "lump" together in certain scale scores, then we can examine the characteristics of those respondents for underlying similarities. In the House, we initially attempted to scale the 10 roll calls loading on Factor I at .60 or better which also had a coiranunality of at least .50. These fell short of producing a coefficient of reproducibility of .90 and we finally narrov/ed the number of roll calls down to 8 — eliminating those roll calls which produced excessive scale errors. This produced a scale with reproducibility of .90. The House 1963 scale produces marginals which conform to the pattern MacRae suggests should obtain when the SGa,le is actually tapping some group cleavage. Table II-& shows a clustering of respondents around scale types I (41 respondents or 34 per cent) and IX (53 cases, or 43 per cent). To identify these "lumps" of respondents, we used a contingency

PAGE 73

57 Table -lI -o FREQUENCY DISTRIBUTION OP GJTTJmS SCM,E SCORES PRODUCED BY SCALE OF ROLL CALLS LOM)ING HIGH ON FACTOR I, HOUSH 1963 Score Frequency Per Cent I 41 34 II 9 7 III 1 1 IV 3 2 V 2 2 VI 4 3 VII 8 7 VIII 1 1 IX 53 43 T22* ino *The number of respondents totals 122 rather than 125 because 3 respondents were dropped for not responding to over half the items in the scale. table analysis crosstabulating scale scores by region and urban-rural. These findings, sumr,;arized in Table II--7, indicate soma interesting features of the interaction of region urban-rural differences in Florida politics. Although the pattern emerges from reading either side of the table, it is perhaps clearest when the relationship of the scale scores to urban-rural is examiined (both alone and in control for regional influences). Reading down the right side of the

PAGE 74

58 pi; 9 H M Di ce; W P '^ d u < ren is; I o H M l-< H P O (D o ni ^ P Eh ^0 •< CO p D Oi H O 03 121 O H g i-1 W (55 o ro H IT3 3 n c! P r(T. T" —^' O n1< ro .., CO ^ n? ^ • • • ?-i! r^ID o-i p| rH r^i ^'^ CO 4.J in G) rH r-M ^ g r-r-H rH -' jj o CO CO 0^ ro 03 l> ro sf OI Csl ro rH c nj M h' +Ji frj c 0) (D ti! Ul C^ (X) LD ,-. • o ^ CM O (.0 ro in rg (N ^— ^ *— V -p rr-i iH in u • '^ IX) CM iH --^ s CN rCD 0) r-l Vi m O 0': 03 s ,0 > H (C! P ro rQ P u o cr a -H rH I o u 4-1 c o u ^zt o o o ,-. ,jJ O Tj O O O rH O rH ro U3 — CO o o o o -— • • 'in o o o --' tN CO VT) CTi VD ^-~. • • • CO ro !^ CO CN ro si' ^^ • > • l> IX) >.D CN ro ro fD y3 r^ !^ ^— V M • • • "=* 4J| CO in in H d r^ ro ro — o >n in o rc-j CO ro 10 — ro r~o --^ • • • CM r-CM o o-i r^ CM ^— ro ro ro m KD o CT) 4-1 in o ^41 • • • CM in oj SI rH CM (X) ITi rH in 00 0) CD I— 1 M iTi u u 03 CO s H > H 1— 1 o O o ^ fa • q u o o O I— T P rH ro IX) ^_C^ rc; c ro st ':? 4J ro • rH p fi r'X3 U5 M !-; CM ro 00 • 03 P ro rO M ft • CM 4J r^ CN] o CM (D rrM v.-^ S r-H o o o ^—^ ro • • in M o o o — -• P CM 00 ^ c o H tr> (D r-\ Cti fD C 1X5 [^ r^-^ 5-i ro • • • •* ^ 4-) .Q CO in in r-l P M oo ro ro 4-1 CD P u tn fi H iH H ro rO „-^ M 5-1 • • > in 4-) 4J ro ys o rH c 0) C71 s — b s; u .H IX) C71 CX) ,"^ ro • • 03 5-1, ro r-03 CM ^ r-i r• — P^i .XI P O in in ^ — V jj ro • • • CO 5-1 ^ CD rCN -—' M ro VX) tg P • M in O in .—. 4-) • • 03 CD CM 1/3 CN — s CM U3 '^ CD (D •-_^ rH U rQ X ro H > H CJ u 00 r;3 (D iJ Cfl CD tn fO 4-> c CD • o CD 54 U CD a, ^ '0 rC o 4-> d 3 rC i-H o -H G .x: -H •> >-> CD g J4 ro M IM H H :; H ^ > O CD 4J rC Eh • H G H g P g w rH CD ^ tn <4H ro 4-) ^ to a o (H CD ra 54 U QJ M o CD ,c; CO O, 4J ra CD C CD rH g G ra 3 CD u ^ rQ CO u Cfi 'en CD G CD m rl M (0 > ra fi ro 4J ^ CO c: CD CD m -H u 4J Vi ra fl -P a, CD C TJ CD c c -H o rH a, rH CO CO CD •H CD U T3 5H rH (D rH rH 4-1 rH !< (0 < ra rH ^ 3 CJ rH ro o

PAGE 75

table indicates that inetrox''olitan .legislators tend toward scale type I (over 70 per cent) rural legislators towards type IX (over 70 per cent) and urban legislators distribute theraselves fairly evenly over the thz'ee categories with soma greater tendency tov/ards type IX. Reading thve next row of the table indicates that this pattern holds up except under control for Northern region-"-~v/here all urban-rural distinctions disappear o Reading the third row of the table indicates a regional pattern v^hich holds except under control for rural _, home county, v;hei"e all regional differences disappear. The pattern uncovered in this analysis suggests that whatever the cluster of interests associated v/ith the rural end of the urban-rural dimension may be, these interests characterize all areas of North Florida whether metropolitan or rural ^ In other words, the rural type interests dominate all of the North Florida region regardless of the level of urbanization. This finding is very similar to one made by Havard cind Beth on the basis of analysis of electoral patterns. In trying to explain v/hy certain urbanized North Florida counties (e.g., Escambia and Duval) vote more like rural areas, they suggest that 'the 'Old South' hisrorical and social bcickground is still strong enough to prevent the dominance of clearly urban interests, partly because the new settlers are raostly from rural Georgia and Alabama."'^'*

PAGE 76

60 In the 1963 Senata, those S roil calls v;hich loaded at .60 or better on Factor V forro.ed a Guttman scale which had a coefficient of reproducibility of .92. Again, the pattern of the scale type marginals (see Table II-S) suggests a group cleavage. There is a grouping of respondents around scale type I (33 per cent) and type VI (42 per cent) To examine the variables underlying this distribution, we used the same kind of contingency table analysis as used for the House Table II-8 FREQUENCY DISTRIBUTION OF GUTTivKJST SCALE SCORES PRODUCED BY SCALE OP ROLL CALLS LOADING HIGH ON FACTOR V, SENATE 1963 Score Frequency Per Cent I 15 33 II 1 2 III 3 7 IV 1 2 V 5 13 VI 19 42 45 99 (see Table II-9) This analysis reveals that a relationship between the scale scores and urban-rural differences exists and that it exists in all three regions of the state (v;ith South Florida legisla.tors from urban and rural counties

PAGE 77

61 > O EH a < h *:— A O Pi ^ nj CO G W W O > 55 H Ci] w w O &4 !J4 Cn H Rj o K 8&1 w !s; 0^ O i W H H tJ O H < M c; ai CO 'J^ i I'^ o o o ^ 4-) • U-) P O O O ""-^ CO CM r/) .-H 1 — I to o n i^"—^ ro 5-1 • • rn M -p O 00 U5 ^ -i fl en VD (5 ^ rr^D — -p • • en M rr-^ rH CO ^ S rc; r-o m ^-^ -P • n P V.0 o en — U3 n r-i CO (D M '"I 1 1 r-H fd f^ (0 CT\ iJD vo — .Q ra 5-1 II • i> M rQ P (N CD CO -^ P =:J' CN CN M o i in c -H ^cjo o o .— rH -p • • IT) i-H M o o o ^ IX) ^ M S +J c >^. o o o ^ -P t • ^ o P o o o o --^ CO H iH fO o o o ^ J-i U • • • en 4-1 4-1 o o O -"OJ C o S rH XJ o o o — +) • (N 5-i o o o — o a rH .^^^ 5 X! rH M H ro H H O H H > W CO 1 iH o Q O ,-^. ra • P in M o O o •— P CO ;-Nj r4 ^ EH ro ro -^ 4-! ro • a • ro P ,Q U3 o ro ^— U VD m CO P O o O — u • • • "=* 4-> O o o >-Q) o S rH r-i O n r-^ nj • • ro !h o ro yD -— :3 ro *X5 cc: c -H &1 a; rH 0:^ nj c (Ti KO vX) -— M n3 • ft • r~!H 4J rQ CN 00 00 ~-O c; M sf CN CM q-4 (D p u tr> s:: H rH rH O O O ^ M u • • ro 4J 4-1 o o O --' r; CU o s rH U <-{ rrVD ^ ro • ro M t^ r^ ^ p 00 -' cc; rC c o o O ^ 4-1 ra • in M .-Q o o O -' !h UD "xi^ 125 P O o O o ^-M t • CN P o o O --' a) o S rH „.— s. 3 OJ (0 ^ rH M H m o H H o u H H > CO CO M 0) CO Q tn nj 4J CO C 0) Q) -H M M fd 4-) 04 ,-\ to (D -H U rH rH 4J f< ra ra rH P o H ra o ^4 0) 0) rH u -H 0) M ro O 4-1 H H g o u MH CO Q) 5-t O U CO (U rH ro o to tn a H > ro to -P £3 O -O PH O CO rH : ,15

PAGE 78

62 being drawn upward towards the metro end of the sca.ls) Regional differences appear rrrac'h weaker and tend either to disappear or be softened by urban-rural control. Unfortunately, we can only note the difference between the House and Senate in the way that the scale scores are related to regional and urba.n-rural differences. We have no explanation, Our analysis of the 195 3 session of the legislature identified one significant line of cleavage in each House. Both the content of the bills involved and the correlation analysis indicated that urban-rural differences were important in explaining this cleavage. Guttman scale analysis clarified the relationships by indicating that urban-rural differences predominated in the set of roll calls examined in the Senate wh?.le the House voting showed an urban-rural pattern modified by regional differences. 196 5 Sess ion — In our analysis of the 1955 session of the legislature we submitted a set of 74 roll calls in the House and 41 roll calls in the Senate to a factor analysis. As in the 1953 session, we rotated the first seven factors. For the House the number of roll calls loading at .60 or better for each of the seven factors was 6, 2, 2, 0, 0, 1, 1. For the Senate the pattern of high loadings was 4, 2, 3, 5, 7., 2, 1 for the seven factors. Since we v/ere interested in those factors with enough high loading roll calls to form

PAGE 79

63 Guttman scales (a iTiinimrim of five; we limited our attention to Factor I in the House and Factor IV in the Senate. Examination of the subject matter of these 6 roll calls v/ith high loadings on Factor I in the House leads to a suspicion that urban-rural cleavage is the underlying variable (see Table 11-10) Roll call nuinj^er 34 "was on a meinorial to the United States Congress urging an amendment to the constitution authorizing the use of non-population factors in the apportionment of bicameral legislatures. Certainly this matter strikes at the heart of urban-rural political conflicts The subject matter of the high loading roll calls on Factor IV in the Senate (see Table 11-11) show no clearly discernable thread tying these roll calls together. In attempting to discover the identity of these factors in the 1955 analysis we turned next to the relationships between associations of the roll call votes in the analysis to urban-rural and regional differences and the pattern of loadings on the factors produced. The analysis demonstrated that these variables (urban-rural and regional differences) had. a lessened impact on roll call voting in 1965 than in 1953. In the 1953 House, urban-rural differences had a significant relationship to voting in 50 per cent of the set of 80 roll calls in our analysis. Regional differences had a significant relationship to the votes in about 36

PAGE 80

64 Table 11-10 SUMIARY OF SUBJECT .MATTER ISSUES INVOL'TED IN ROLL CALLS WITH HIGri LOADINGS ON FACTOR I, HOUSE 1965 Roll Call Roll Call Mumber Loading Description 41 -.72 Creates Redlands County. Vote on motion to lay bill on table (y=69, n-36) 40 -.70 Creates county of Miami Beach. Vote on motion to lay on table. (y=--61, n=-40) 48 .68 Deals v/ith acquisition of lands for state buildings and facilities for the capitol center; authorizes issuance of revenue certificates. (y=56, n=45) 34 .66 Memorial to U. S. Congress urging an amendment authorizing a state with a bicameral legislature to utilize factors other than population in apportioning one house if approved by the electorate. (y=77, n=33) 11 .61 Memorial to U. S. Congress proposing an amendment to Article X of the U. S. Constitution specifying that exclusive jurisdiction of schools shall be reserved to the states (y=62, n=48) 56 -.60 Appropriations bill. Vote on amendment which provides that no district board surplus funds shall be used to acquire a land right-of-way for the Cross Florida Barge Canal (y=56, n=50) 25 per cent of those roil calls. In the 1955 House the votes on about 10 per cent fewer roll calls (39 per cent of 74) showed significant relationships to urban-rural differences. Regional differences are related to the votes on about the same per cent of the roll calls as in 1963 (34 per cent) The

PAGE 81

65 Table 11-11 SUMMARY OP SUBJECT I-IATTER ISSUES INVOLVED IN ROLI, CALLS WITH HIGH LOJVDINGS ON FACTOR IV, SENATE 195 5 Roll Call ^ Roll Call Number Loading Description 10 -.79 Provides for regulation of lie detector examiners. (y=18, n=21) 11 .70 Amendment to state constitution relating to the District Court of Appeals. Provides for the number of appellate districts, number of judges in each, and composition of court. (y=21, n^=21) 40 .67 Allows the sale of Florida citrus or goods promoting the state along the Florida Turnpike; allows advertising promoting Florida attractions. (y=21, n=15) 37 .55 Extends dates for racing and jai alai; provides that part of the proceeds are for charitable purposes. (y=25, n=9) • 36 .64 Provides that attorney's fees in eminent domain proceedings shall be set by jury. (y=28, n=9) reduction in the impact of the independent variables is even more dram.atic in the 1955 Senate. Whereas votes on about 26 per cent of the roll calls in the 1963 Senate analysis were significantly related to urban-rural differences, the votes of only 10 per cent of the set of 41 roll calls in the 1965 Senate analysis demonstrated this kind of relationship. Similarly, while regional differences were significantly related to about 21 per cent of the roll calls in the 1963

PAGE 82

66 Sena.te analysis, only 12 per cent of the roll calls in the 1S65 Senate analysis v;ere so related. The reduction in the impact of urban-rural differences in the House and both urban-rural and regional differences in the Senate in 1965 suggest an explanation for the much v/eal-cer structure of the factor analysis for the two houses. The factor analysis shows less because the bases of voting which underlay the 1963 factor analysis have a lessened impact on the 1955 legislature. A preliminary analysis of the relationship of the Senate 1955 factor analysis to the association between the roll call votes and urban-rural and regional differences gave no clues for factor identification. That is, high loading roll calls demonstrated a very erratic relationship to the urban-rural and regional variables. For example, roll calls ten and eleven load on Factor IV at -.79 and .70 respectively. But, roll call ten has a coefficient of association of .19 the urban-rural variable while roll call eleven shows a coefficient of .38 for such a relationship. We thus abandoned our analysis of Senate Factor IV. V7e did, however, pursue the identity of House Factor I by zero-order and partial correlation analysis relating roll call loadings to the coefficients of association between the roll call votes and urban-rural and regional background. Table 11-12 shows that

PAGE 83

67 Table 11-12 MATRIX 0? INTERCORI^L/^TIONJd AMi.^NG MEASURES OF ROLL CALL ASSOCIATIONS^ AND LOADINGS OF ROLL CALLS ON FACTOR I, HOUSE 1965^ I II III Roll Call Roll Call Roll Call Association Association Loading on wi th Urbon-Rural with Ragion Factor I ic 1,00 0.43 0.58 II 0.43 1.00 0.42 III 0.58 0.42 1.00 ^Coefficients of association relate the "yeas" and "nays" of each roll call vote to the urban-rural or regional breakdown Cell entries report Pearson correlation coefficients. ^Roman nuraerals on the vertical axis of the table refer to the variables associated with the same Roman numerals on the horizontal axis. The table thus shows all possible correlations among the variables listed across the horizontal axis. the roll calls loading highly on the factor also tend to show an association with urban-rural and regional differences. The covariation of roll calls loadings and roll call associations with urban-rural differences persists even under control for relationship of the roll calls to regional differences (the zero-order correlation is .58 and the partial is .48). On the other hand the relationship between loadings and association with regional differences disappears when

PAGE 84

68 assoGiation w.lth urban-rural differences is partialled cut (the zero-order correlation is .42 while the partial is .23). This part of the analysis suggests that Factor I in the 1953 House taps the influence of urban-rural differences in the roll call voting of this session. The third phase of our attempt to uncover the pattern of voting in the 1955 House was to Guttman scale the six roll calls with high loadings on Factor I. These produced a scale with a coefficient of reproducibility of .90. The distribution of legislators among the scale types (see Table 11-13) is generally similar to that we obtained for scales Table II-13 FREQUENCY DISTRIBUTION OF GUTTMAN SCALE SCORES PRODUCED BY SCALE OF ROLL CALLS LOADING HIGH ON FACTOR I, HOUSE 1965 Score Frequency Per Cent I 39 35 II 8 7 III 7 6 IV 11 10 V 14 13 VI 33 29 112 100

PAGE 85

69 in the 1963 session. This pattern suggests some sort of bloc voting, but the greater representation of respondents in the middle scale types suggests that the impact may be less than in 1953 In order to examine the nature of these scale types more fully, we used a contingency table analysis of the scale types by urban-rural and regional differences. In the initial tables, v;ith no control variables, a, clear relationship between both urban-rural and regional differences and the scale types appears (see Table 11-14) Hov^ever, controlling one variable for the effects of the other demonstrates that there is an interaction of regional and urban-rural differences in relationship to the scale types. Within North Florida, for example, metropolitan and urban legislators are closer to the rural end of the scale while in South Florida rural legislators are closer to the urban end of the scale. This suggests the presence of a kind of regional subculture which mediates and softens differences between urban and rural legislators, within the region. The urban-rural variable seems to be the more basic, but its impact is modified by regional differences. This modification is not, however, as great as in the 1953 House where the effect of urbanrural differences disappears altogether within North Florida.

PAGE 86

70 H o Eh O < o Pi (0 H Q O m 13 > li] m c^ CO H Q O < ^ O '^ rH W O I l-H HMO w ft; r-Jj vi' f^ 0-! '^ '^f O) JNJ 1 a fo ^' (N p O] CO '^^ CN 5-i! en ^ cN CO o CM rH „^ u • • • CTi p CO "* r•^ 0) en LO V—-P 1 O 53 r; CM CM VX3 -^ !-> = • • CO ffi fO ^ CM m h :; rH n in -^ O P M CM O H B m 02 .H s fi rc o r~CI .^ o M • • • rH H H +J C3^ CO CN n ^ bC C-J m 00 -' < (D — ^ 5 (U ^ rH M H ra H H u u H I— i > w CO ra I c P U o 4-1 tn G H rM iH o O U rC =* y^ O ,-^ -P • • H S ^0 ro O rH ro >X) — m .H rH m o O O ,— rtJ M < • • in 5-1 4-1 O o o •— 3 C o O:^ Q) rH U ^ o \D o — ~ 4-! • • • M G\ [-^ o — ^D CN m S ^ O O o — 4-! • • ^ j:3 in in o ^ O CN CN in C/3 M C fD o o o ^ m U • • • o fi. 4-J o o O H M CI m m CN •~' P (U o ^ o o o ,-. 4-1 • • •=* M o o o --^ in in 13 ,c; o i> en .-^ 4-1 • > en 3 o rH CO (N CN r•--' CO ^ C nj fo CO o — V M • • vs -1_ -p iX> C^J O rH (1 c: ^^ in -' IS u rc; o o o — p • • o M o o O r-i ro in CN -— a j^-^ ^ '*,--' o o ^ rH h n ro hi H O o \-\ H > m CO o -H !n Pi u o CH H rH rH O U P o rH NtCD O ^ fD a • r-l P W3 n O H r^ CO VO *. — r5 ^ c O o O ^ p fD • B • 13 rC5 in in o ^ P CM CN in CO P o r00 ,~~ p • • • m 4-1 o r-\ CO CM a) CN r--S r~l o O o -^ fD a • • in P o o O •-' o Pi r-1 iH tu C O o o -— P fD • • • o P rQ o o O H c P f^ in CM ^' CD D U m CO o — P > • • CD P ^ 00 O rH 0) •vf in -~^ S rH o CD '^ ^ fD • • a\ P en [^ m CM P >X) CNJ — C^ rC d o o o ^ ;J fD • • • -si^ P ^1 o o o •— p UT in S p o o o ^-p • 1 • o P O o O rH Q) 0-) in CM —' S S5 N,^ Q) 0! ^ rH p H rs Ol hi H CJ C)| M H > u-j CO ID P fd p C (!> P (!) • O, p ri: rG 4-> ^ fi V Xi •H 3 rC rH 5 s -H p ip P fD h > (0 p fi Eh • H c H e • p g to rH p tr> U m fD P ^ to c d) fD p u p o rG CO 0^ P fD c H g c fD 13 rH ^ to o to Gn C to -H p > fD rG fD P .i:^ to c: to •H p P P fD G P 04 c ri d G -H iH a. rH CO to -H u -d p rH !-i H p v-H /< fD < fD rH ,Q 3 CJ t~i fD o

PAGE 87

71 The 1965 session provides us, then, with only one dimension of voting v^hich can be clearly identified, and that for only one house. The urban-rural cleavage appeared as a force in House voting, albeit to a lesser extent than earlier. In the Senate, v;e are not able by methods we used to identify any social bases of voting at all in 1965. 1967 Session --In the 1967 session of the legislature we analyzed 61 roll calls in the House of Representatives and 3 9 in the Senate. As for other sessions, we rotated the first seven factors produced by the factor analysis. The number of loadings of ,60 or better was 21, 1, 0, 1, 1, 0, for each of the seven rotated factors in the House. For the Senate, the respective numbers of high loadings were 13, 3, 2, 2, 2, 1, 3. This structure of loading points to the dominance of a single influence in the set of roll calls we analyzed. We therefore devoted our attention to the identification of Factor I for both House and Senate in 1967. Examining the subject matter of the roll calls involved in these factors did not provide much help in factor identification. The subject matters ranged over a number of issues (see Tables 11-15 and 11-15) This suggested that the common bond underlying these factors was not a single attitudinal continuum, but was probably some kind of group loyalty.

PAGE 88

72 Table 11-15 SUMMARY OF SUBJECT M2\TTER ISSUES INVOLVED IN ROLL CALLS WITH HIGH LOADINGS ON FACTOR I, HOUSE 1957 Roll Csll Factor Nuinb e r I L o a d i na 13 15 31 57 27 49 42 ,89 89 86 ,84 S3 .32 80 Provide.s appropriation for the junior college Mininram Foundation Program (y=^78, n=34) Appropriations for fixed capital outlay buildings and improvements. (y=75, n=42) Provides for group health insurance for state employees (y=86, n=29) Provides appropriation for State Department of Public Welfare (y=63, n=48) Creates a state department to provide medical assistance to the needy. Provides an appropriation. (y=68, n=43) Determines amount and procedures for expending funds for school lunches (y=81, n-32) General appropriations bill for various state agencies, (y-70, n=46) Creates the .Florida Bureau of Law Enforcement. Florida Sheriff's Bureau to be transferred to it. (y=80, n=35) Provides an appropriation to the Florida Board of Forestry. (y=64, n=47) Provides for an appropriation to the Florida Public Service Commission (y=63, n=44) Provides that polling places should open at 6:00 a.m. and remain open until 8:00, thus extending the closing hour. (y=85, n=28)

PAGE 89

73 Table II'.--on-xnued Roll Call Number Fact". or j Loadiriq 10 19 35 39 53 78 73 71 70 70 6 7 S5 65 ,63 {-. 2 S ub j e G t Ma. 1 1 e r 2\utl,ioriz83 exercise of urban renev/al power for counties with population of 390,000 to 450,000 and over 900,000. (y=64, n=36) Provides an appropriation for the station and laboratory of the Institute of Food and Agricultural Products of the University of Florida, (y=68, n=27) Gomprehensive public education bill. Vote is on whether to table a, motion to reconsider the vote by which the bill was defeated (y=46, n=56) Relates to the Department of Public Safety's Driver's Handbook. Provides an appropriation. (y=58, n=52) Allows county school boards to build school facilities on land leased from a government agency, (y-85, n=22) Increases sales and privileges taxes to provide funds for the reduction of ad valorem taxes levied for school district purposes. (y=64, n=53) Provides that school district tax monies may be used for the provision of recreational facilities. (y=60, n=48) Provides for control of junkyard regulation. O'unkyards may not operate within 1000 feet of the highway right of way unless screened from view (y-56, n=43) Provides that Florida will enter the national compact for education and establish an educational council. Provides an appropriation, (y-82, n=2l) Deals with city planning. Provides for regulation and codes. (y=55,, n~59)

PAGE 90

74 Table 11-16 SUFU^'lcARY OF SUBJECT MATTER ISSUES IWJOVmD IN ROT_,L :ali,s WITH HIGH LOADINGS ON FACTOR SENATE KO-li v.-a..l-!. Number 20 22 6 23 Factor joading .97 95 95 .96 ,95 95 94 Q^, ,93 Subiect Matter Provides for the apportionment of funds to county boards for the trsnsportiition of school children. Vote is on motion to waive the rules and read for the third time in full. (y=26, n=20) General appropriations bill for state agenGies~-for salaries and capital outlay. (y=28, n=20) Relates to junior colleges. Limits the required local share of the junior college Minimum Foundation Program. Vote on motion to waive rule and read third time. (y=28, n=19) Provides appropriation for fixed capital outlay, (y-24, n=18) Increases the amount for instructional salaries in junior colleges. Provides an appropriation. (y=2 7, n=19) Increases the amount for instructional salaries in public schools; provides an appropriation. (y=28, n~18) Authorizes expenditures for fixed capital outlay projects at junior colleges, vocational-technical centers, and institutions under the Board of Regents. (y=29, n-17) (same as 22 above; y=30, n=18) General appropriations bill to pay salaries and capital outlay for state agencies. (y=28, n-20)

PAGE 91

75 Table II-IG Continued Roll Call Factor ,w_ .„-,„._..__: —-_-.._-— Ninrtoesr Loading Sub j act Matter 24 .93 Provides an appropriation to the Florida Board of Forestry, (y-^=28, n=20) 3 ,92 Changes formula for determining capital outlay funds for education. Limits county matching funds to two-thirds the amount of state funds. (y~29, n=17) 31 .86 Extends the closing hour of polls from 6:00 p.m. to 7:00 p.m. (y^28, n=17) 7 -.86 Resolution approving the amount of tuition to be collected from students during 1967-69. Vote on amendment which would have increased the registration fee for state students, (y-17, n=30) To explore the basis of these factors further we looked at the pattern of association of the roll calls to urbanrural, regional, and party differences. Urban-rural differences show a reduced influence in this session compa,red with earlier ones. In the House only 16 out of the 61 (or some 26 per cent) of the roll calls produced a chi square value significant at ,05 level or better when related to urban-rural differences. This should be compared to the 39 per cent in 1965 to the 50 per cent in 1963. In the Senate significant associations between roll calls and urbanrural difference dropped from 26 per cent in 1963 to just 10 per cent in 1965 and 1967. In the House, regional

PAGE 92

76 differences snow a much more stable pattern of associations to roll calls from session to session. Regional differences produced significant relationships to 34 per cent of the ...roll calls in 1963^ 36 per cent in 1965, and 33 per cent in 1967. In the Senate regional differences were significantly related to 21 per cent of the roll calls in. the 1963 analysis, 12 per cent in 1965, and 13 per cent in 1967. Thus, it seem,s to show less fluctuation than urban-rural differences as was the case with the House. Party differences were added to the analysis in the 196 7 session because this was the first time the Republicans had had significant membership in either house of the legislature. Party differences were significantly related to 41 per cent of the roll calls in the 1967 House analysis and 46 per cent of the roll calls in the Senate analysis. To try to sort out this pattern of associations we related thera to the pattern of loadings on the factor analyses. The correlation coefficients between the factor loadings and the coefficients of association betv/een the roll calls and urban-rural, regional, and party differences is shown in Table 11-17. These correlation matrices indicate that, both in the House and the Senate, the strongest pattern of covariation is betv/een the factor loadings and degree of association of the roll calls with party. In other wor-ds

PAGE 93

Table 11-17 MATRIX OF INTERCORRELATlOMS AMONG MEASURES OF ROLL CM,L ASSOCIATIONS^ AJSFD LOADINGS OF ROLL CALLS ON FACTOR I, HOUSE AND SENATE, 196?'^ House 196 7 I II III IV Roll Call Roll Call Roll Call Roll Call Association As.sociation A.ssociation Loading on with with with Factor Urban -Rural Region Party I I'^ 1.00 0.44 -0.14 -0.33 II 0.44 1.00 0.14 -0.04 III -0 14 0.14 1.00 0.93 IV -0.33 -0.04 0.93 1.00 Senate '. 1.957 I II III IV Roll Call Roll Call Roll Call Roll Call Association Association Association Loading on with with with Factor Urban -Rural Region Party I I^ 1.00 -0.25 -0.34 -0.49 II -0.2 5 1.00 0.33 0.35 III -0.34 0.33 1.00 0.92 IV -0.49 0.35 0.92 1.00 '^Coefficients of association relate the "yeas" and "nays" of each roll call vote to the urban-rural, regional, or party breakdown. Cell entries report Pearson correlation coefficients. ^Roman numerals on the vertical axis of the table refer to the variables associa.ted v/ith the same Roman numerals on the liorizontal axis. The table thus shows all possible correlations among the variables listed across the horizontal axis for both House and Senate high loadings on Factor I and high correlation of the roll call votes to party differences appear together. Moreover, it should b£i noted that as the degree of -i,.;r£^-t^-c^.-iT!

PAGE 94

78 association betv/een roll ci-3.1.1s and party increases, the degree of association with urban-rural differences decreases, Party differences displaced urban-riaral differences as the primary line of voting cleavage. Party differences v/ere not siiriply an extension of urban-rural differences. Instead, when party loyalty was invoked by a, roll call, that loyalty cut across urban-rural differences. The relationships, araong these variables, along with the identity of the underlying dimension itself, are further illuminated by G'uttraan scales of the roll calls with high loadings on the factors we considered. In the House the set of 2 1 roll calls with loadings of .60 or better on Factor I scaled with a reproducibility of .92. As seen in Table 11-18, the group opposition pattern appears in the frequency distribution although there is a large number of cases scattered between the tv/o polar groups of the scale. The pattern of group opposition is even more marked in the scale for the Senate. The set of 13 roll calls with loadings of .60 or better produced a scale with a reproducibility of .99Table 11-19 makes the pattern of polarized group opposition clear. Although these 13 roll calls have a potential for producing fourteen scale types, only six actually appear. Moreover, a full 87 per cent of the legislators are accounted for by the tv/o extreme scale types.

PAGE 95

79 Table 11-18 FREQUENCY DISTRIBUTION OF GUTTMMT SC7LLE SCOPJ^S PRODUCED BY SCALE OF ROLL CALLS LOADING HIGH ON FACTOR I, HOUSE 1967 Score Frequency Per Cent I 59 50 II 2 2 III 6 5 IV 1 1 V 1 1 VI 1 1 VII 3 3 VIII 1 1 IX 2 2 X 5 4 XI 2 2 XII 1 1 XIII 3 3 XIV 3 3 XV 1 1 XVI 7 6 XVII 2 2 XVIII 3 3 XIX 2 2 XX 14 12 119 100 Table 11-19 FREQUENCY DISTRIBUTION OF GUTTMAN SCALE SCORES PRODUCED BY SCALE OF ROLL CALLS LOADING HIGH ON FACTOR I, SENATE 1967 Score Frequency Per Cent I 27 56 II 1 2 III 1 2 IV 1 2 V 3 6 VI 15 32 48 IQS)

PAGE 96

8 Sxainination of the House scale type,s in terms of urbc.nrura 1, regional, and party differences produced sorue interesting relationships. As Table 11-20 indicates, only urban-rural differences show no relationship to scale position. The clearest relationship is between party differences and scale position. Moreover, the table shoves that the correlation analysis v/a s accurately displaying the ability of party to cut across urban-rural differences. The table deinonstrates that regionalism bears a relationship to Republican party voting but not to Democratic party voting on these roll calls. Central Florida Republicans are pre-' dorninately spread through the eighteen intermediate scale types while South Florida Republicans are mainly in type XX. Nevertheless, the primary dimension tapped by this scale is clearly that of party. The Senate scale shows a somewhat different set of relationships to these variables. As with the House, only urban-rural differences appear unrelated to scale position in the first-order (without controls) tables (see Table II21) 1'he parts of the table v/hich examine scale type in relationship to party under control for regional and urbanrural differences validate the primacy of the party variable as suggested by the correlation analysis. Neither of these control variables disrupts or alters the form of the

PAGE 97

81 H rQ .^ lo r^ ro ,.— •p a ^0 ri o rH CO -N? LH C^S TnI — cn c rH fO in •—i ^x -rt M • (-i tr -P tH vO :m xi' (11 sf in •^^ Pi rCi -:!< VD o ---P • • CS] iH 0^ o O 01 in =* • — a rH W5 "* o ^ ro • • • CX) ^ in <* O r-\ p in "si' V_^ d:; rH ro u P C o in in ^-p:i ro • • •^ 1 .Q o rCsl CM C U LO m r-1 ^ru D ,Q M C3 rH rfO .^ !h • r•P CO rvi" r-0) ^ en rH -*-H A '^ ID CTi .— • • cn !-H IT/ ro i IX! ro •— > -P V4 (tS fo • m in O — e • • O CD r H X X! w EHi o -H o; !h o 4h b^ -rH O M •p o u C •£) -Si-Di • • en m\ o rH (X) ~^ Kl 00 1.0 '-' .-G 4.) ;r! i/5 • fN m O ^-v g r(U in -sf o rN Q CO 7H ^• ro in ro ^~n| • • • CTi (D in (Ti in rH (3^! CO — iH ro 5h 4-J -*-< d) u • r~~ ro o .— ~ g CM cN ro oj Q r^ eg --' • O O O --S &! rH 0) O O O — ft! O rH ^ -p M l-pt^ • ro t^ o .— g rH (S iH CO O ro Q \0 ro ^-' a (D •— rH (!) fd a ^ X o >1h X X CO E^ 1 ro u P PJ I !:: ro A M O 4h in •H iH rH o JH JJ c o u o o o ^Oj -. ro a) o o o -^ P::; o rH r~l rtJ M ^ Ct:; • rro o ,~v Eii • 'in 0) 1.0 ro o rH ^ Qj I.D ro — • C O o ^ Q\ yo 1.) O O O --^ p:^ mm & ro ^ M !=> • t^ ro O — g • • • CO (D U) ro o rH Q U) ro -— • ro o r-.— > ^ • o 0) ro O 1.0 ro Di UD ro — u p 0) lt~l fA • VD ^ O -— S • • ra; IX) ro o -^ Q [-. CN •— 13 --^ rH ro .o| ^ X O > H X X 00 Bl i 0) (D tn ra •p S-fl 0) o 5-1 ai a Q) .c! -P (D U ,C (D ^ -H rC O 5 (D tJ g P o rH M u ip c: -H k CD r U ra 0) ,c: X Eh • H c X g • p 03 rH •P ro .^^ ro p ^ to C CD (D CQ -rl M +J U ro c P O. CD C -d 0) en G -H rH Cu rH CO tfi CD -H O M n3 rH CD rH v-i -I-J rH < ro <: ra rH ^ 3 U rH ro
PAGE 98

82 I H H o El o:) O M fa a > fa H fa K H H Q >. CO Eh 0:3 < O CU o CO Q fa O O H ri'^ t^ hfO WD Pi Eh cr^ i fa m Eh S iz; fa 0:1 < o H H H o Eh ^j ;-! c^ r,fj :3| H CN u> to CO a) o (Ti VX) IJ3 -— • • • 03 CO m m r^ rC; o ro f,-^ -P • • • CN !HKri CO U5 r-1 O i> H -' r-H r^ o ro ro • • M ^0 O ro Pi 01 iH m vD -— ^^i' r~sf CO r-l Li'T i-H CN --O CO o^ ro <=* c-j 04 en IT) CT) (111 1 CO tH CN (l)| in rH n oil -H CN UD 0^ • 00 O^ ro ^gj • • H H > CO B i i VD VO CTi 'Di • • • r(D 00 CO Cs] —' p:; CN CM '^^ 'fl .1.) P CO • 00 CN — el • • • ^ CD rH CO r-t Q! CO r-H — • c —^ DJ • • • (D C3 rH P4! ^ .H CO >— r~i ro M +J a 0) u • ^-v g ... CO (IJ ID in --^ PirCN • ^~~ Ci! • • • r^ 0) ^-' Pi! in in .i: -p M 2i • .-. g (1) r-H Q en r-{ -~^ ^_^ S -•— 1 Q 1 ^ H \ H H > W EH • C") o o CD O C3 O rH • 000 Ei • • • (U o o o P o CN 000^• • r~) rH • O Oj 0) o Pi r-CN rH ^ g • • • iH a:) CN 03 CTl rH P !^ <--i ~— o rn X— • • in o (O m t-\ CN CN in ^-^ • in o in .—V £| • • • vX) Q) rO CN rH P! CO rH -' !z; Q m u CO Eh fit M f-f H CO Q ro 4J Q) O a, >U-\ i-i > Q) 0) fi M 4-1 ^ V 13 H ffl .c; -73 ^ ;3 g a ^ •r< in = U ^ ro > (U ,c tn -p g H • 13 H CO rH g ijl u (D ^ +J rc: ip C u (U 03 to CD 5H Jh rC 0^ -P V CD m fi g G 3 Q rH T-\ rQ fO u V cn 0) to (D m rC U 0) -P ro rC -rl -P 5 m C (y (D CO -H !h -p !H ro c -P 0. C ^ 0) fl c -rl rH a, rH CO CO 0) -H ri M rH (U rH rH p ^-^ < ro ri: fO rH (U CD !h (D 15 rQ

PAGE 99

83 relationship betwaen party and scale positi-on. It seems clear fha,t this dimension is party cleavage. In the 1967 session, then, \ve v/ere able to identify one factor for each house, and that factor was demonstrated to be party cleavage. The point which raight bear reemphasizing is the fact that urban-rural differences and party loyalties v/ere not part of a developmental sequence here, but rather they tended to be mutually exclusive. Al chough urban representation had been increased, urbanrural differences were displaced by party differences in influencing roll call votes in this session. 1969 Session — In this last session included in our roll call analysis, we were faced with an upsurge in the number of roll calls. In the House 320 roll calls met our ititial criterion that at least 20 per cent vote for the losing side of the roll call. In the Senate 132 roll calls met that criterion. By increasing the required dissent to 35 per cent and allowing only the 3 roll calls on a single bill to be included we narrowed the number dov/n to 100 roll calls in the House, Increasing the required dissent allowed us to reduce the number to 80 in the Senate. Because of the trem.endous expense of attempting to factor analyze 100 roll calls we were forced to reduce the number of rotated factors from seven to four. Since the

PAGE 100

84 first few factors had been of greatest importance in previous sessions we took the risk of rotating fev/er factors with the expectation that our analysis would not be severely hampered. As will be seen shortly, nothing of great importance seems to have been lost. The number of roll calls loading at .50 or better was 12, 4, 2, 5, respectively, for each of the first four factors. In the Senate the pattern was 8, 6, 7, 5, 2, 3, 7 suggesting the operation of more significant variables in influencing roll call voting in that house. Using the techniques and variables used in analysis of previous sessions, we were able to make sense out of only Factors I, II, and IV in the House and Factors I a.nd II in the Senate. Because these were the only factors amenable to subsequent treatment, and because of the need to conserve space, we present summaries of subject matter issues for these factors only. (See Tables 11-22 through 11-25,) These subject matter summaries indicated only the need to probe deeper through alternative techniques. They did not suggest any clear identity for the factors „ As in earlier sessions, we went back to the roll calls and examined their associations with urban-rural, regional, and party differences. As with the 1957 session, party differences seemed to play a significant role in roll call voting. In the 1969 House, 53 per cent of the roll

PAGE 101

85 Table 11-22 SUMMARY OF SUBJECT JUlTTER ISSUES INVOLVED IN ROLL CALLS WITH HIGH L0.AJDING3 ON FACTOR I, HOUSE 1969 Roll Call Factor Numbejr 65 19 18 89 17 13 ,83 79 74 73 .72 .72 -.71 -.70 .69 Loadinq I Subject Matter Provides part of revenues from vehicle tags to be deposited with State Road Dept. Provides for statewide construction of roads not on highway system. (y=61, n=40) Appropriations. Provides for withholding funds from students who violate laws, advocate overthrow of government, etc. Motion to lay on table. (y=45, n=53) Appropriations. Provides for withholding funds from students who violate lav/s, advocate overthrow of government, etc. (y=48, n=49) Deals with county charters. Provides that charter commissions shall not spend funds without authorization from Board of County Commissioners (y=44, n=57) Comiiaunity Development bill. (v=64, n=39) Local self-goverrmient for counties. Motion to accept conference committee report. (y=65, n=4l) Authorization for county adoption of local charters. (y=32, n=54) County self-government. Appointment of conference committee. (y=57, n=47) Reorganization of executive branch. Motion to reconsider vote on amendment to create a Department of Law Enforcement. (y=62, n=42)

PAGE 102

bb Table 11-2 2 Continued Roll Call Number 91 56 82 Factor Loading -.69 .63 -.60 Subject Matter Relates to classification plan for state highway system. Provides for statewide construction of roads not on state highway system. Motion to withdraw from committees on transportation and appropia™ tions (y=53, n=54) Appropriations. To provide appropriations for industry. (y=51, n=54) Mandatory imprisonment for selling drugs to minors. Motion to reconsider. (y=56, n=40) Table 11-23 SUMMARY OF SUBJECT MATTER ISSUES INVOLVED IN ROLL CALLS WITH HIGH LOADINGS ON FACTOR II, HOUSE 1969 Roll Call Number 61 11 3 98 Factor Loading .77 72 .68 .62 Subject Matter Method of fixing millage for property taxation. (y=58, n=53) Deleting date for referenda on millages higher than 10 mill limit. (y=65, n=44) Safety inspection of automobiles. Amendment to change fee to $3.50 rather than $3.00. (y=41, n=64) Executive Reorganization, Amendment to make governor and cabinet head of Department of Natural Resources, (y-53, n=53)

PAGE 103

87 Table 11-24 SUMMARY OF SUBJECT MMTER ISSUES INVOLVED IN ROLL CALLS WITH HIGH LOJiDINGS ON FACTOR IV, HOUSE 1969 Roll Call FBctor Number Loading 8 10 Subject Matter Courthouse and jail for Clay County (y=59, n=39) CoiLrthouse and jail for Clay County (y=69, n=39) Construction of Highland County Facilities (y-69, n-39) Bonds for Highlands County to be financed by race track funds. (y=39, n=39) Authorizes Osceola County to issue revenue certificates. Vote is on an amendment to set a ceiling on amount which county can pledge in payment. (y=54, n=59) calls demonstrated a significant relationship to party differences (compared with the 41 per cent in 1957) In the 1959 Senate party differences were significantly related to more than 38 per cent of the roll calls (compared to the 46 per cent in 1957) Urban-rural differences were significantly related to only 21 per cent of the roll calls in the House (comipared with the 26 per cent in 1967, the more than 39 per cent in 1965, and the 50 per cent in 1963) In the Senate 10 per cent of the roll calls were related to urban-rural differences in 1959 (compared with 10 per cent in

PAGE 104

Table II~25 SUMr4ARY OF SUBJECT 14/\TTER ISSUES INVOLVED IN ROLL CALLS WITH HIGH LOADINGS ON FACTOR I, SENATE 1969 Roll Call NuiTiber Factor Loading Subject Matter 26 -.80 27 -.79 38 3 7 .65 59 -.54 36 62 .61 Rules change to \vithhold all bills from committee of reference after 45th day of regular annual session. Vote is on amendment to require waiver racomm.endations to be reported back to the Senate the day following their receipt by the Rules Committee (y=24, n=17) Rules change. Ainendment to change "shall" to "may" in the last sentence regarding wa.iver motions being referred to the Rules Committee (y-24, n=19) To enumerate criteria for retaining personnel of public schools. (y=24, n=12) Relates to provisions for insuring state buildings. (y=24, n=19) Relating to purchase of insurance by state— allowing Commissioner to institute state self-insurance program that he deems economical — motion to reconsider. (y=24, n=17) Motion to recall "Dirlcsen Memorial." Motion to waive rules and withdraw from Committee on Constitutional Am^endments and place on calendar. (y=12, n=35) To provide for commtinity development. (y=20, n=16)

PAGE 105

89 Table II-2S SUKOVLARY Or SUBJECT MATTER ISSUES INVOLVED IN ROLL CKLLS WITH HIGH LOADINGS ON FACTOR II, SENATE 1969 Roll Call Nuinber Factor Loading Subject ^"fetter 42 -.85 Amendmant to state constitution providing for use of single-meinber districts in legislative apportionraent (y=:17, n=29) 43 .31 Araendment providing for, single-meraber legislative districts. Motion to reconsider. (y=28, n=18) 41 .81 Amendment providing for single-member legislative districts. (y^27, n=16) 65 .68 Limits number of charity days at racing and jai-alai operations to the number authorized in the prior season. (y=19, n=20) 51 .62 Amending Florida Statutes relating to amount of reserves required to be maintained in the fund. (y=23, n=19) 48 .60 To prohibit taking food fish with certain nets within or without Florida waters-tuna exempted. (y=ll, n=24) 1957 and 1955 and v/ith the 2,6 per cent in 1963) Region dem.onstrated a significant relationship to about 47 per cent of the roll calls in the House analysis (compared with the 33 per cent in 1957, with the 36 per cent in 1955, and with the 34 per cent in 1965) and about 21 per cent of the Senate roll calls (compared v/ith the 13 per cent in 1967, with ]2 per cent in 1965, and with the 21 per cent in 1953)

PAGE 106

90 In this session we included one refinement of the party variable. At the beginning of the 1970 session of the legislature, a group emerged in the Rouse known as the Florida Conrarvative Democrats. This group was sufficiently organized and visible that nieiubership lists ware available from newspaper articles. ^"^ Based on these published lists we subdivided the Democrats into our coding so as to examine the role of this group in roll call voting in the 1969 session of the House. When we crosstabulated votes by the tripartite division—Democrat, Conservative Democrat, Republican — we found about 18 per cent of the roll calls in which at least 6 per cent of the Democrats voted in opposition to 60 per cent of the Conservative Democrats. Thus, although the Florida Conservative Democrats did not leave their colleagues in even a majority of the cases, they did play a role in the voting patterns. This role will become clearer v/ith later elem.ents of our analysis. Regional, urban-rural, and party differences showed a relationship to the individual roll call votes as they had in previous sessions in our analysis. We therefore moved on to relate coefficients of association expressing these relationships to the factor loadings. As indicated above, inspection of scores revealed the fact that some of the factors would not be identified by this analysis. That is.

PAGE 107

91 high loading roll calls did not show a clear relationship to any of the coefficients of association. Thus, we focused on factors I, II, and IV in the House and Factors I and II in the Senate. As seen in Table 11-2 7, this analysis does provide sorae help. The correlation coefficients suggest that in the House, Factor I taps party cleavage. Factor II taps regional differences, and Factor III taps urban-rural differences. While the pattern of correlation coefficients makes it quite clear that Factor I in the Senate taps party voting (since the only positive coefficient relates factor loadings to roll call associations with party) Factor II shows a correlation with both urban-rural and regional differences. Partial correlation coefficients provide little help in untangling the influences involved. Controlling for roll call associations with party and region reduces the correlation betv;een loadings and associations with urbanrural differences only from .53 to .37. Similarly, partialing out associations with urban-rural and party reduces the correlation between associations v/ith region and loadings only froxTi .55 to .47. To seek further cleirif ication of the factors underlying these roll calls we moved next to the use of Guttman scale analysis. Selecting those roll calls with loadings of .60 or better, we scaled votes from Factors I and IV in the House

PAGE 108

92 aj B Iz; o H % u OrQ CO CO CO < Pi o iJ Eh S < I H H v^ O Pi o CO M Oi CO I a Eh u w h1 cyi CO cri C5 ru O Eh 2 CO O < K is o CO Eh U CO ^ Q O rtJ K w CO O D O CO m K o Eh H o X H ft! EH H o Q ViD CTi CD W C5 > rH H! rH fci cn M CJ S O (Ji I"!> Lil O H -H -iJ rrH (N l-^J O O > ^ ^d u rH ro fU O O O O CD r--! O iii ( s Pi i-^ G H tH H rH fD tn ^ Q5 O LO o o in ^ U C 00 vD CN] ro o c:> > -H 4-1 rH 'Ci U CD O C O iH O rH ra fO 1 1 fa p; i-:| C rH H rH roo si' o o rro tr. M r-j rH cri o 00 cN ^ O C > -r-', +; O O O r-H o o H ^ I'd a II i r-H ro fC O fe Pi i-q G rH >1 H -H -P I""r-H o v tr; r^ ro 4J 5-i Oil rH O 0^ C-'i CN H c) ra to H -H Pi O CD iH C; O O H rH O 1 i t 1 rH rC CO 4J iX Ul -H < 15 C G rH rH -.H -H fD 4-1 tn <>] O rH CO O CTi U ru QJ si' O rH rH ^0 iH H -H Pi H .H c; O rH O O O O rH rG 1 m 4-> Pi r/1 -H < ^ H fO d u rH 3 rH -H Pi iXi -P ) O CN CM r~00 O u fD a o -sj* CM CN ro r-H -H ro rH ^ rH O O O O O rH !H 1 1 CO p Pi K 4_' "H ^ C HI fH f-H !-i 03 CJI 00 in in CM o CJ G U if) in 00 m o > H fH 13 -P O O O O rH (— i ro U 1 1 o ro Pi i-:i fci a rH H rH ro 00 [^ O CN fO tn u ro rH r~o 00 ^ u G C > -H .p O O O rH O H i-i T! U i 1 1 r-4 ra ro g fa Pi iJ en UD G rH r-H •H +J r-CO o [^ in ro -P !h CM o o r-00 1-) u ro ro -p H -H Ph O O rH O O ro H rH a 1 1 1 c rH fi m 4-> CO Pi m -H G C H H H -H fC! -P tr UD o en 00 in o ro CD 00 O O rH in H H P^ H rH CJ O rH o o o rH ^ 1 1 03 4-> Pi Cfl -H < 5 rH ro G M t-\ O J3 ^ •H p; rti -p f O V.0 f00 ro U ro c O ro CM ro uo H -H ro r-H u ^ rH o o o O iH u 1 1 (n p (0 < .G -P is o H H H H > H H 1-1 i-H > > H O H hH > H HH H H >

PAGE 109

0) 93 r-^ r-l w M (d rS :3 -u O 0) +-' H rH ^3 H rQ 0) -H ra M j^ ^ H ^ M fa rC ^ fd J-J > (D <^ (D ^ 0) .a 4J IH -iJ CO DO • s +> m M (Q -rH a St M X (ti fd • 0) fd C 10 m -73 = -p 0) iH CD M ro -P 'd "^3 0) -P CO C H (U C -H Ti fd o ^1 tH i1 •H rQ N -H r i-W ro -H CO -p CC Kl 4-1 4-1 !-l 0) c ro G CD rH 1 A m t; 4J 0) -H r,-^ d A u tN H +J U 4-> c H A i-d CO 0) i-H H ^ 0) +J >i o; X CO +J .H ra -P u in rH .JQ rH i-i u fd tn fd CL' fd rH Sh C Eh M 0, u (d ro o •H rH 01 g -H ro P-i OJ -P C ..a P^ ro ro 4J 4J rH to -H !-l O 0) m G". d g fH a) a fd jh 4^ M (D CO ^ i-l to u iH (D CO r-i &T ro rC U) 4-* fd !H 4J rH a 5-f •H 0) ^ Q 3 J-l 2 ^ -H •H M -P iJ 4J CO o 1 G a -H CO -pH f-l G) r> m fd G D^ tw rQ r-H ro T3 (K M rH S 0) rH 13 CJ 4-) rH u u Oh ru ro fO ^ c associ shows axis

PAGE 110

34 and Factors I and II in the Senate (Factor II in the House was not included because there were only 4 roll calls with loadings nf .50-~a number we regarded as insufficient for scaling) All scaled v/ith a coefficient of reproducibility of .90 or better with the exception of the scale forined from Factor I in the House analysis. The 12 roll calls dravv'n from that factor produced a scale with reproducibility of .89. The grouping of legislators was clear enough to convince us that this was not an accidental configuration, and that further tarapering with the scale to improve the coefficient of reproducibility was unnecessary. The frequency distributions of scale scores indicate that the scales produced in the 1969 analysis show the kind of polarity which we have seen to be associated with bloc voting (see Table II--28) To see what further light could be brought to bear on the nature of these group polarities we related the scale scores derived to urban-rural, regional, and party differences. Relating the scale types to these variables provides some clarifications of relationships suggested by the earlier correlation analysis. In the House, the correlation analysis had suggested that one dimension of voting was party, another was urbanism-ruralism, and still another was regionalism. Because Factor II provided an inadequate number of high

PAGE 111

95 Table II~28 FREQUENCY DISTRIBUTIONS OF GUTT.m_N SCALE SCORES PRODUCED ON SCALES OF ROLL CALLS LOADING HIGH. ON SELECTED FACTORS, HOUSE AND SENATE, 1959 nouse s cale I Scale IV Score Frequency 53 Percent Score Frequency Percent I 45 I 54 45 II 2 2 II 18 15 III 3 3 III _A1 40 IV 5 5 119 100 V 1 1 VI 4 3 VII 5 4 VIII 5 4 IX 1 1 X 1 1 XI 3 3 XII 5 5 XIII 29 119 24 100 Senate S< i
PAGE 112

96 into the saraa sca.le claEsif ica tion as the regular Democrats, thus showing unity across a publicized fault. Although some cell frequences ere quite small, when controls are imposed the table suggests that this relationship to party cuts across urban-rural and regional differences. The analysis of Scale IV in Table 11-29 shows that these roll calls do tap an urban-rural cleavage. The scale makes clear one point not revealed by the correlation analysis. The urban-rural cleavage suggested in the correlation analysis in reality finds the underlying division legislators from metropolitan counties pitted against legislators from urban and rural counties. The table also suggests more of an interaction between regionalism and urbanism-ruralism than the correlation analysis indicated. Although the relationships are difficult to disentangle and some cell frequences are distressingly low, it seems that the basic pattern emerges from urban-rural cleavage but is softened and altered in some categories by regional differences. In the 1959 Senate also, the analysis of scale types illuminates and confirms the correlation analysis (see Table 11-30) The analysis of Scale I confirms that the set of roll calls loading on Factor I was most influenced by party differences. Also indicated is the fact that while the relationship to party is constant over urban-rural differences.

PAGE 113

ra H CO W P4 o o ^ P-1 M o S i &H P &^ W << > cu H 0^ P CO W iJ Pi sC O (^ CTi U D ; CO Pi 1 i H W is H u 2 Q) tH CO D rQ ru C?i Eh '!> hJ CTi r^ H 125 O W H 02 O i^ 9 O K K o P-4 Eh o > 0.) H H ffi (X! CO O £5 Eh O U H a: ^ r-^ .< 1-3 Q Di <: J3 H I i O CO rC LO rCD .^^ +J • • • ^j. P ID CN H ^ ^ C'J 01 — CO O ra tJl -P (Ti CN C> ^ — • • OT ) "d* O -M 00 00 (-0 — ^1 0-1 4-^ I t CX3 O CN 5-: U3 ro o 00 a! ro o r^ --> • • • ir> 00 o vr> iH en -^ CM OO ^— s: • o o o .--. Cli CN CDl O rH CJl -^ Pil on vo -' > pi CO OJ o -P mi • M f'l ID 00 O ft ul 00 o o o — g o CD o o c; ni" PI CO CNj "^ CU rH ffl 03 a a >! CO &^ u H H H H H > X 0/! o o o D O O in in • LO in o — gj • • • oo 00 rH O M P| 00 IX) —• •, o o o P4 0) o o o cr; Ln in CO r-~ 00 o -~. • • !—i iJD 00 O CM UD 00 ^— • O O O Oi • • CN QJ o LO in 00 di CM r^ --' rH en o g • • ro (U CTi O O •=* p f^ CN ^' fe; (U lii CO Eh O H H H H H > X • O C^ CO --. Q\ CO 0) O CN r~rH Pi! CN r— • CTi GO g • • CD VO 00 P rCN CN o CD Pi o o o .-• o o in in CM ON r^-' • CN CO O Ln '^ \D m m CN • o o o --. pj ^ 0) o o o -^ Pi o • 00 ro <-gj • • "CO (D Kf in o OS' P|CD 00 — 13 CD rH 0) ra a C) > CO Eh H O H H H H > X CO CD Cn ra p c 0) o u CD CD •p ,a o -rH 5 O a CD Eh g rH o o m o CD ra &i o ro -P rc; fl -P Q ra o CD u c CD (D g (D P CO rH (D o ^ U -P (D O 5-1 P ro ra m (D G H -H p P 0) C -H (D iH Q) M +J (D rH o ro rH a iH rH < ra ra u CD p (D 5 H p o aJ a ra S o p 4h ra CJ o p c -H 'O ra o rH I rc: H rC g o p m (D fi p o MH Cfl -H H CD rH ro u CO CD p '0 G) -cJ P rH o c •H (U p ro H H X o -p H g o p iw (Q 0) p o o CO CD rH ro u CO p H to -p c (U n3 C O to CD Pi o

PAGE 114

98 •H o u H H Eh ^ CN =;? in .— ^J -ti J C^) r-i m %'' ^^ ^— ( sj ^^ CO M ro CN! iJD Oi ^-M • • • 0") 4-! r-.-1 rH +>' m CD rH rH rrs C\! CN m •~-' • CM n in ^ & • • CN un -^ o ^ oi ^' rH -sf -Q m I-; '^< in .— G • • rr-H un o") 00 U CO t-H — • o o 03 Q) LD in '^ Q M CN] tX) a "> — ^ Q iH (D H ro P H H u > H H H 00 E-i -. c;::! o„ j .13 I U c •^ ^j Pi rH jc; c --. +J ra • > • l^ P /5 CD -^ }..4 CO P ^-^ in r-CO --M • • •si^ -P U5 •nJ< 00 01 (D CN r-{ in '-• S rH on i> — ro • • OT 5H en in --' 3 on in Pi i-H iC c r-~ rH rH --SH ra • • "^ -P ...Q in rr~rH r; M CO — b P CJ !h in '^ <-\ .-. +J • • • CD Q) rH m Cn fNj a rH H r~•— rH .-ro • • • CTi Jh ~^ 13 Pi rH ^ G in un ,— 4J ro • • CO VI A ^ csi >--• G U CO rH S P rrr,— p. f • in -p vo CD KD rH CD CvJ '^ CSI •—s ^ — ^ a •---' 0) .H (U H ro D H H > H H H cn Eh 1^ H O +J o s o IH Cfl H rH ro o u -H r) CO I -& rl !H 4H CD g M O UH CO -H H H CD CD u 13

PAGE 115

99 i-H ra w Pi w o u E^ iZ! f'^ H |J4 S fa C H Oi Q fa i>^ Q tH > < H ill Pi w n c S < o CO ro fa 1 Di 1-1 H O < H u a w p Pi iH H 1 fU E3S E-i U ffl W Pi p CTi ID CTi |J iH <, s M O F^ H <; o W Pi CO o fa £h o H Oi H H IXl Pi t/) O E3 EH O U H < tH fa f< 1^ Pi H 0) rH u CO ^ fsj O 00 .-!J > • CO p C^l o r^ ~H O tNJ LO C>v! ~-' CO c 1 — ro (Ti U3 ^£) ,-H u • • • CO +J CO IT) '.n rH 0) c ro LO ~— Pi ri:: r-r~ r,--v 4-) • • > (N M VT) \D ^D r-H rH r-\ ID -^ hv 13 rH O O O ^ ro • CM M O O O ^ p O Di rH t ID M C ^n i-O O -— c^ ra [-! ^ r^ cn CA T-\ c M i^M CM in •— fO P rQ M P o o •^ --5-1 0^ -P r-', r~rH r-j d) cn rsi ^'•>-' S in CO CO -— CO CO H Pi!^ rH rH -' > -P VI ro P^ -s}rH IT) ^. S • • CN d) o\ 03 f^i m Q rj \o — "a cu o — rH 0) H ro 0H H O > H H > C/J Eh o crs o ^ & • • o o o o o — Pi rH rj ^4 :3 rt o o o ^ g • '.M Q) o o o — a o rH •, o o o -~. p) • • iri (U o o O '— CEiivD CN (N id ro ,i5 VI p ., n o r.—V e • • • r^j (T' CO in CD rH Q ! H H H > CO Bl 1 j 00 o m 00 • • "cD o CD ^^ Pi OT IT) rH ^: -p :3 CO i> o ro .-^ S • ft • r^j CD CD o ro iH Q rH • uo ro — • CO o CNJ ---. Q k • (T> Q) !> o f\J v_Pi! tfN] rH ro 5h +J id 0) CJ • O r-i C> ^ g • > • CTi O H 03 ^ p! rH CD • o O O --d • i • ^ do o o —• Pi o rC rH P U g5 H CN r^-. • • rH en 00 l>J rH Q r-{ [~• ^ i :-i H > CO B 1 m 0) &i ro 4-> cs M U) &. OJ ^ -p • • ^ H (P o VI •H VI 0) rC rC; ^ P O '^ g fd g • 3 rH CO rH rH p (D tn o u H ro H -P rid tn C o id g (U ro -H a Ct) T3 VI M ro tp CD .id 0^ p rH to fci 1 o; d (P S^ VI g (d ^ pi (U -H u rH rQ rid to CO g 0) rH (D to SH ro !h (!) <4H ro rc; CO -p T3 CO sd (U A o (D g -p H !h M rl M ro ^ -P P-i m C CO 0) id to +J -rH H id rH CD rH to H d Q) -H id U O -d H D4 rH Q ro to rH p o (B f< ro CO Pi ro ^ .Q ri u r-\ ra u
PAGE 116

100 c •H P o o o ro H H GJ H o ri^ CM r-I-^ ,— 4J • CO P C\': \D rH — O! iH y5 ^ C/j c r-l fO r-CO i^l .^ H >-, • • CO tJ -p u:) r^ in rH (D c >X) CM — cri Q) U ^ [^ r-r^-^ jj • • r^i M WD -.0 V.0 .H 'X> iH nH •--• iZi .H O O o ^ m • CN u o O o — p) ID in ,0^ r-H to M :::! c in LO o ^ p:; ro • • • r! ^ VT' ro O H c 5-) r^ (N — fO P A u P in i> CO ^ M • • • CTi 4J '^' o ^ CM (U ro CM ^ •^ s O m in .--. D • • • !>• 1) o C-i r^ f4 in r-\ fO — • > -P M (D d, • o O o ^ g • b • CM (U o in in ro P in CM rsj -— B CD U rH 0) H fO P H U t>\hi H > CO E^ 1 1 H O O O ^-^ ro • • O u o o o --^ 3 Oi f: rCO o ^^ -jj ra ft • 00 ;:! ^ 1.0 ro O ^-' M U3 ro w P O CO ro 00 — V-i • • • in 4-> ro ro ro rH 0) rH rH r~>-' S rH O O o ^ CD • • o U o a o ^ a P< rH fO a o o o ^ 5-1 j • • o -P ,Q o o O rH C u r00 — (1) p u in o in .— M • ft • CO 4~l CM in CM ^-' (1) WD CM rH 2 rH o o O ^ ro • • CM M o o O — 3 in in cr; rC c o o o ^ 4J ro a • • h? i-! A o o o •— S-l o ia p rH o ro r— M • • 1.0 -P O OT 1.0 -^ in ro iH S 5 O rH (IJ H tu & H > H H > CO Eh • H 0) H SH (U f-l .i: 4-> T! u H rH i-\ +J u H H !T> C s -H o d 5-1 ro m rH w 1
PAGE 117

it is modified by regional differences. It is interesting that both Democrats and Republicans in the Southern region are in the intermediate scale types while they are prepoiKler— antly in the polar types in the two other regions. The analysis of Scale II does not go very far in helping to resol^'e the question as to whether region or urban-rural is basic in these roll calls. The regional relationship appears somewhat clearer, and it might therefore be inferred that it is basic. Certainly a degree of interaction between the two variables seems to be indicated. One additional pattern should be underscored because of what it reveals about the relationship of urban-rural differences to party loyalty. ViJe emphasized the fact that the party factor identified in both houses of the 1957 legislature cut across urban-rural differences, and that the same is true for Factor I of both houses in 1959. This relationship should be viewed along with the fact that Scale IV in the 1959 Houses--which we have identified as primarily tapping urban-rural dif ferences-~splits the Democratic party fairly neatly into its "regular" and Florida. Conservative Democratic components (some 50 pec cent of the regular Dem.ocrats are Type III while about 81 per cent of the Conservative Democrats are Type I) Although it would be premature to draw firm conclusions about party and urban-rural differences on the

PAGE 118

basis of the analysis of two sessions, it seems worth hazarding the guess that this Xind of break in Democratic party unity will become more common unless or until their is some shift towards greater homogeneity in Democratic constituencies. In terms of our emphasis on apportionment, the re-emergence of the urban-rural cleavage in the 1959 session is significant because it indicates that party loyalty probably will not permanently displace the influence of these differences. Thus, reapportionment seams to reemerge as a factor of political significance. Perhaps it should also be noted that the five roll calls v/hich went into the construction of Scale IV have a clear comimon substantive content. They all involve authorizations for construction of county facilities. These capital outlay expenditures would be financed against state subsidies. These roll calls place metropolitan legislators, vrhose districts 'benefit less from such subsidies, in opposition to urban and rural county legislators whose districts benefit more. This kind of bread and butter fiscal issue proved to be an insurmountable obstacle to Democratic Party loyalty. Sumiria.r.y of Roll Ca^ll Analysis — Our analysis has provided some indications of the political significance of reapportionment with regard to voting patterns in the B'lorida legislature. As v/e noted in the introduction to the analysis,.

PAGE 119

one imporhant question about the pclitical relevance of rea;op.ortiorjraar.'t VJ'a s v/hether or not the urban-rural cleavage did indeed influence the structure of roll call' voting. V7e were able to identify sets of roll calls with strong relationships to urban-rural differences both before and after reapportionment,. The other diraension of voting which looiried large in roll call voting was party. The presence of this line of division in the legislature is itself not unrelated to the fact of reapportionment. One study has indicated that by applying the 1957 county-by--county legislative election returns to the 1955 apportionment plan it can be demonstrated that continued malapportionment would have reduced the nuirber of Republicans in the Senate by five and the po number in the House by eight." This seems to be a very conservative means of estimating the partisan implications of reapportionment. Given the fact that the number of Republicans in the Senate actually jumped from two in 1955 to twenty in the first session after reapportionm^ent (1957) and that the number in the House similarly jumped from ten to thirty-nine, it seems safe to speculate that the disruption of existing elite structures by the redrawing of district lines offered a, golden opportunity for challenge of which the Republicans took good advantage.

PAGE 120

104 Assessing the relative weight to by given to these implications of reappor tionrnant involves an exercise in judgment rather than objective measurement, bxit this kind of judgment seenis iraportant to make. In noting the fact that the urban-rural cleavage underlies roll calls both before and after reapportionment, two limiting factors should be considered. First, fewer roll calls seera to have their base in urban-rural differences by the 1969 session of the legislature. Second, the urban-rural differences which appear in the 1969 session place metropolitan legislators in opposition to urban and rural legislators. This may mean a kind of rural-suburban coalition ivhich would probably be as conservative as the old rural bloc. The rise of party voting, along with a boost in the metropolitan representation amiong the Deraocrats, probably hold the greater political significance | and the greater portent for change in Florida legislative politics. CgiTCTiittee .§t:Qic;bur_e A nalys is — A treatment of the political significance of reapportionm.ent would not be complete if we relied exclusively oii roll call analysis. While understanding the voting alignments and their relationship to urban-rural cleavage is essential, there are other politically significant aspects of legislative activity which need to be included in cur picture. Other than floor voting.

PAGE 121

105 the point of decision-making which is commonly regarded as next most important is in the committees. The committee system serves as a filter through v/hich political issues must pass on their way to floor voting. Moreover, since committees have had a greater tradition of real influence in 29 Florida than in som.e other state legislatures, it seems important to turn our atte^ntion to them. Our examination of the effect of reapportionment upon the comjiiittee system focuses on several questions. First, we want to know if the committee system reinforced and/or exaggerated the rural over-representation which characterized the legislature before 1967. That is, given the distribution of seats, we v/anted to know if any group, shared disproportionately in influential committees. Second, we are interested in v;hether or not reapportionment caused a shift in the assignment of coiTmiittee positions which led to a power shift from rural to urban legislators. In other words, what effect on power relations resulted from simply changing the distribution of seats so that urban, rather than rural, legislators predominated? VJe first looked to see if rural domination went beyond a surplus of seats. Then we looked at the consequences of reapportionment for the committee system. The first phase of our committee power structure analysis rests on the assumption that the number of committee

PAGE 122

106 assigmvients is an important aspect of power for an individuai legislator or a legislative group. Presumably one way to limit the power of a group wcnild be to limit the nurnber of positions which they hold through Vv'hich they might exercise influence. We might expect then that if urban-rural cleavage was a significant factor affecting legislative behavior (as our roll call analysis suggests that it was) then the dominant rural group might try to limit the influence of the more urban groups by limiting their access to committee positions. In order to test this hypothesis we calculated the number of committee assignments per person for our urban-rural and regional groupings. These calculations for the House are shown in Table 11-31 and for the Senate in Table 11-32, The data in these tables cast real doubt on a rural bloc domination hypothesis. The variation by regional or urban-rural grouping is quite small for both houses. The only really striking vdeviation occurs in the House in the 1965 session. There the Central region shows a depressed number of committee assignm.ents compared to either the Northern or Southern regions of the state. It seems unlikely, however, that this is really related to urban-rural cleavage. In the 1963 session, the rural members had almost two more committee memberships per person than did the metropolitan legislators. This discrepancy is corrected in the 1965 session, however.

PAGE 123

Table 11-31 HOUSE COMMITTEE MEMBERSHIPS AS RELATED TO REGION AImD URBANISM-RURALISM, 1963-1969 .-CimimittLgi Reqion Year Nor hh Central South Total 1963 5.47 5.11 5.21 5,28 1965 5.40 3.53 5.21 5.24 1957 5.34 5.66 5.43 5.41 1969 2.91 3.02 3.02 2.99 Urban-Rural Year Metro Urban Rural Total 1963 4,85 5.15 6.82 5.28 1965 5.35 5.33 5.29 5.24 1967 5,42 5.38 5.44 5.41 1969 2,95 3.03 3.00 2.99 Table 11-32 SENATE COI'#IITTEE MEMBERSHIPS AS REI,ATED TO REGION AND URBAN ISM-RURAL ISM, 1963-1969 CoiTiraittee Meiuberships Per Person Reqion Year North Central South 1963 1965 1967 1959 9.35 10.60 6.00 2.83 10.08 9.25 5.67 2,68 10.08 10.42 5.83 2.45 Urban-Rural Year Metro Urban Rural 1963 1965 1967 1969 9.78 10.33 6.00 2.59 9.53 10.00 6.07 2.71 9.91 10.24 3.67 2.50

PAGE 124

108 •which V/3S hold before re-irpportionrnexit We reject, therefore, the notion that a rural bloc b-attre3so:d its political power in the legislature by holding a disproportionare share of the committee assignments. It might well be arg-aed, however, that an analysis of coiTimittee assignments and reapportionment v/hich treats all committees as equal in power is unrealistic. Recognizing that this was true, we developed a procedure whereby committees might be given unequal pov/er v/eights or influence score. The easiest v/ay to derive such a set of values is to use a. set of expert judges. This is what we did. In this case the most obvious experts on the relative influence of various coiranittees would be legislators themselves. However, not all legislators would be equally qualified to assess power. Moreover, regional or urban-rural differences might affect perceptions of coirimittee influence. Taking these things into consideration, we selected from each house, for each session, a group of legislators making up about twenty per cent of the legislature to serve as expert judges of relative committee influence. Each of the judges was asked to assign each committee a, number from one to five representing his estimate of the committee's importance or influence in the legislature. For each committee the influence values assigned by the judges v/ere summed to produce a total value.

PAGE 125

109 The comiTiitteas were then rariTieci according to the total influence values. This ranking v.'as divided into fifths. The conmiittees in the top fifth of the ranking v'ere assigned an influence score of 5, those in the next fifth a score of 4, those in the next fifth 3, those in the next fifth 2, and those in the ].ov/ast fifty were assigned a score of 1, Thus each comrriittee, for each house, for each session was assigned an influence score from 1 to 5. With a method of assigning differential power weights to comrnittea assignments we could probe somewhat more deeply into cha,nges in the committee assignment process and their relationships to reapportionment. Our next step was to compute committee assignments par person by regional and urban-rural groupings talcing into account the influence score of the corr-anittea assignments. The distributions of comiTiittee assignments per person by urban-rural and regional classifications and influence score are shown by year in Tables II~33 and II~34 for the House. These tables do not suggest any clear patteri^'. of rural dominance of the high value corajiiittee assignments or exclusion of urban and metropolitan members from such high value committees. In fact, the one variation which seems significant is in the Urban group in 1965. Urban legisla.tors had 2.39 committee meirfoerships per person at the value 5 vyhile Metro had only

PAGE 126

110 Table 11-33 GOM^IITTEE ASSIGNMENTS PER PERSON, BY INFLUENCE SCORE AND URBAN "RURAL, HOUSE, 1953-1969 Year High 5 Low 1 Metro Urban Total 1953 1.11 0.66 1.09 0.87 1.13 4.85 1965 1.59 0.88 1.12 0.80 0.95 5.35 1957 1.26 1.09 1.55 0.75 0.77 5.42 1969 0.68 0.92 0.52 0.44 0.40 2.95 1963 1.97 0.91 0.91 0,79 0.56 5.15 1965 2.39 1.05 0.83 0.78 0.28 5.33 1967 1.50 1.00 1.29 0.83 0.75 5.38 1969 0.83 0.55 0.59 0.48 0.59 3.03 Rural 1963 1.43 1.11 1.25 1.21 0.71 6.82 1965 1.12 1.00 1.20 1.16 0.31 5.29 1967 1.28 0.94 1.67 0.61 0.94 5.44 1959 1.00 0,47 0.47 0.67 0.40 3.00 1.59 per person and Rural had only 1.12 per person at that value. Not only does this fall short of any pattern of systeinatic bias in the committee assignment process, but it also occurs in an unpredicted grouping. The data on power of Gomraittee assignments and urban-rural and regional groupings for the Senate is found in Tables 11-35 and 11-35. Again, there is little to suggest any bloc domination of the high influence committee assignments. The next step in our examination of committee assignments and power involved the calculation of overunder/

PAGE 127

Ill Table II~34 COMMITTEE ASSIGNMENTS PER PERSON, BY INFLUENCE SCORE AND REGION, HOUSE, 1963-1959 High Low 1 North South Total 1963 1.79 0,96 1.09 1.06 0.57 5.47 1965 1.86 0.81 1.28 1.12 0.33 5.40 1967 1.28 1.00 1.44 0.66 0.97 5.34 1959 0.94 0.63 0.56 0.41 0.38 2.91 Central 1963 1.46 0.97 0.91 1.00 0.77 5.11 1955 1.19 0.72 0,67 0.51 0.53 3.63 1967 1.54 1.05 1.39 0.95 0.73 5.65 1969 0.72 0.84 0.49 0.51 0.47 3.02 1953 1.09 0.88 1.26 0.84 1.14 5.21 1955 1.39 1.08 1.05 0.93 0.76 5.21 1957 1.33 1.09 1.&7 0.53 0.72 5.43 1959 0.56 0.84 0.55 0.50 0.48 3.02 Cc nribined 1953 1.45 0.94 1.10 0.97 0.82 5.28 1965 1.64 0.96 1.11 0.94 0.59 5.24 1967 1.31 1.05 1.51 0.75 0.79 5.41 1969 0.75 0.78 0.53 0.48 0.45 2.99 representation scores for committee assignments of a given influence score. The under-over/representation score is calculated by subtracting the ratio of the actual to the expected proportion of committee assignments at a given influence score for each urban-rural or regional classification from one. The expected proportion is the same as the

PAGE 128

112 Table 11-35 COiyLMITTEPJ ASSIGNMiNTS PER PERSON, BY INFLUENCE SCORE KSD URBAN-RURAL, SENATE, 1963-1959 Year High 5 4 Low 1 Total Metro 1963 2.44 2.44 1.78 1.33 1.78 9.78 1965 3.00 1.89 2.44 2.00 1.33 10.33 1967 1.52 1.10 1.26 1.10 1.03 6.00 1969 0.62 0.96 0.31 0.17 0.52 2.59 Urban 1953 2.40 1.87 1.73 1.80 1.73 9.53 1965 3.20 1.43 1.79 2.29 1.29 10.00 1967 2.00 0.86 0.93 1.57 0.71 6.07 19S9 1.65 1.06 0.24 0.24 0.53 2.71 Rural 1953 2.52 1.52 2.48 1.62 1.76 9.91 1955 2.62 2.24 2.14 1.24 2.00 10.24 1967 0.67 1.00 0.33 0.67 1.00 3.67 1969 0.50 0.50 1.00 0.50 0.00 2.50 proportion of the total membership falling into a given classification. The actual proportion is the proportion of the coiTunittee assignments at a given influence score which fall into a specified classification. For example, if metropolitan legislators made up as m.uch as 23 per cent of the House mcembership and we found that exactly 23 per cent of the committee memberships with an influence score of five were allocated to the metropolitan members, then the over/ under-representation score would be .00 for that crossclassification. It should be noted that this procedure is

PAGE 129

11 Table 11-36 COMvlITTEE ASSIGNMENTS PER PERSON.. BY INFLUENCE SCORE AND REGION, SENATE, 1963--1969 Year High Low 1 Tota 1 North 1963 2.25 1.80 2.15 1.30 1.95 9.35 1965 2.70 2,05 2.30 1.50 2.05 10.60 1967 1.83 1,00 1.08 1.17 0.92 6.00 1969 0.75 0.83 0.33 0.25 0.67 2.83 Central 1963 2,86 1.77 2.08 1.69 1.69 1965 2.92 1.50 1.93 1.75 1.25 1967 1.44 0.89 1.11 1.33 0.89 1969 0.56 1.17 0.22 0,22 0.50 10.08 9.25 5.67 2.68 c )Outh 1963 2.42 1.92 2.00 2 0.8 1.67 10.08 1965 3.17 1.83 2.00 2.08 1.33 10.42 1967 1.50 1.11 1.11 1.18 0.94 5.83 1969 0.61 0.89 0.39 0.17 0.39 2.45 basically the same as that used to compute expected values in the chi square statistic. The procedure thus has some 3 1 Standing in the field of statistics. The over/under-representation scores for the House for each session in our study are shown in Tables 11-37 and II38. The regional data have an erratic look to them. They show no strong or consistent relationship between committee assignment over/under-representation and region. While the urban-rural data do not support a simple rural dominance -.1. ~,^i^::^y^

PAGE 130

114 Table 11-37 distribution of 0\?ti:.r/under~-pje;presehtation scores for comj:>iittee assigistments by influence score and by URBZ-^iN -RURAL, HOUSE, 1963-1969 Year High 5 Low 1 Metro 1963 -.24 -.30 .00 -.10 + .37 1965 -.16 -.08 + .01 -.15 + .62 1967 -.04 + .04 + .02 + .01 -.03 1969 -.10 + .18 -.02 -.08 -.11 Urban 1963 + .35 + .16 -.17 -.18 -.32 1965 + .45 + .11 -.26 -.17 -.47 1967 + .14 -.05 -.15 + .11 -.05 1969 + .09 -.30 + .11 + .01 + .32 Rura" L 1963 -.02 + .19 + .14 + .24 -.14 1965 -.01 + .05 + .08 + .23 -.47 1967 -.03 -.10 + .11 -.18 + .20 1969 + .33 -.40 -.12 + .39 -.10 hypothesis, they do reveal some' interesting variations. A clear pattern emerges if only the highest influence scores ^ are examined. At the highest influence scores, the clearest pattern that emerges is that the metropolitan legislators have been consistently under ~r epresonted The urban legislators have consistently been over-represented. In both cases the tendency has decreased throughout the study period with the degree of under/over-representation dropping ,.

PAGE 131

115 Table 11-38 DISTRIBUTION OF OVER/UKUER-REPRESENTATION SCORES FOR COMMITTEE ASSIGNMENTS BY INFLUENCE SCORE AND BY REGION, HOUSE, 1963-1969 Year High 5 North Central South Low 1 1963 + .23 + .02 -.01 + .10 -.30 1965 + .13 -.15 + .15 +.19 -.45 1967 -.06 -.05 -.05 -.12 + .33 1969 + ,24 -.20 + .06 -.15 -.16 1963 .00 -.42 -.16 + .04 -.06 1965 .00 + .05 -.16 -.24 + .26 1967 + .20 .00 -.08 + .27 -.08 1969 -.05 + .07 -.07 + .07 + .05 1963 -.25 -.05 + .15 -.13 + .38 1965 -.14 + .13 -.05 -.01 + .29 1967 -.05 + .03 + .11 -.16 -.09 1969 -.13 + .08 + .04 _.14 + .07 sharply in the first reapportioned session in 1957, The rural group has not been seriously over or under-represented except in 1959 — the second session after reapportionment. We should note, however, that this is only an expression of over-representation, not an overall raeasure of power. In assessing these findings concerning the rural group, we must recall that they declined in power because of the decline in their membership in the house. Even if over-represented

PAGE 132

-LX on influential coiniriittees in 1959, the rural grouis was losing power. We would have to conclude, overall, that the data suggest a pattern of metropolitan-outstate cleavage which gradually declined. No clear pattern of urbanrural differences appears in the coiranittee assignment process in the House. The over/under-representation scores for the Senate coramittee assignments are shown in Tables 11-39 and 11-40. These break down the over/under-representation scores by regional and urban-rural classifications. An examination of these tables does not lead to any conclusions of bloc dominance. The variations seem to be more random and unsystematic than those in the House. We would conclude, therefore, that the committee assignment process in the Senate has operated during our study period on the basis of factors not included in our analysis, if any systematic pattern exists at all. Our analysis thus far does not lead us to conclude that the coirimittee assignment system systematically exaggerated the dominance of a rural bloc (beyond its surplus of seats) nor that it perpetuated rural power after reapportionment. We should not, however, be stampeded into a, quick conclusion that reapportionment had no effect on the committee system. We m.ust remind ourselves that simiply by changing

PAGE 133

117 Table 11-39 DISTRIBUTION OF OVER/QNDER-REPRESENTATION SCORES FOR COMJNnTTEE ASSIGNI'iENTS BY INFLUEMCE SCORE AND BY URBAI^I-RURAL, SENATE, 1953-1969 Year High 5 4 ....J Low 1 Met] lO 1963 -.01 + .34 -.14 -.18 + .02 1965 + .04 -.16 + .17 + .15 -.19 1957 -.06 + .07 + .14 -.09 + .10 1969 -.01 -.01 -.01 -.17 + .03 Urb< an 1953 -.02 + .03 -.17 + .11 -.01 1965 + .11 -.22 -.14 + .32 -.18 1967 + .25 -.15 -.15 + .30 -.24 1969 + .04 + .08 -.25 + .13 + .34 Rural 1953 + .02 -.16 + .18 .00 .00 1955 -.09 + .21 + .02 -.28 + .23 1967* -.59 -.03 -.70 -.45 + .05 1959* -.21 -.50 +2.16 + 1.38 + 3.20 *Because of a small n, quite small absolute variations in actual from expected values produce large over/underrepresentation scores. the mix of urban versus rural legislators reapportionment could have a profound impa.ct on the committee system. We can turn row to a direct consideration of this change. We would emphasize that the analysis preceding this has taken the distribution of seats among urban-rural and regional groupings as a given ai"id has looked for evidence that any

PAGE 134

118 Table 11-40 DISTRIBUTION OF O^TER/QiroER-REPRESEIsfTATION SCORES FOR COMMITTEE ASSIGN.MENTS BY INFLUENCE SCORE AJMD BY REGION, SENATE, 1963-1969 Year High 5 Low 1 North 1963 -.09 -.01 + .03 -.20 + .05 1955 -.07 + .11 + .10 -.13 + .25 1967 + .17 .00 -.02 -.05 .00 1969 + .20 -.15 + .07 + .20 + .33 Central 1963 + .15 -.03 -.01 + .04 -.04 1965 + .01 -.19 -,12 + .01 -.24 1967 -.07 -.11 + .01 + .09 -.03 1969 -.11 + .19 -.29 + .07 .00 Sou th 1963 -.02 + .05 -.04 + .29 -.05 1965 + .10 .00 -.04 + .21 -.15 1967 -.04 + .11 + .01 -.05 + .03 1969 -.02 -.09 + .25 -.20 -.22 of these groupings managed to exaggerate their power beyond any advantage or disadvantage suffered in the distribution of seats. V7e now back up to examine the distribution of seats and its effect on the access to influential positions a. group might have. Since the number of members a particular group has in a legislative body acts as a constraint on the number of influential positions which can be occupied, it would be wrong to ignore the distribution of seats .per se

PAGE 135

'^ 119 In order to examine the effect of reapportionment's redistribxition of seats on the access to power in corninittees v'/e calculated for each seBcion the percentage of the coinir.ittoe assignments at a given influence score held by each of our three urban-rural and regional groupings. These calculations" provide us with informa.tion to answer a question such as,, "Of all corniTiittee assigniTients rated five (high) how many v;ere held by legislators from metropolitan counties? urban coxanties? rural counties?" We v/ould anticipate, of course, that v/ith reapportionment urban and metropolitan legislators would account for a. larger proportion of all committee memberships. This in turn v/ould increase their power potential. For the House, the data on per cent of committee assignments held at various influence levels are shown by urbanrural groupings in Table 11-41 and by regional groupings in Table 11-42. The table for urban-rural groupings shows some dramatic before ana after reapportionment changes. Most of the change appears in the metropolitan and rural groupings with much less change appearing in the urban classification. The important thing which has happened after reapportionment is that the metropolitan group increased its total committee m.embership dramatica.lly while the rural group diminished

PAGE 136

120 Table 11-41 PER CENT OF COMMITTEE ASSIGNMENTS, TOTAL AND BY INFLUENCE SCORE, FOR IJP.3J\N"RUR/^J. GROfJPINGS, HOU; 1953-1969 High — Low Total Year 5 4 3 — 2 1 Per Cent N Metro 1963 28.6 26.5 37.2 33.9 51.5 34.5 228 1965 37.0 40.2 44.4 37.1 71.2 43.0 252 1967 62.2 67.2 66.1 65.2 62.8 64.8 417 1969 56.7 74.2 61.9 57.9 56.6 62.3 222 Urban 1963 35.8 31,6 22.6 22.3 18 4 27.4 181 1965 23.4 17.8 11.9 13.3 7.6 16.4 96 1967 23.1 19.2 17.2 22.4 19.1 20.0 129 1969 26.7 17.2 27.0 24.6 32.1 24.7 88 Rura 1 1963 34.6 41.9 40.1 43.8 30.1 38.0 251 1965 39.7 42.1 43.5 49.5 21.2 40.6 238 1967 14.7 13.6 16.7 12.4 18.1 15.2 98 1969 16,7 7.5 11,1 17.5 11.3 12.6 45 sharply. The ta.bla malces it clear that this change affects comm.ittee assignments of all influence scores in about the same v-^'ay rather than shcv/ing up disproportionately in high or low value assignments. Nevertheless, this has the effect of increasing the raetropolitan share of top value committee assignments (those with value of 5) from just 37 per cent in 1965 to over 52 per cent in 1957. Conversely, the rural share of top value assignments drops from about 40

PAGE 137

121 Table 11-42 PER CEKT OF COMMITTEE ASSIGITMSHTS TOTAL AND BY INFLUENCE SCORE, FOR REGIONAL GROUPINGS, HOUSE, 1963-1969 Year High o Lov/ 1 Total Per Cent N North 1963 46.233. 5 37.2 41.3 26.2 38.9 257 1965 43.5 32.7 44.4 45.7 21.2 39.6 232 1967 25.3 25.6 25.6 23.6 33.0 26.1 171 1959 33,3 21.5 28.6 22.8 22.6 26.1 93 1963 28.0 16.2 23.4 1965 27.7 29.0 23.4 1957 38.0 34.5 31.7 1969 ... 34.4 38.7 33.3 Central South 1963 35,8 32.5 39.4 29.8 47.6 33.9 224 1955 28.8 38.3 32.3 33.3 43.9 33.8 198 1957 24.4 40.0 42.8 32.6 35.1 38.2 250 1969 32.2 39.8 38.1 38.6 39.6 37.4 133 per cent in 1965 to around 15 per cent in 1967. These data suggest, therefore, a sharp shift in power in coira-nittees from rural to metropolitan areas as a simple by-product of adjusting the relative representation of these areas in the House Table 11-42 shows a similar, but much less marked shift from North Florida to South Florida in committee power. The fact that the shiftsare less dramatic than those in the

PAGE 138

122 urban-rureil table suggests that the change is really a rural to iFietropolitan shift, and that it appears in the regional breal^downs because of the fact that South Florida is more urbanized than North Florida. The same kind of data for the Senate is presented in Tables 11-43 and 11-44. It seems unnecessary to belabor the point that the same pattern appears here as in the House data. The shift is from rural to metropolitan and from North to South with the urban-rural shift appearing more dramatic and profound. Our analysis of the committee system leads finally to two conclusions. On the one hand we were unable to find clear evidence of urban-rural or regional groupings being able to dominate the committee assignment process and systematically over-represent themselves in high influence committees. On the other hand, we note that simply by shifting the distribution of seats reapportionment had a profound effect on the nature of the committee system. This is a miatter of no little significance. .J Conc l usions and Impli cations — In previous research there have been conflicting results concerning the importance of urban-rura.l cleavage in legislative politics. Derge stimulated new interest in this area several years ago when his I findings found the cleavage to be of minor sicrnif icance in \

PAGE 139

123 Table 11-43 PER CENT OF COM24ITTEE AS S I GNMS i^TS TOTAL AND BY INFLUENCE SCORE, FOR URBAN-RURAL GROUPINGS, SENATE, 1963-1969 Yea.r Hi.gh Low 1 Total Per Cent N Metro 1953 19.8 26.8 17-1 16.4 20.3 20.0 88 1965 21.3 17.3 23.9 23.7 16.7 20.8 93 1967 61.0 69.4 73.6 58.6 71.1 65.0 186 1969 50.0 59.6 60.0 50.0 62,5 59.6 75 Urban 1963 32,4 34.1 27.7 37.0 32.9 32.6 143 1965 3 5.4 24.7 27.2 42.1 26.0 31.3 140 1967 36.4 24.5 24.5 37.9 22.2 30.1 85 1969 36.7 38.3 26.7 40.0 47.5 36.5 46 Rura 1 1963 47.7 39.0 55.3 46.6 46.8 47.4 208 1965 43 .,3 58.0 48.9 34.2 58.3 48.0 215 1967 2.6 6.1 1.9 3.4 6.7 3.9 11 1969 3,3 2.1 13.3 10.0 0.0 4.0 5 32 roll call voting in Illinois and Missouri. By Gontra.st a recent study of the Georgia legislature indicated some increase in the degree of urban-rural voting although it could not unambiguously be related to reapportionment -^-^ Our findings deraonstrated that the urban-rural cleavage was significant in roll call voting both before and after reapportionment. We would also say the urban-rural cleavage has probably lost some significance because of the rise of

PAGE 140

124 Tab.le 11-44 PER CENT OF COMMITTEE ASSIGNMENTS, TOTAL AMD BY INFLUENCE SCORE, FOR REGIONAL GROUPINGS, SENATE, 1953-1969 Yaar High 5 Low 1 Total Per Cent N North 1953 40.5 43.9 45.7 35.6 45.8 42.6 187 1965 42.5 50.6 50.0 39.5 56.9 47.3 212 1957 29.3 25.0 24.5 23.7 25.0 25.8 72 1969 30.0 21.3 26.7 30.0 33.3 27.0 34 Central 1963 33.3 28.0 28.7 30.1 27.8 29.8 131 1955 27.6 22.2 23.9 27.6 20.8 24.8 111 1957 34.7 33.3 37.7 40.7 35.4 35.6 102 1969 33.3 44.7 26.7 40.0 37.5 38.1 48 South 1953 26.1 28.0 25.5 34.2 25.3 27.5 121 1965 29.9 27.2 25.1 32.9 23.2 27.9 125 1957 35.0 41.7 37.7 35.6 38.6 37.6 105 1959 36.7 34.0 46.7 30.0 29.2 34.9 44 partisan conflict which has tended to displace it. The mixed conclusions, to which ours are added, indicate the need for case studies to give way to comparative legislative behavior studies before the conditions under which the urbanrural variable has significance can be clarified. Our finding concerning the displacement of urban-rural conflict by partisan conflict adds to previous similar findings. Buchanan found for California, Grumm for Kansas,

PAGE 141

1.25 and Wiggins fcr lowa"^ that significant increases in the strength of one party tends to increase the incidence of party voting. This certainly was the case in the Florida legislature in the 1967 session, when the Republicans made their greatest increases, and continvi.ed into the 1969 sesgion. The phenomenon of party voting cutting across urbanrural differences resembles a pattern which Salisbury observed in Missouri, In that case Salisbury attributed this phenomenon to v/hat he called the political blandness of the state; i.e., the fact that industrialization had not advanced to the point where real social class conflicts corresponded with urban-rural differences. Moreover, he points out that the parties failed to develop distinctive 37 ideologies. Florida socio-economic development has been described in similar terms, and we might, therefore, speculate that Florida parties v/ould remain unattuned to ideological differences for some time to come. The significance of our findings, in relationship to the question of the political importance of apportionment systems, is to lend credence to the reformers' belief that this was both a legal-moral question and a political question. While the point has been fairly well documented that reapportionment is not syste^-oatically related to certain

PAGE 142

125 types of polxcy outputs for all states,"^ this research does not prove that apportionment is never a significant political variable. We have already seen that research reports differ considerably from state to state concerning the importance of urban-rural differences. It may be that the political significance of apportionment depends upon whether or not this is a significant line of cleavage. If it is, as we have demonstrated for Florida and Hawkins has for Georgia, then apportionment assumes political significance. Particularly should the pattern of partisan conflict in Florida begin to disintegrate once the newness of political opposition wears off, urban-rural differences may re-emerge as the primary line of cleavage and the nature of apportionment v/ould assume even greater importance. We believe that our findings and those of Hawkins foreclose any simple, premature dismissal of the importance of apportionment.

PAGE 143

Footnotes to Chapter II ^Kobert S. Friedman, "The Urban-Rural Conflict Revisited," E§.gJL§i^Ii Ji^2liti-cai £ii§-dL^^ -'^^'^ (June, 1961), 481-495. ''For a discussion of the problems of power structure studies see Willis D. Hawley and Frederick M. Wirt (eds.). The Search for. Coia muni ty P ov/e r (Englewood Cliffs: PrenticeHall, 1963) especially the articles in Parts IV and V. "^Duncan MacRae, Jr Issues and Parties in. Le gislat ive Voting: Me thods c>f. Statistical Analysis (New York: Harper & Row, 1970) p. 91. '^Harry H. Harinan, Modern Facto r Analysis Rev. Ed, (Chicago: University of Chicago Press, 1967) See chapter six. ^ Ibi d. pp. 14-15, ^IM.4-/ P136. "^MacRae, p. 131, See his footnote 1. ^M^cRae, p. 41. ^John G, Grumm, "A Factor Analysis of Legislative Behavior, Midv/est Journa_l of Political Science, VII (November, 1963), 339. ^^MacRae, p. 260; Grumm, p. 344. "^ "Region" as it is used in this study is a kind of shorthand for subcultural groupings in the state of Florida which vary from the fairly conservative Old South North, through the moderate Central area, and finally to the more liberal area of South Florida. It is because we conceive the regional variable in these terms that it ordered for purposes of calculating coefficients of association. For a discussion of regional subcultures in Florida which provides some justification for this way of looking at region see William C. Havard and Loren P. Beth, The Politics of. Mis "Representation (Baton Rouge: Louisiana State University Press, 1962), especially chapter two.

PAGE 144

128 13 The Repablice^n party did not hold enough seats m the legislature to rr^ake the party variable useful before the 1967 session. Party is included in our analysis, therefore, only in the 1957 and 1959 sessions. ^^MacRae, p. 223. For a brief discussion of the problem of variables which load lowly, their impact on factor scores, and a method of avoiding this problem see Ira Sharkansky and Richard I. Hofferbert, "Dim.ensions of State Politics, Economics, and Public Policy, American Pol iti cal Science Review LXIII (Septe-mber, 1969), 867-679. See footnote 9 on p. 859. For a general treatment of the use of Guttman scaling in roll call analysis and a discussion of the two approaches to its use see Lee F. Anderson, et al,. Legi sl ature Roll Call Analysis (Evanston, 111.: Northwestern University Press, 1955). See especially chapter six. '-Y7, J. Dixon (ed.) Bi omed ical Computer Programs (Berkeley: University of California Press, 1968) pp. 169184. 17 IBM, Syst em /36.Q Scientific Sjj^jciCLiitiiias Rg^ia-gs (White Plains, N. Y, : IBM, Technical Publications Department, 1958). For a sample factor analysis program see pp. 429-431. ^^Raymond B. Cattell, "The Meaning and Strategic Use of Factor Analysis, in Raymond Cattell (ed.) Handbook of Multivariate Exp erimenta l Psych ology (Chicago: Rand McNally, 1955), pp. 174-243, The point made here is at page 232. At this point in the analysis we were using the BMD03M program which has a maxim.um capacity of eighty variables. The coding of the 1969 roll calls had not yet been done, so we had not yet become av;are of the fact that the increase in the number of roll calls in that session v>?ould raake continued use of this program infeasible. Later we developed the larger program described above with a larger variable capacity. We did not re-run the 1953 data because of the expense involved. Further, we believed that the roll calls in this analysis were sufficiently representative of the entire set to make re-running with the entire set unnecessary.

PAGE 145

129 20 An eigenvalue measures the proportion of the variance in a matrix Vv-hich is accounted for by a given vector. More specifically, an eigenvalue of 1.00 indicates that the vector accounts for 1/Nth of the matrix variance. In this case, an eigenvalue of 1.00 represents 180th of the variance. See MacRae, p. 144, 21 The coding of counties on these two variables follows the criteria set forth here. All Standard Metropolitan Statistical Areas are classified as '"Metropolitan." counties with 40,000 population or greater, but which are not SMSA s are classified as "Urban." "Rural includes counties with under 40,000 population. The "North" includes Alachua, Baker, Bay, Bradford, Calhoun, Clay, Columbia, Dixie, Duval, Escambia, Flagler, Franklin, Gadsden, Gilchrist, Gulf, Hamilton, Holmes, Jackson, Jeff erson, Lafayette, Leon, Levy, Liberty, Madison, Nassau, Okaloosa, Putnam, St. Johns, Santa Rosa, Suwannee, Taylor, Union, Wakulla, Walton, and Washington countiesThe "Central" region includes Brevard, Citrus, Hernando, Hillsborough, Lake, Marion, Orange, Osceola, Pasco, Pinellas, Polk, Seminole, Sumter, and Volusia counties. The "South" includes Broward, Charlotte, Collier, Dale, Desoto, Glades, Hardee, Hendry, Highlands, Indian River, Lee, Manatee, Martin, Monroe, Okeechobee, Palm Beach, St. Lucie, and Sarasota counties ^"^Cramer's V, the actual measured for the tables crosstabulating "yeas" and "nays" against region and urban-rural differences, is based on chi square. It is designed to measure association in N x N tables. For a discussion see Hubert M. Blalock, Social Statistics (New York: McGraw-Hill, 1960), p. 230. We recognize that the problem of multicollinearity is involved in this analysis since degree of association between roll calls and urban-rural differences correlates at .55 with degree of association between roll calls and regional differences. Distinguishing unambiguously between the effects of two correlated independent variables is difficult and sometimes impossible. See Hubert M. Blalock, Jr., Causa_l Inferences in None xper im enta 1 Res earc h (Chapel Hill: University of North Carolina Press, 1954), pp. 87-90, for a discussion of the problem. The pattern of partialling, however, suggests a developmental sequence leading from region to urban-rural to roll call voting. This sequence squares with previous research on Florida politics. See Havard and Beth, pp. 20-21.

PAGE 146

130 24 *Havard and Beth, p. 16. ^Ssignif leant means, here as above, producing chi square values significant at the .05 level or better. 26This is in contrast to a line of research which has related legislative party systems and party voting to differences in constituency makeup. For a clear statement of this approach see Heinz Eulau, "The Ecological Basis of Party Systems: The Case of Ohio, Michi^st Jouriia^jL of Political Science I (August, 1957), 125-135. For otlier examples see Thomas "a. Flinn, "Party Responsibility in the States: Some Causal Factors, i-jae rican P oli tical Siii_ence. Reviev{_j_ LVIII (March, 1964), 60-71; Thomas R. Dye, "Comparison of Constituency Influences in the Upper and Lower Chambers of the State Legislature," Mst erii P olJJ:J,c,a 1 Qua^t erj^ XIV (June, 1961), 473-480; Duncan MacRae, Jr., "The Relationship Between Roll Call Votes and Constituencies in the Massachusetts House of Representatives, American Politic al Science Re_view, XLVI (December, 1962), 1046-1055. 27Barbara Frye, "Coalition Flexes Muscle," Tampa Tri bune (April 14, 1970) 6-A. ^^Manning J. Dauer, "Florida is Different; So Is Her Politics Different," pp. 35-36. .Unpublished manuscript. ^^Havard and Beth, p. 148. ^'^These data were drecvm from the Florida Data Bank of the University of Florida Political Data Laboratory. The actual selection of judges was done by Dr. Manning J. Dauer. Drawing upon his years of observation of the Florida legislature, he chose legislators on the basis of their perceptiveness At the same time he tried to maintain a regional and urban-rural representativeness in his selection. The collection and coding of these data was carried out over a period of several years. Unfortunately precise records concerning selection procedures have not been retained. It was possible to perform statistical tests of interjudge agreement only in the case of the 1969 House and Senate rankings. The average Pearson's product moment correlation coefficient considering all pairs of judges was ,46 for the Senate and .31 for the House. ^^For a discussion of the method used to calculate expected values in chi square see Blalock, So cial S tat xs tics, pp, 215-216.

PAGE 147

131 ""'David Derge, "Metropolitan and Outstate Alignments m Illinois and Missouri Legislative Delegations," _Ameri Gan Political Science Review, "LII (Deceifnber, 1958), 1051-1065. ,. Brett W. Hawkins, "Consequences of Reapportionment in Georgia," in Ricliard I. Hofferbert and Ira Sharkansky (eds.), S^tate arid. Urban Politic s (Boston: Little, Brown, 1971) pp. 2 73-298. ^'^^Villiam Buchanan, Legislative Par tisanship : The Devian t Case of Califo rnia (Berkeley: University of California Press, 1963) See especially chapter eight. ^ Gruiran, p. 3 54. "^^Charles V7iggins, "Party Politics in the Iowa Legislature," Midwest Journal of Political Science XI (February, 1967), 86-97. -J ~ "Robert H, Salisbury, Missouri Politics and S tat e Political Systems (Bureau of Government Research, University of Missouri, 1959) 38 Manning J. Dauer, pp. 4-11. See literature reviewed above in chapter I.

PAGE 148

Chapter III ^ POLICY OUTPUT ANALYSIS — I At the basis of much of the reform writing concerning reapportionment was the assumption that the mode of apportionment was systematically related to the performance of the legislature. The expectation, at least implicit, in these v/ritings was that fundamental changes in legislative output should accompany a redistribution of legislative seats in accordance with population. A study of the effects of reapportionment on the Florida legislature v^ould not be complete, therefore, without giving som.e attention to the policy outputs of the legislative body. What sorts of changes were expected? Malcolm Jewell offers this list which is reasonably representative: A legislature dominated by mieiribers from the farm^s, towns, and small cities is not likely to be sympathetic to the needs of the larger cities. Home rule legislation and laws to deal with the specific problems of the larger cities--~slum clearance, metropolitan transit, annexation, for example — often fail because of the ignorance and indifference of rural legislators. Urban underrepresentation affects the outcome of votes on labor and welfare legislations and in the South on segregation questions. Cities are often seriously handicapped by the state legislature in the type of taxes they can levy 132

PAGE 149

and the sisa of tax rates perraitted. Perhaps the most direct evidence of discriraination against urban areas lias in the formulas established by state legislatures for the distribution of state aid or of certain proportions of state-collected taxes. When state aid is distributed in greater amounts to counties with lov/er total assessed property, urban areas sometimes suffer because their, rate of assessment is generally higher. The distribution of school funds often neglects the needs of urban areas, particularly those growing school districts burdened by heavy construction costs. Probably the most widespread discrimination is in the distribution of some portion of the gasoline tax for road-building. The formula used often favors rural counties because it gives little weight to population, while in some states none of the tax receipts are available to cities for their roads. We should note the fact that Jewell includes in his list a. sat of factors fairly directly related to taxing and spending (allowable forms of taxation, revenue distribution formulas, expenditures for schools, etc.) and a set not so directly related t.o state fiscal affairs (home rule authorizations, authorizations to tahe part in various kinds of federa,! programs, etc.). In the analysis of the policy consequences of apportionment we want to give some attention to both of these types of legislative output. In this chapter we intend to deal with fiscal or expenditure rratters. In the next chapter ve will deal with nonexpenditure types of output. In this treatment we will be dealing with two broad areas: (1) an analysis of state taxing and spending during the period of

PAGE 150

134 the reapjportionment revolution, and (2) an analysis of the state of Florida's impact on the capability of its local governments to itiobilize resources to deal with their problems adequately. State Expenditure Analysis Methodology -In this part of the analysis we tried to come to grips with the question as to what, if any, consequences malapportionment has for the way a state taxes and spends. There is the implicit assumption in Jewell's analysis that there are real differences between city dwellers and others in the state as consumers of governmental services. If this is true, v;e might expect these different service demands to be expressed in state action after reapportionment. To examine this possibility we selected a fairly small set of taxing and spending variables for analysis. Selection of taxing and spending variables for analysis was influenced by several considerations. First, variables were chosen because there was reason to suspect that they might bear soma relationship to urban-rural differences. Second, variables were included which had been used in a number of other studies in hopes that there might be the 2 possibility of comparing with the findings of those studies. For a given policy area, these variables are intended to measure several things: (l) absolute level of expenditure.

PAGE 151

135 (2) expenditure effort in relation to df;-,raand (per capita measures) (3) expenditure effort in rcilationship to resources (expenditures expressed as percentages of personal incorae) and (4) priority ranking of expenditures in resource allocations (expenditures expressed as percentages of total expenditures) We can note in Jewell's statement the assumption that city-dwellers require more and different types of governmental services than do others. With the exception of highway policy, we therefore propose the following hypotheses for our variables: (1) as states change from malapportioned to fairly apportioned legislatures they will increase their expenditures significantly (i.e., they will try to provide more services) (2) per capita, expenditures will increase to meet service needs after reapportionment (3) expenditures will rise in relationship to personal income as reapportioned states mobilize resources to meet service needs (4) priority changes v/ould be expressed by greater parts of total expenditure being taken up by these services. Highway expenditures would not necessarily change in the manner predicted by these propositions. Rural areas have traditionally had need of highways for marketing purposes.

PAGE 152

In attempting to analyze change in patterns of expenditure in Florida and relate it to reapportionment, we encountered the problem of trying to differentiate between '^normal" change and exceptional change such as might occur from important change in the political system. We would expect, if for no other reason than persistent price inflation, that expenditures v^ould go up' over time. Given this expectation, how could we distinguish this type of change from the kind of increases predicted by reapportionment reformers? The problem is analogous to the problem in experimental research of controlling for the influence of extraneous variables. In experimental research the prescribed solution is to use a control group which does not receive the experimental treatment as a baseline against which to measure the effects of the experimental treatment. We needed, therefore, a statistical analogue of the control group — some baseline against which to com;^are. To acquire this we useddata on expenditures from all of the states. We try, by using all of the states, to develop norms against which we can evaluate changes in Florida expenditures. To get at the pattern of changes in our expenditure variables over the years of the study period we used regression analysis on the expenditure data for all of the states. Working through pairs of years, we set up the

PAGE 153

137 regression equation so that we predicted each year's expenditure by the previous year's expenditure. This provided several measures which we could put to use. The correlation coefficient produced indicated the degree to v;hich the distribution of the states from one year to the next was stable. A notable drop in the correlation coefficient from one year to the next would indicate some changing of positions on the variable was. taking place among the states. If the coefficient remained high we knew that the states were maintaining their relative positions. The second measure of interest to us from this analysis v/as the beta coefficient. The beta coefficient (which is the "b" term in the equation 'y = a + bx) is a measure of the slope of the line passed through the bivariate distribution of points (in this case representing expenditures for the states for the given two years). This beta coefficient, or measure of slope, provided us with information about the rate of change between the two years in question. ^ If the beta coefficient was large, it meant that the amount of / change was great; if it was small the change was small. A third type of information provided by the regression analysis was a set of predicted expei;iditure values derived from the equation. These values were predicted for one year from the previous year's expenditures taking into consideration

PAGE 154

the overall pattern of change among the states. Since this regression line was estimated by a formula which miniiiiizes deviations of the actual values from it, it seemed reasonable to taKe it as a measure of normal change and use it as a baseline for evaluating Florida's changes on our expenditure variables. By using the line estimated by the regression equation as an estimate of normal change, we had a means of distinguishing between it and more extraordinary movements. This baseline made it possible to examine Florida for patterns of spurts on the expenditure variables. We, of course, wanted to see if these spurts came soon after reapportionment. In the next stage of the analysis, we tried to focus m.ore directly on the r-ela tionship of apportionment to the expenditure variables. We could have no confidence in whatever relationships we might find between apportionment and expenditures unless we had some notion about what was happening v/ith other political variables of possible relevance. We did not want to leave open the possibility that any changes we v;ere interested in attributing to apportionment were in fact the"' result of change in such things as partisan control of the legislature or governorship of a state. We collected, therefore, measures of partisan control of the legislature, change in partisan control of the legislature, change in

PAGE 155

139 either legislative bouse in partisan control, partisan control of the governorship, partisan control of the legislature, and per cent of the legislative seats held by the minority party for each h.ouse of the legislature as well as measures of apportionment, (See Table III-l for a listing of the political variables.) Since this part of the analysis was not at the crux of our research problem, we did not expend the effort required to develop sophisticated measures of these variables. With the exception of the measure of within-legislature-competitiveness (per cent of seats held by the minority party) all of the variables have been "dummyed up" to simple dichotoraous measures. As will be seen below, because of this crudity in measurement, we imposed some rather stringent rules of interpretation on our correlation coefficients relating these variables to state expenditures In this stage of our analysis, then, we computed zeroorder correlations between each expenditure variable and the political variables for each year. In interpreting these correlation coefficients we were interested not only in the value of the coefficient but also in the consistency of those correlations from one year to the next. If a variable is related to another in a direct and causal sense, then a change in one should produce a change in the other. This

PAGE 156

140 Table III-l POLITICAL VARIABLES ENTERED INTO STATISTICAL ANALYSIS OF EXPENDITURES Variable Name Partisan Control of the Legislature Change in Partisan Control of the Legislature Change in Partisan Control of the Lower House Change in Partisan Control of the Upper House Partisan Control of the Governorship Change in Partisan Control of the Governorship Congruence of Partisan Control of Legislature and Governorship Appor t ionment Minority Strength in Lower Legislative House Minority Strength in Upper Legislative House Possible Values Democrat-Republican Change -No Change Change -No Change Change -No Change Democrat-Republican Change-No Change United Control-Divided Control Unacceptable apportionment-court acceptable apportionment Per Cent of seats held by minority partyPer Cent of seats held by minority party

PAGE 157

141 means then, that if the variables are really related tha value should remain about the same over repeated measurements and certainly the sign should remain the same. Flip-flopping signs on correlation coefficients signal the fact that the variables are moving independently of one another. Thus we were interested both in the size and pattern of correlation coefficients for a variable for each year of the analysis. But more than the relationships between political variables and the absolute value of expenditure variables we wanted to know something about change in expenditures and its relationship to reapportionment. One way of approaching this would have been to compute a gain score for each variable for adjacent pairs of years and examine the 4 relationship of this gain score to independent variables. We found the arguments against this presented by George W. Bohrnstedt to be compelling, and, therefore, adopted an alternative method suggested by him. Bornstedt argues that the use of gain scores over-corrects for the influence of previous standing on a variable. In his view either partial correlation coefficients or partial slopes should be used since they have the effect of setting the covariance between two measures of a dependent variable at zero, which is the real desired effect. The choice between partial correlations and partial slopes depends upon whether or not one is troubled

PAGE 158

142 by the standardization of coefficients to the sample v/hich is characteristic of correlation coefficients but not n3cessarily characteristic of regression coefficients. Since vie were working v/ith the population of all states (except for Nebraska and Minnesota which v/ere dropped out because of the inapplicability of our political variables) and because of greater simplicity of interpretation, we chose to use partial correlation coefficients. To suminarize, our method of analyzing change involved computing partial correlations between a given year's expenditure variable and each of the political variables for the previous year, partialling out the previous year's values on the expenditure. This partial coefficient becomies, in effect, a correlation between the political variables entered into the analysis and the cha nge in the expenditure variable between time "t-1" and time "t." In interpreting the coefficients produced by this analysis, we' had to lean even heavier on the notion of pattern than was true with regard to the zero-order correlations. Sharkansky has well documented the fact that the correlation between current and previous expenditures is quite high, especially as the two measures come closer in time. We would expect, then, a reduction in the size of the partial correlation coefficients simply because of the drastically reduced amount

PAGE 159

143 of variance in the dependent, variables. As a result, we had to rely on size of coefficients relative to others for given variables and, raore specially, upon the consistency of relationships expressed in the coefficients from one change period to the next for criteria of interpretation. Debt and Taxes.. With this explanation of the methodology set forth, we can now present the analysis of individual policy areas. The first area we want to discuss is that of 'tax and indebtedness. The particular dependent variables which we chose here were per capita tax revenue, per capita. general revenue, per capita debt; and tax revenue as a per cent of personal income. This policy area goes directly to the question of the state's willingness to tap private resources to produce public fgoods. Since it is generally expected that urban areas are in greater need of public services than are more rural areas, it might be expected that tax efforts and indebtedness would be related to apportiorjnent, i.e., that change could be expected with reapportionruent As the first phase of our analysis we looked at the relationships of the distribution of the states on each variable for pairs of a..djacent years. Table III-2 presents the correlation coefficients and slope measures (beta coefficients) for this set of variables. Looking first at the

PAGE 160

144 CO Pi H ^ i rH 1—1 rH rH ro to u ft; M X OD CO E^ t-in 0) ro M O o cri Oi CTi ft Eh • ft • • ro $3 H in VD m o LD rH ro iH ri) a^ o o C?i 0^ ra a • • • (D 0.) (^^ r(T^ 00 CD KO 'X (Ti CTi en cyi (Ti ^-1 M • > • • • y. o fO E-i ^ ro CD rcri Lfl in U3 P 0. o o O o O iH • • a • • Oj +) ,H rH H H r-i rH ro rQ W U O "H C>^ CTi cn CO 0^ £i >.o •^ KO 10 ro 1 1 1 1 U 03 o-i ^ LO VD r-H >H 3 '-0 UD ^.D xi ro C> (Ti C3^ cr> CTi ft iH M rH rH rH o H P fU H CJ H (0 U (U M Q) :^ •H 'd i -P <4-l H M-l rH H

PAGE 161

145 correlation coefficients we find that the states maintain their standings relative to one another to a very high degree (the correlation coefficients are high for all variables throughout the period) When we look at the slope measures -v/e find a pattern of gentle rise during the period, Jin other words the slopes and the correlation coefficients tell us that although, taxes and indebtedness were rising gradually during the period, there were not many changes in the rankings of the states on any of these variables. Next, we looked at the relationship of taxes and indebtedness in this period to ourpolitical variables. Table III~3 presents the zero-order correlations between our political variables and the tax and debt variables for each year. For each of the variables, except tax revenue as a per cent of personal income, the coefficients are low and/or unstable (i.e., signs change from year to year). In the case of tax revenue expressed as a per cent of personal income, there is some consistency of correlations with the measures of party competition in the legislature (per cent of seats held by the minority party) with the valines ranging 7 from lov; CO moderate in strength. The very range of the ^ correlations over the years argues strongly against making much of this relationship. Nevertheless, it might suggest some hesitancy to tax and incurr indebtedness v;hen the

PAGE 162

146 CO iH I m CO Eh Q P Eh 1 H H CO H H > O H t^ H i-q O H H pq O H E-i < Pi Pi O u Cn •M (U I -p y\ o j-i o u-i fS -p ft; O I o -H 4J -U M C O 0) O !h O P S -i-j -p M P CO m O 0) B o •Pj Du ro Id ,H CO -H (J (l) 4-> +J -H &i M C O Q) a M H +J s; en CO O) 4J C u m to A Cm O tfl (D -rH D^ -p c:; 53 •r-l fi r-Q ro i-H (D o;! O bl -H 5-t C -P -P .13 M C ^ oi O O Oj u C fd 01 O rA V) P -P fD CM o ro -^^^ Oi cn n O O Cs! o I I I CO VD -J 00 O rH O O O CM V.0 1-4 li'^ rCM rO CM O O CNJ o rro V.0 v£) iH CM O O O r-l O rH E I c?i cr. r.T UO 1— I r-i o CM CN cm; 00 rH (N CT\ ^' 00 "si' -^ r-J rH o VD U-i O O CD l~CM fM rH ^ ro CO CO in r~ CN t.D o O rH O CN rH ro I i I I I I o d H CO 00 ro OHO O ro rH rH sf 00 O CN O CN •sf "^ UO CN CN rH IX) o^ Lfi r^ in rH iH rH O CN CN C^ 00 "si' CN vo ro O CM O rH rH CM u o o O'j -sT c?i in rH rH O rH o to Pk 01 •P: I I i i i (D (D > p:; O • CN rH in <^ ><; O rH O CM CN ro ro CM Ol O KO rH O rH O O I III O CSl •^j' 'q-' rH in O O rH rH rH CM u Ml 03 Pm| ml m ro > X ro till ro •P 00 o 0) c > (D ro CM r-00 CM 5^ O rH rH O gj o C7^ V£) CM O ro, u 5H o H ro CN t^ rH H H O ro O o in o 00 CM O ro p -H ro U !h CU I i I I I I I I I CM ro O O rH O -r-^ CM 00 O CN I I I O O^ '^' CO LlO rH O O CM O CM CM CD oo r-o oj 00 ^ O yo (T< r-i O O O O 00 r-i. CM O rH O CM rH OJ rH "^ rH in st' CO 00 C^l O O rH i I I I CO CN ro >* O CO nI* "^ ^ rH rH CM I I •^ r-\ K£> CM r-J 00 O -* rH rH rH iH I in o uD in t^ cTi rH H rH CM CM rH i I I I I I CO ^ in IX) r-^ CO \X) ^ VD ^.O VT) U5 t5> cTi cTi CTi cj^ cn r-\ r-{ rH r-i z-{ r-{ I I I i I I ro ^ in V5 [^ 00 V£) VO 0 VO VD 0> cn O^ rH t — I — \ f — I { — i E — I I I 00 -^ in ix> r-00 IX) UD VD U3 UJ 'X) cn cTi cn cTi o^ cTi I — 1 rH t — 1 t — i — I I — I

PAGE 163

147 •a s -H iH O u I H H Eh G Oi CO fi •H COr-> 4J 5-i (D !~f O M fO fl ?3 a^ !-) •P QJ fO m > r-H U) w -iH OJ tn D 14-1 0) S 1-1 (U ^ rH 'Ci M a Cr. u ro C -p d: o u to •H| -H > rH A3 M U G .1.) ro & i '^ 6 -p M a, c •H O +J -H o u o rH fi o u o u 4-) o u c fa D1 •>H ra Q M O > o o u CM -^ (Ti r-CN in O O O CM rH o t.0 iH CM O rH CO rn ro iH H o iH iH r~~ CO nH rH Ci rH n-l O 0\ CTi O O O O ^ O IT) rsi o i-H I t I I I I I I I I I roj o Ln 1X1 P' O nH o u Q) a +J > m s^ O O O u M I I o c H H CO u 0) tw; -p iH ro C O CM in c3^ •si' O O iH ffl i > rH V 00 rH CO in rH CTI rH in CTi r-1 vo ijD CTi rH (7. rH 00 rH fO rH >* \D .H m rH \£) CTi rH rH CO IX) en rH

PAGE 164

148 tn rH > tfj flj -H 13 -H UH a ro (D > 03 CO tn U) '^ tn • ti 0) ^ C m ^ -H 'r-i G = >1 >i 03 ra 4-1 tn ra 4-1 03 rC ai fD 4-1 4-1 ^ -H p 03 •p p .£< rH U W 4J 03 13 03 03 UH ro rH = ra fO ro 03 ro 5= "• fO a cu 15 > C ra 5 03 CD > T? 03 tn 03 ro u C Q) >i >i C •H 03 p ^ ;3 0) ^d • 4J 4-1 & 4-1 03 ro 4-1 P 05 -!-> rC 'u 03 03 H -r4 -r4 ro 03 S ro 4-1 p fO -P t: O en 0) M U tn 4-1 ro 4J ra 4-1 H fC p 3 (n CO 03 rH ^ ro w Q) r\ C m (D 03 4-1 03 4-1 •H >i u ^ £ -r-; fU 0' ro ra •H .i::; 03 0^. -i-J r-t ^ tn 4-1 CD -H "? 5-1 M u 4-1 03 4-1 H 5-1 03 0) Q' 03 0) H Xi ^^ M-) 13 > 0^ >i^ S^ "^ >i 4J 03 4-> (D -n O D.< 4-> P 4-3 ra A 13 P rC ,13 tn H, 3 c U 4-1 4-1 C •?[ P P g 0) C >i >! fO ra 'J3 ro 4-1 T! P CD OJ 03 ^ ,Q rH = a> 03 ^ c C ro H rC ,c r^ a • 13 i-H C P 3 "i-t ra U -p 4-1 ^ 13 CM o rC c c O fl 'TJ > [H H 4-1 : H 4J ro ro > H W rH (D C 01 0) ,c 4-1 ,G 03 03 -P +J ra CO +J -P rC rG 03 03 ro G •H ^ tn a nj rC 3 g P u ro C 4-1 03 4-1 Q T3 ^3 T3 (33 03 c M V O ra P P u CO OQ CD ^ 03 03 K) 03 o ro -H 03 ro "+4 fi c G •p -P P P -H > g rH •H a< Xi H 4-1 en i 0? 0! a rC p^ 2 ^ rH 03 P rH -}-' 4-> 5 a fO 13 4-1 ,c! = 03 Cb tfi G P P O a OJ ro tn u ^-^ a H >( CD .-a O .Q rQ rH i-j 13 c iw C Oi 42 -P Ph 15 03 tn 03 ro 13 P (U H (D ; o 0) P -H ra ^ -rl >i <-{ EH M CD H M ,3 rQ rG rC 0. W 03 5 O 13 4-1 ra >i rC (D 4J +J 03 ro -H > 4J r-1 -P 5-1 rd o ro ^ p 4-1 CU P p rO P IS M !-i c C o o -H O ro g '•^ G A •H ro 03 p G C ^ n f, ^ h-! 03 4-1 U-i 5-4 o P 03 ro 4-1 ra U 03 cn 03 44 Q} c 13 P g m tfl en P 4-1 5-1 ^ p > -H 13 rd, T3 = C fO ro ro fO ro P O 13 C fC 4-1 rH B ra O GJ 03 C > tn M G cy -^ m 03 03 03 o: U ro o a 13 tn >i ^ : 'p-i g g rH tn •H > C -H ^ -P = ra n: 03 03 r: cu rd ^ ro 03 W M • CQ w rC! ^ ro 4-1 4-1 03 tn "03 13 p p ro CM 4-1 4J ro C p p ro (J rH 04 = 03 03 4-3 UH rH f-i ra rC rC! 1+4 m p rC o •H C rC a 03 rH u > a (13 -p 4J O o g P 4-1 p ra o •H p ^ p p 13 5 P g 0) iH tfl tn 03 03 4-1 03 U C > MH > 4-) 0) ^ 03 rc ro ro cn tn -H C ra O -H G p -p CTi > ;$ 3: ra ra 5: .a •H tn = tn 4-1 p G 4-> 4J 4-1 r 13 ro u (D Tj to (U tn tP C C 03 -H rH C oj 4-1 ^ CD .^ ^ fi C G) 03 0) 5 ^ = rC ro = rG 03 4-' S O 4-> H •H V U 4-1 4-1 jj 4-1 p = ijl ^^3 ti M 5-1 ro = • m -H 13 O 03 01 4-J 03 144 P IH rH P m rG S CM H Cfl El 03 u CJ 04 fli Cfi H rH h-{ •-{ H 4-1 = ro ^j .Q C u 'cs 03 m 'cy ^ rC ro -H ro •r-i fO -H ^ > > p >i P 4-1 P P CO U) 4-1 Ci4 rH >

PAGE 165

149 opposition can use the fact against you in campaigns. There seems to be no relationship between these variables and reapportionirient. The correlation coefficients are quite low and the signs changeable. This strongly suggests that what variation there was from year to year was independent of the reapportioning of state legislatures. Although we have demonstrated that these political variables show little or no relationship to the tax and indebtedness variables on a year-by-year basis, it might be possible that they would correlate with the amount of change. Therefore we pressed on to look at the partial coefficients which are presented in Table III-4. It seems safe to say, however, that there are no reliable relationships expressed here. The coefficients are low or unstable for all of the variables. For all the states, we have shown that there was not a systematic covariance between our tax and debt m^easures and political measures (including reapportionment) either in absolute values or in change values (expressed in the partials) It is still possible, however, that the variable which did not systematically affect all states can have impact in a single state. We looked next, therefore, at Florida's pattern on these variables in the residual tables of the regression analysis. This is presented in Table III-5

PAGE 166

150 o a !, 1 H c: p-j t4 -H tri iT^ r-o ^ ^ rv^ CO in rn-^ ro c^j 00 o "* n en KO VD o +J +J .H^OCNO OOrHCC rHOnHiHO OOOrHO r^ M q 1' s' i' r r r r r r r i* i r r i Eh (D S Oj g i O Oj 1 ^ < RES 0) iw 0) >i,c; ffi -P -P O'^vOiHO onmooinro 'jJ'cncnmco cq in ^ r~id i M !-! c; r-i rH rH CS! OrviroHO oorooo 0000 rH 1 1 Q Q) a. c ^ en a. -H jj 0} CQ P S c/5 'K &3 (D c Q ^^1 CO H ^' ri pi CD S co^ >1 rC! '-' ^ b1 SH IXl +J 4-1 r-LOiH-^OO i^CTiUD^'J^r-CNVDrHLTlCn iH^V£>CNiH -H CTi 00000 fl OOCNOrH rHOC^JOO • • • 1 • (U CN Pi-P ^ H H ^:j H ^ cn ^^ 5-1 JH fi (D • u 1 1 w 1 P r 1' :? a !-i 5-t fl m -rH -P 0) f a) a '^ m a) 4-1 P. m K 5-1 P Q 4J VX3 .^CTi-JtEHr.yDC^Lnfl 05 c s > ro CD 4-1 in • ^ \D rH ^ \0 <( p^ <'"^ QJ 5-i -H !ji •d^ ^1^ --^ -^. ^^. ^. -: r'^ rH n rH „-. >-, ^ a 1 CD • • • HJ 1 41 1 4-) a fO ^ U (^ •r(0 4J To: TICAL VIOUS fiortyfC D P^ O u 0) roi ^ •H Q) CM JV & w fOl 5;^ >-< POLI R PRE for : ^' 13 C tn H !H ^ (U <^i^OrHO rHCO-^^VDrnCM 0:1 CD (D r-1 CN r1.D 13 rH CN VD rM -H tn rH C OOOrHCM rHOOOrH OOOCJU a O i-r; (D -P C (D 1 1 tH .^„5: M ra II 1 > 1 I ra .i^ OJ i-:i Pm a CrJ H X m •H C^ re EH H -2 S nj r-t (D w L'toq'vDooo >xiv.ocncMijn r-coinror.HfMLnrHr;&1 -H M rH rH ,-H rH CN) O rH O O ^ ^ ^ rH ^ ra S-i c { i \ r 1 1' •H < ^. ra 1^ 04 8 U3 o-|i,OnH(NCri n-sfiHvDrH mrocTiior-OrarHrHV^ c i -H M 4-J 4-1 rH rH rH (M O O O O rH rH O O O rH CN U U ra r r r i* i" r r r i i i i i 4-1 H ra 04 CM !-i s;'inwDr--co ^irivDi>03 ^unuDr-co -^ m kd rCO ra KO '^ ^ ^ ^0 iXJV.OUDin^O v^VDvr)^D>^ U5U5iXi^.OV.O C13 o>a^cncri(Ti crlCTl(TlC^^o^ cjicncricTicri o^crio^cyicj^ >^ rH rH rH r-^ rH rH rH iH jH rH r-H rH rH rH rH r^ t-\ T-K r^ r-\

PAGE 167

.51 T3 •-< 5 •r-i b o "* H H H (U rH fO d ft iii -iH •rH tfl -p i-l O-i -H M O -P ra M Q) (0 u MH > 4-1 O ro O rH en U <4H -rl C O !Ti (D ,d OT 4-) 5-1 d fi-i iH d 01 !h -rH 0) P > J-l fO u (U n O (Ti r-i r~{ O i I I o 4J ,Q O -p -H C4 fU U M (D ll4 rH ^ CO O O rH cy\ CN rH CD Ln o o o o o O "Nf O CTi C5^ r-m w5 cTi [~^ r-i O O O O O O rH O O I I n 0) O u d H d o 0) u CI, 4H O 4J d EH 3 rol col rcii CO 0) d > XI ra D O M CM CO UD O O fO o o O O rH O O I i I I I I I I I M ^ ^ in VO rm ^ IT) V.0 rCO ^ LO VD r^ CD •=* in VD r-00 m VD 10 ^O VO VO VO VO ViD VD VD VO V£) VD VD VD VD VD VD VD VD (1) cTi fVi m o-^ m n> m cn C3^ cyi 0^ 0^ a^ 01 Ol 0^ 0^ 01 01 01 J>4 x-\ rH rH rH rH r-\ rH rH rH rH rH rH rH t-H rH r-{ rH rH rH rH

PAGE 168

152 O t5 rH H n5 • .' = ri 0) p fl tn -H f>J iH .^ .fl -r4 M = (Q P fl (d 44 4-1 W ro iTJ iH tn > 05 > tn ^ fd -H T! lij 4-^ .-3 ro > 05 05 b CQ 5 O • OQ rfl fl ro 0) •H iji: >l >( ro 4-1 tn 5 M ro .C S2 • 4J 4-> rfl -H p ^ -P 13 03 iH i-l U 4J 05 'n 05 05 :-\ ,c: = ro fO ro 05 4J > P ra o & Dh T) > C ro ra 4-1 (D > ij tP 4J rfl ro U 'd OJ >i >l fl -H 05 05 p 4-1 5 0) P. T3 • • -P 4-1 tn 4-1 05 ro P rfl 03 4J Xi fO 0) -H •H -H ro 05 ^ 0' 4J 4-J (D -P en m M M 05 4J ro .fl ro rH P 03 05 4-1 1— 1 A .fl 05 in (u c r-i fi3 05 4-1 05 4-1 •H >i = u ^ ^ H fd fd td 05 •H p tn -P 0) B g rfl 5 4-1 tJ tn P (1) -H (W [5 !-i u 5-1 4J 05 c rH M CD fd iH fl T3 5: a rfl rfl 5 >i 4J rfl fl p 0) -i-i (U (B a 4J 4-1 4-1 fd rfl -C fa 3 tn tH 13 fl U 4J 4-1 fl ,fl > 4J g rH a >1 >1 ra ra 03 fd 4J ro rd ,Q ^ iH r a, 05 0) .fl tn 'a C fo > fii ^ ^ Oh • 'd H 4-J p> •H o +J 4-> TJ T3 (N u r^ fl 4-1 'O Q) rH rH 4-1 z H Pl ra P y, ^. c m t-i A (u fl 4J ^ ra 05 H 4-> •H -P o 4J in 4-1 4J ,fl rfl ro c a, -P tn rCJ P g P p ro fl 05 ^ c (D C! T! ^3 fl rH u ro P fl m CO Q) ^ 0} m O ro -H 05 tn • •r4 CJ C d 4J 4J :3 •H > g rH -H fl rfl rH 0) (D fd tn (D H -H 4J /5 4J fd = 2 "^ ^. a -H G g S rfl ,fl M Q P P rfl rfl Dh 1 +J -M [fl r4 •H rfl Oh S u 4-1 Dh H r~n to rH rH 5-1 M Dh 4-1 PU P H m x) ro >i a ri:; Pi 13 .Q rH H 3 rH 4-> 4-> s Oh ro TJ Pi -H (d >i 0-> ifl c P Dh ro tn ^ 44 > 44 (D >i (D fd fii ^ •-{ 13 t3 fl 1+4 fl P ,-i +J Oh ^ c tn 05 15 Oj ro re •H z 4-J •H fd ^ H >i .fl Ph M QJ rH Vi 1q ,Q .fl rfl Dh 05 05 ^ V rfl 4-1 +4 E-i r^ 0) 4J 4-J 05 ro 05 •H •n 4-1 M rC u fd ^ M 4J P P 'd rC fO P 15 M M fl fl u O 4-1 £ m G •H -H (D p fl fl fl -r^ fl H f^ M-l m M P P ro tn >i fd U 0) tn m 4-1 c 'a g H ^ tn cn CO 4-) 4J M S M > H > 05 T3 = C ro fd fo fd ra ;3 ro ro 03 fd 1 • (d U fl > tn tn fd Q) rH CQ m 0) u ro Dh ^ rH ^ = -H s 1 — i tn H fl 05 >H 4J = fd ro fl Dh .^ rd .fl = ro ra W 0) S-) • CO m rfl rfl fT3 4J P 05 -P • > P P p f!3 CN 4J 4J ro C P cn' p 4-1 ra iH Dh = -P IH 44 fl M (^ ^ ,fl lJ-4 M-l 3 XI -i H fl -H fl 4-1 U > d (U P 4-1 p p rd U H J3 .fl p P T3 t:! p +4 S (D 1— 1 cn 00 +J 05 u fl > fl ,H > 05 CD fC^ (1) ro rd fD tn Cn -r-1 fl ro ro fd p Q -P tn > q 5 ^ fO 4J 4J is rfl 4-1 -H :; tn = > tn CJ Ti fC 0) tn Di fl fl 05 -H iH rH 44 CM X! CU rC ^ fl fl ^ rfl = rfl rfl rfl = +3 C U -P H H y O 4J -P 4J 4-1 44 >i P tri "Ti "Ci J4 u ro 44 P IH -H n3 O 4J 05 44 P 44 P ^ 44 M 4J P H m S Q^ u u Ah Ph m H iH H rH H rd ra rH ro ifl ,xf a o ^ 4-1 tn ,fl rfl ro -H fd fl •n Dh rH fd td 4J > > tn -H 03 -H > M m T3 "• 05 g tn nS m C! rfl CN] .fl 05 ra rfl g ns to -P = 4-1 ro 05 .-4 P

PAGE 169

153 Table III-5 CHANGES IN LEVELS OF DEBT AND TAX'ATION IN FLORIDA AS RELATED TO TRENDS OF CHANGE FOR ALL STATES Year Mean Predicted Actual Florida 1964 1965 1956 1967 1968' 1954 1965 1966 1967 1968 1964 1965 1966 1967 1958 1964 1965 1966 1967 1968 .ir-'er S129.17 140.75 154.48 170.05 1 ,40 Ca p.i.ta.. Debt $ 93.73 120.40 140.06 148,38 162.05 Tex: S^sverrue as a Per Cent of Perse 5.~78 5.51 5.92 6.09 6.15 6.17 6.06 5.76 6.25 5.91 Tax Revetuie P^^: Capita 127.88 112,89 137.14 133.18 154.26 147.89 163.90 146.46 180.62 163.22 General, Revenue Per Ca pita $110. 45 127. 00 133. 82 148. 98 • 133. 76 Income 5 94 5 .94 5 .83 5 69 5 .69 124 .21 131 .34 .137 .88 146 .23 157 .98 221.45 241.73 271.39 297.15 326.15 157.75 190.23 219.51 213.91 251.92 177.54, 188.12 203.60' 221.00 234.21 It should be recalled that the predicted value for each year is a value which would be expected if Florida moved in step with other states. The information this table provides deals with whether Florida has spurted ahead of other states (in percentage increase) or lagged or simply kept up. None of

PAGE 170

154 these variables show a spurt which could be related to Florida's reapportionment in 1967. In per capita, general revenue siid oer capita tax revenue there is a spurt in 1954 and a holding pattern thereafter. In tax revenue as a per cent of personal incorae not only is there no spurt but there is a general losing of ground during the period. In per capita debt there is simply a holding pattern throughout the period. Our analysis indicates that, for the period studied, apportionro.ent is not statistically related to taxation or indebtedness, to the measures of partisan control of the legislature or governorship, nor to minority strength in the legislature. Moreover, it is clear that no dramatic changes in these variables occurred in Florida at a time v;hich would permit attribution of causality to reapportionment. gduoation. The next policy area which we shall discuss is education. We might expect that the urban areas with large numbers of children would ta]<:e the opportunity offered by reapportionment to get the state to offer better services in this area. We examined data on six variables in this part of the analysis: expenditure per pupil in average daily attendance, average annual salary for instructional staff in elementary and secondary schools, per capita exoenditure, expenditure for elementary and secondary schools

PAGE 171

155 as a per cent of personal incoiTie, expenditure for higher education as a per cent of personal income, and education • expenditure as a per cent of total general expenditure. Table III-6 presents the correlation coefficients and slopes i relating pairs of years in the analysis. This table reveals more irregularity than did the one for taxes and debt. Four j 1 of the variables (per pupil expenditure, education expenditure as a per cent of personal income, average salary, and higher education expenditure as a per cent of personal income) demonstrates intercorreldtion of pairs of years and slope i measures V/hich suggest gradual increases. Tv/o of the i variables — per capita expenditure for education and education | expenditure as a per cent of total expenditure — show lower correlation coefficients for the pairing of 1965 with 1966 and 1956 with 1967. This suggests that there was some i shifting of relative position among the states during this 1 three year period. It should also be noted that the slope values are lower for these pairings suggesting a lower rate 1 of change. This implies that this was a time when education was losing ground relative to both need and resources. These expenditure variables are related to the political variables in Table III-7, At the very outset we can note that apportionment is not related to these variables in a significant or stable v/ay over these years. In fact, the

PAGE 172

W rH U CU fa fi3 G w u JJ H CD C 4J rG -P H -H tn -H (t! 'X3 -H 4-> ^ C fi ;X! (d C) CQ Oi M 3 s^ ;-! X-' 1) CU H ^ verage S Salary a P w < H < 03 ft fO r-H ch rH < 0} m a H rH C W S M kD H 13 tfi 1 U !3 iw C ^ M H H O fO O CU H fe H^ CD CO CU 1-H fe E-! M >i w < P U > UH (D O U -i-J fD M rH U ID -r-l 4J "^ „ ,Q Q 13 C TJ -5?. Q) ro M M a CD C g E-f P^ g ro O O ti a, a) o a (-:; O X .H oj tfl s en W M en (ti H en AND YEAR 0) (d u 4J 5 125 -H 4-1 O fo D-i -H H O TO Tj E-i O fi << 0) i3 CD X o o Pairs of Per Pupil Years Expenditure 156 u r'.O CTi t^ rH rH Oj O o o • e • • rH ^ rH H CO r-l 3 CD CO • • • CJv • n o rcn H O o O (Ti CO CTi CO in CO en in (Ti o o ID CO en CO o r00 CO CO rcri en en en 0^ co en cn cn rH cn CO r-IT) cn rH cn cn CO cn o o • CD in CD • • cn • cn en • -^ Ln ^£> rCO KD CO t 1 CD 1 CD ro '^ in CD rCO CD CD CD CD cn cn a> cn O iH r-^ iH rH r-^ j< O •r^ •*-! (d rH P CJ rH fC Q) tn CU rd • JJ ^ CD a 0^ -H .H T-i T? 0) c nrl m P rH 00 n CD p: 0^ •H H tn M 'CI) O 4J M-^ fd 4J tn CO !5 -H >i UJ 4J w IM H -H s 14-1 fd rH +J H (T3 rtl (-!

PAGE 173

157 00 i rH CO 52 o H < 'd p U -H cn 0) t^ m -u I i Eh H .p H O O fi, u O H W n3 S 4J H fd W Q m — cn o H o u o P j u c (U ;3 m O >i,i:: !B -P -P M 5-i r: CU O Q) D 21 cn m y 0) O >i ^ in -P +J •H Cn !-l M C 0) O 0) ^ C M -H -P tJ S CO O C W ra COT! M -H a) 0) -P m O^ ro O t3 Ci. W o C in M -H tri (D -P C 5 M ro ro rC hJ eu U O M -P c o U Ph o o ra o 5-! 05 >H -P -r-j t3 C Q) a w iH fd M m P O 9h ml 01 -p c u PM ro| G O -H -P fO u P o a w CO OD rH ,H O .H CO ro CO o! o r^ c> O O O CN O rH r-n O vo CTi r) 00 o KD rrH rH rH n o o rH r-i O o rH > 1 I 1 13 1 iH -H o M ip !H U > rH rH ^ cn •P (7^ • in 0^ CO in ; rH O o rH •H rH o 01 rH o • i 1 13 C a • 1 1 H rH -H P ;3 m A4 -P -H !H D) (Tl m rH ITl fo CO r^ oo VD cr. O PM iH o 1 O 1 M o o • u o 1 o O l' O rH U 13 .p -H Ti C D <1> '^ m rH C?i r~o CTi O m W r-1 iH • o o • rH • o o cn rH rH 03 rVD CTi O rH rH O I i ^ rH CO ^D OJ (Tl CM r VD vo vo ^iD C"-. O'^ O^ CTi iTi CTi rH rH rH .H rH rH ro in r^ rH O CN O O CM CTi O O rH 00 O CM r-i O rH O I I 1 I g () VD CTi ^ E^ CJ vo VO CO CO vD r~in in in "nT' c CM rH CO CM rH CM • • • H r r r r r r o M vDrHCOr^UHicoooojc^Jcriin LO LO ^ =* 0| fN rH "^^ CO CM CM 4J G O 5h ft I I I I I I VD m rH "^ r-i cn (D t-i rol CM cnl CM o CO CM vo CM O CM O C o H -P u I I I [^ CM in '^ f0 CM o in !H • • • • rH "sf rH vo fO '^ ,H O O CM rH CM • •••• I I I CO CTi CM vo CO "^ O ^ }-; •p •H d "^ rH CO o r^ O O O rH rH I I I "^ CD \0 in rHO<*inCMOO rHCOCMCO C^, C=* in VO rCO vo vo >vO 'XI vo vo CTi c?i cn vTi
PAGE 174

158 P c H o u I H H H (D rH Eh 1 H tJi P -P M fi <3) Q^ e a, < i^ ffi -p -^-i 'H tri i-i M C Q) (U 0. C !-i & -H ^J D S W -a M -< >i^ -P P -r-i tP fO ro rH in t^ en O O CV! iH H H r-oi m CO CM c-i nH O O I-! O O ) i I i rn LO CM CN O O O O O tn 1^ o ^ cTi '^ in f-o sf vo Lo LTi I I (i) O (D ^' G M O •r^ .p ^J Sh W ro •H ra > CD > -P rH in -H do Ui P O G •H tn -P C 5-1 ra fO .G P^ u fd r-l O CO cn -H M C -P -P fD 5-i C rC ra O O Oh U ro r-i tfi o •H M -p -p U G ra o ro u CO H Ui -P C u ra ro .ri W H 4-> ra o m m en o r-l o re P TO •^ (T) -^ O O O -H i I I I 4J ro n o ro 4-i| rH O O rH rH [^ "cf r-CN) CO 00 in >x> "* m •si' "^ I I kO vo cTi r-rH ro O ro i-D ^ CM o O CM o rH in U fd u CD a^ 'Cf (71 CM '^ IN VD ^ CM iH iH ^-^ O CN] rH 00 iH "q^ O rH rH O CM '^ O in U ro CD CTi CO t-H in t-"^ O rH O ro 00 CM ro s^ in >x) [^ 00 1X5 u? vD ',0 1,0 ix; en en en (71 CO ^£> ro a! ip r-H u o u CO !^ •H M Q) IJH tn a CD rH c h^ ro 0) >H M P d cu M U C G tn -P ro CD G a O u o rH ro -P 4H c -p rri G tfi (U -H (D-H o -P rG ft !H -P •H V^ ro rd (P (1. > CQ PM M a O ro H rH C M en 5-1 Q) ro tn -P > S c O G ro o .G C) -H u -P ro o P iH w cG ft H !h ro rH O r00 O rH O CM en ro CM U CD Cfl 0) ,G !h !H G -P P ro G -P IXJ '^ '* rH tn iJH M •H O O CM rH rl ^ • • • P > f:! M (D ro O U (U 00 =* in IT) r~CO U) VD V.D VD >X>
PAGE 175

159 o u r^ i H H H CD rH A .T3 Eh r^ (M O rH in •^^ 04 iH (U • • e o u H It! a o p u (D VD CO CM iH •n}^ o cyi cTi H CM CM n m a CM O CO CM CM CM • • 1* i' r 1* 1* 1 S! to QJ M 13 P H d C X rCM '^ CO vD O tH O O O M CO '^ JO 1.0 i^ CO CO •vt^ ID (D ra) m W3 yo v,0 >X) VO UD VD V£) 1X3 U) UJ '.O (11 CTi CTi CTi f?. G^ cr, CTi (Ti 0-^ O 0^ o^ !>^ rH H iH rH iH rH rH iH rH rH rH r-H 00 "^ in U3 r^ CO ^£> V.0 vX) VD "^ V.0 cyi CTi a> cTi cTi CT^ rH rH rH rH rH rH 00 ^ in kD r^ VD U3 >X> y5 VD Cn CT' 0> C?i CTi iH rH .H vH

PAGE 176

.60 '73 CU H -P C o o I H H H O a> [ m C 'H ti fO rC^ ^ T^ m uj Q) H M-r-i o C ^P 0) iH Cn f-i f^ M -a -H rH (0 Jh S (tf to Cm O -P •H to m rH rt! = ro O CO > • m rH •r-i m 03 V.0 rH (Ti o CM ro ro > rd 0) IM &1 U r^ r^ OJ rO r-l rH -H rd CO (D • P u >1 Congru! Contro! and ro ro ft -P rH • • +j P d p ro 13 •H -H !T> 03 > r-* ; [fi CO M 5H •H rH ra • P to W O rH d m to -H ^ Z rd rd •H ro td M-! tn +J >H g g 'w (P = ts 5H M 4J H 0) u >i m ^d ^ a rd ^ ch Q) +J O A •p P 15 ( ^ -ri u rH pi r-j +J M <0 >i >i C n3 (D ,Q rQ ro = d -n rH !-i rd rd rH • •H (U ro dJ +J -p ri 'd ft r^ n. Partii 1 of thi norstiip •H Jj o 5^ to P ro 0) n > 5 ro 'T3 'd tn pi d T3 rd rH rd P d 4J CO >H C to to d ro m CM fO iH H M c H (N iH CN (NO 0) ^ •H U H -H S C rd •rH -P U rd C'ha nge Cont; Gov* 1 P W m ri -H m ro o W ro tn rH rH 5 -P ft ft -d ^^1 -M CD J3 p ft ro > >1 r-\ -S-> rQ ro rd cQ rQ H 3 d M ^ = z -H S^ Oi rH fl ^ ^ rd rd +J to iT3 rn -P -P ft to •n Pi u ro L-j 1 ro -p >1 !h u d d u (D £ nj o O -H U rd ;3 ij o fd S rH CQ CO 4-1 to 113 P-i (D ro ro rd tn ijl -H Q > d • 5 S ro ra •p ^5 c Q) Q) ro IN tn iy d d to -H fi rd rd = d d :s to 4J A-^ u -H •rH o u 4-1 -p T3 d iH u ro fO ^-i 'xi P 4J to Q H a) ;2; rH U o CM ft en ro C ^ ro > U n3 4H tP rd 4J U ro -^ in IX) r^ cx) -r^ -d d fO V.0 i~D ^.O VD V.D >i3 m -^ r-i rH r-H H rH rH fd -P ro

PAGE 177

161 ll< (!) (D M = 1 P ::3 • 1 (U rA CD -H 3 'l •H -H -d -p m cn to CD J tn m fi Ti ^ ,c; 1 ra fO tn CD (D 4J 4J w CO to Cn CD M-l 13 ro ro to -H 5 ^ : fO to c m CD to tn J Q) Q) to nj J-l -H 1 +) 4J ro 13 CO CO 1 TO (ti ^ to -P P CO -P •P (U m ro 1 CO CO 1 tfi CD CO 4-1 4J TS fi CD M rti Sh c 4-> Q) rC O 4J 1 (T3 m a: P to 1 a^ to -P 'd m C TS O C C to o ^ Q CD 5-1 -H ,c: J f^ tn fO +J +j = tji (D fl & Vt H rC fd CD rQ ,H H 4-1 ^ (D &i = f^ H V iy U-i 4J m C CD 0) O 13 !D tjl P tn iH •H ,£:; c >i H a A CO -rj u ra -P ro -H fU ra ^ H > to B > 0. -P u M -H o Oi u rC a 13 T-l A P-J CI -H 03 4J G M 13 13 g J-i rH 13 >i Q C 'Ci Cl, (D O 4-* > i-i H id 54 Q D. ^ 13 tn nj tn > -^^ to s -H a rC JH fQ to 4J &i to LT C M a M ra A (I) ^-1 4-! ^ C G CO P u > M It! >i 3 Q) CD ^ .Q u 4-1 > Cn > a -H 13 H tn CD tn 4-1 CD t3 .c CO ^ C Q) 4-i ID 4J .H rc; (U ^ A to +J 4J M-l 4J u = -H 4-) U-! H CM <+-) IW .a G H H 13 H 4-1 ^ •H C •n (D 1 a a +J iH iH Z • V\ H (C ro rH CN CO > > Oa 3:

PAGE 178

162 o'nly relationships of any notice are between the strangth of the rLiino?rity in both houses of the legislature and expenditure per pupil in average daily attendance. These coefficients remain moderate and stable over the period. Change is examined in Table III-S and III-9. The partial coefficients presented in Table TII-S indicate that our' political variables, including reapportionment, do not statistically covary with change in these expenditure variables.. Change for Florida, as presented in Table III-9, v/as examined by comparing actual and predicted values for Florida on these education variables. The thing most notable about this table is the spurt above the predicted vali-ies shown in 1968 for salaries and expenditvire per pupil in average daily attendance. While it is tempting to attribute this solely to the new apportioned legislature, we must instead call attention to the fact that there was intensive lobbying by teachers' groups for more money for education during this period, and there was a teacher walkout in 1968, Nevertheless, it should be noted that the teacher walkout came after, not before,^ the appropriations had been passed. We can, however, safely go no farther than to say that reapportionment probably had some impact on this outcome. Although our analysis is indeterminate on this point, it is suggestive that a notable national lag in the

PAGE 179

153 o H 13 H G H D E-t P H < a-CO fc OJ H > CO u H CO U Eh H H nq O 0) CM t-H E-i H m CO o H Eh < O o < H E-i CJ << A4 CO O H > -P K in Q) Pi I O >i Pm +J O O 'Z M-l H o Eh a; 4-) o Id CO CO f=3 t c o -r-! tf. A-> -P O i rC (0 P -P -p H tn O M fi Eh cu C M <4-l -H 4-) s; CO +J (U o d u rut3 (y 01 O ft H Bi -P C ra M fC ra rS^ CO 04 U ra ro u rn CO H bi P G, ra rfi ft u O -H -P (ti u H c^ ra rH (D tn fi -P 4-* rd M d ^ r5 O U ft o s ra en -r^ u -p 4-> M C ro o ft o GJ >H O Cn CVn I—; O rH sT ui CO CO O cH O C • 4 • t i I rH iH O 0-1 rH CM O CO X) yD U3 iJD ijl +j "^ o in -H H -^ ir> LTi 4-1 CM rH o 4J O o H O < r 1 ra o ;:! !^ 44 1 1 > -H ro P ft 1 4-1 -H a X w ro l' 1 H iH a ft a^ •Nf ai 4-1 CO ro CO CO o rH o -H O O CN o i>4 • • 1 & it) U (D ft 1 • 1 ft VI B <-^ rH o rH 00 CXi en a o C\l CO O O CM o X • • • • • • w CO f\' CO W O CN r-4 m rin h o o r-l o o 00 ro CM 1 CO ^ rCO CSJ • 1 rH .0 rsonal IncoiTi r-4 CM 1 rH • 1 • o ft 1 CO O <-i CD CM rH rH CTi o o O rH ft 00 s? r-\ o rH r m 1 r 1 -p c w u M

j • • • ro| • CO rd 1 • • C -H 4J ra o 00 00 o •^ iH rj CO CTi o rH o u ro CM rH o o • 1 1 W O >4H (D 4-1 • 1 > 1 • 1 '^ ^ (Ti -H o rH "^ •^ ro CM O n3 ro CM o ^ rH 1 • 1 (D a X 1 • 1 • • o ro O *£i in 00 in rH rH rH O • • rH • o t • o^ en r-l r-i ^ m uD r-CO V£) (.0 V-O VD vo (Ti CTi CTi CTi CTi rH rH rH rH rH I I I <;]' m ID rCO ID '-D UD IX) VD cn cTi cr> cTi cri iH I — I I — 1 I — I I — I '^iin ? r-~ 00 3 u) vo uD :> cr^ (ji en en a^ rH rH rH rH iH

PAGE 180

164 P a -H ^-) o o CO I H H H o; rH fD Eh i d o -H ij> O Q) ft (0 m o >i ^ K -P -l-J •H tri !H M C O O Q) P S W m (D iH rQ ro -H 5-1 > > -rH +J ro 1 — i m o o -H 5-1 O c -H 5-1 -P cn CD en o !.( 0) ft ft !=> M CD -H tn 4-1 C 5-< ro fO rC CD D^ CO (1)' 1-5 g O O H ra o Cfi CD 0^ ml o! p r; CD U M X) ro Or* O r-1 4^1 r-1 V£) CN >£l V£) V5 rH rH iH O O • • • B I I I CO r(.0 O O rH I > u m rH m H ro O (Ti G ,-H O C CTi H CTi r^4 r0-3 rH O O O i I CD & ro u CD > rq LO ;J< m ID t"^ CO 'XI ^ v.0 UD VD O Ci CTi o"^ Ci rH rH H rH rH D. r; -H 10 rC Gl M H }HJJ 5.-I c CO VH a. 0) r> tw o a CD m '; C (D rH :3 5h 5-! irr^ -P G G O o o CD 5h P P ro rH CO H tjl !H O ro O u P -p •H d X ro d o ro -P MHl Oi -P d u 5h di ml CO' ro d O -H p ra u M 5h O UH u P -H ri d a C^I VT) •* ro O rH in o CO o in r-j rO CN rH 00 CN O 's)' O .H rH O 5-1 -din 1-0 rCO '^ in ID r-*co m ^ ^0 VD v£) i£) ^ UD <,0 VD o CD CTi CTi 01 CTi CTi cn 01 01 01 0^ >H rH rH rH <-\ rH rH r-H iH r— 1 H

PAGE 181

•d G.1 -U C o u CO i c fD ^ n ai m a! "l—i M S^ +J 3 !^ a -P ra n fO fe Q) iH > m ^-^ n "n o tr. (U m 1^ s t) o 1-1 C ^ ro ;-j M tr. iJ G C O O -rH ^C &( +J 4-) -H ro M-i 03 fii O M O •H O f-4 M 0) 4J r; b fO O a u (D +J M r-l -H i ™ p u > (D -U > i-ji a C\ !^ O 'Hi O -P o u ro p. -H -C w 0) M ,d o -i-j d K) 4-1 0) -H O > -l-i o Id ft u •H Oi p' ft u ro +) O C ra +j nj Q u =* in V.0 r-CO sjin 1,0 l^ CO m VD vo 1.0 U3 VD \D wj 1.0 H >H r-1 rH iH rH rH rH tH rH rH ^ m ^i) r~ CO vo VD VD VD 3 O^ CTi CTi C5> CTi r-l iH rH rM rH s!' m CD r-00 KD \D '^ ^ \0 (Ti (Ti CTi 0^ 0^1

PAGE 182

166 o t! JH = tj C P 3 • fd rH g 4-J CN :: > rH CO rH ^^ •H H fd CO CD = -d CN > H -rH CN £fi z ^ tr. rH ^4 ro (!) c 4J 4-> (D ro 13 tn rG r-l > > 03 CO iH •H ^ T3 4J MH (ti (C fd CO rH w 15 ^ • > CO fd G ^ rG ^ •H >1 >1 fD > tn 4-1 4-1 ,c 4J 4J -H C iH 4-> y cyi U !h rd CO CO tn MH U H c fC3 fO 4-> fD ^ CD -H ^ ^ fd rn Oj a 5 4-> (D CO G (D > .c; 13 CO CD tn j-i >i >( 13 CD fD M -H (D • 4 4-1 4-1 c +J fd CO CD -P ^ ^ •H H tn (D G :S CO 4-1 CD fD +J c CO [0 M 5-1 -H 4-) tn fD fd fD iH ra • p 13 CO CO •H ^ rH fi 03 ^ iH C •l~i CO CD 4-1 CO CO -H >< :: ,a td •H fd fD CO fd ri ^ fD Cn 4J rH g g rG fD 4J 4-1 tn 4J 15 -H z ^ M M 4-) CO fD .-M M M CO 4-> •-{ ,Q o m 'd ^ cu ^ ^ >i fD CO 4-1 (U -n Cli 4-) 4-) 5 4-i ^ rG ^ fO ^ u rH p U 4-1 .G 4J ^Ci 4-) g >i >i C fO rG 4-1 CO "fi rH rH •H CD fD ^ 4-1 +j m r-l 4-1 = +J ,G G CD a 4J a) fD T! +J P ^ ^G c • fD fd ^ ^ (U rG (N U rG G rG CO 4J Q Q T5 cn -a 'd g = U 4-1 fD :=s G w m a CO CD c CO G m 5-1 00 G C fO 4-1 -P :3 3 g G •H fd rG H 1 (D ro tn ^c; H •H -H 3 fD 4J CD > H ^ u H u g g ,£: rC 4-1 rH Q M -H rG H +) •H CO -H -H M fD -H fd -P 4-1 = DT H i-H CO rH iH M J4 > rH Cu u IW O m CO &, ^ -S fd rQ H ( fd rC ^ ^ rH ^ 4-1 <4H ^ C (13 4J ai ^ 'TS Di fd tn P Di Eh -H : r 13 ,G G >t rH C J-l rH s ^ ^ rG r^ 4-) CO fd U fD 4-1 fD -H r- 4-i 4-) &< G fC • rG •H > CO n 4J M tn 5 CO 4J M c ra -p >1 u M G c U -H fD a, g fi-l c .H •H -H u CO M S G ^ •n ^ a, H 5:; 4-t m fD CD -H fd 4-1 (D u -Jl CD fd G !H Ti ^3 g CO CO +J P 4-> !h H u >f 'xi c fd fd ra fO U G HD Cl^ fd 4J [~^ fd J-l <-< 1-^ U > !h -H G M o i-i CQ CO CO CU,G 13 tn Id 1^ -H rC g g 5 tn > H CD G -H ft 4J 5 fd (d rG 5H fd CD U3 J-) CO CO rG rC fD c 4-1 tn CD CO 4-1 fd CO 4-) 4J fD C U G M fd rG (D rH Oi 4-> rH u 4-1 5-1 to CO rC ,c; MH m 13 Dj rQ ^ G G CD U > d fd 4J 4-1 g 4-1 M > !H (d >i -H U rd fd 5 15 ,Q g CO CO 4-J u UH > tn > ^ fd fd tn tn •H ^ c -H 13 Q 4J tn :: 5 ;<: fU fd <: u -H tn tr -P c 4-) 4-> J3 T3 ,G fD <-i 13 fO CM ij tn G fi CO CO G 4-1 4-1 rH ^ rc: ^^ C rG i-d ,G .G CD +J c u -H •H 4-1 ^ 4-1 4J MH 4J U Cn Ti T3 u M fD 4J = •H 4J IM -H :3 4-1 •H 1+4 .H <4H ip rG G H 0] !z; rH U U ^ flj CO 5 H H 15 H 4-1 ro (0 03 (U 3: ^ fO > '5 'TS 4H tn CO ,G rH (d > H C fd r~i party were c

PAGE 183

167 Table III--9 CHANGES IN LEVELS OF EDUCATION EXPEi^TDITURE IN FLORIDA AS RELATED TO TRENDS OF CHANGE FOR ALL STATES Year Maan Predict ed Actual Florida Expenditure Per P up. .1 in Ave rage Daily Attendance 1965 $460 $. $. 1956 505 460 439 1967 553 482 479 196S 602 528 554 Average Ai-^nua 583 5 1 Sa lary For Inst: ructional Staff 1964 5873 6227 1965 6030 6423 6270 1966 6351 6584 6577 1967 5725 6950 6700 1968 7180 7155 7500 Exr endi' ;are for Educa tion as a P er c ent of Personal Income 1964 • > • • • 1965 4.97 • • • • • • 1965 5.07 5.00 5.15 1967 5.23 5.30 5.32 1963 5.41 5.49 5.53 Expend i tv:re .|2. r Higher 1.38 Eg .ucation as 0.69 a. £e r Cent of Personal Inc. 1964 0.74 1965 1.43 0.88 0.89 1966 1.51 1.02 0.94 1967 1.81 1.09 1.00 1968 1.94 1.13 1.12 Education Expenditure As a. Per Cent of Total Expenditure. 1966 1967 1968 1954 1965 1966 1967 1968 33.63 38.36 33.70 31.38 36.33 38.50 34,. 59 41.22 40.70 36.39 42.14 41.20 3 7.14 41.55 45.00 Per Capita Expenditure for Education 76.29 62,69 66.57 82.47 73.34 75.91 104.76 99.12 82.78 118.72 101.72 93.72 132.99 109.30 109.62

PAGE 184

168 years 1965-67 is caught up after reapportionment in P.lorida Welfare Just as problems of the young and education are associated with the cities, so are problems of economic insecurity and welfare. The demand and need for welfare services seems to be mostly closely associated with urbanized areas, and we wanted to see if with greater urban representation legislatures would move to improve these services. The variables examined in this part of the analysis were (1) average weekly benefit to total unemployment (2) average monthly payment for Old Age Assistance, (3) average monthly payment for Aid to Families with Dependent Children, (4) average monthly payment for Aid to the Blind, (5) average monthly payment for aid to the totally and permanently disabled, (6) per capita welfare expenditure, and (7) expenditure for public welfare as a per cent of total general expenditure. The trend patterns for these variables are set forth in Table III-IO. For unemployment payments, per capita welfare expenditures, and welfare expenditure as a per cent of total expenditure the pairs of years correlate at a uniformly high level throughout the period. In the case of Old Age Assistance and Aid to Families with Dependent Children 196566 between pairs of years is lowest in the 1955-66 and 195667 pairings. For Aid to the Blind the correlation drops into

PAGE 185

169 V) O Pi O CO < 13 W U O fH H f=q H H Pi O CD |ii i-l ro E^ H O H w c o M A^ O en Q O H EH P^ p3 O u ro -P o Eh 0) O •H >1 rH -P CD c g 5-1 P^ ^3 C M o '^' rCN H O CT. CO O O rH rH rH M? '^ ^ CTi on 0"^ C7i CO CQ 0^ •P T! O -H 4-1 <->. en (D H rH H ro o -p •H •P H in c?i '^ o CTi cr\ o rUD ^ rH VD (Ti CTi (Ti CO (T" "d O rH •H -P U •H ^^ CO >Xl rH O O CO H :-{ r^ CO CTi IX) 00 CO J^ 0^ O^ CO CO (Ti (1) u CU G tn ro i=d -p H O w u rH ID rPCM o cr. 03 cj^ o CO "* vo in n CTi CTi 00 00 CTi 4H o ro !h fD -H tU ro >H P4 •<* m vi) rCO UD VO VD IX) U3 i I I I I 00 "^ in 1.0 rVD VD 'X) 1X3 VD <7i 0^ CTi (Ti (Ti rH rH rH rH rH fi) (!) u ^ U-i !-f P -P 'H -H -n -p 'T1 fD C a C u 0) Q) C C) 04 u X >;; ttJ H !H w 0) CL. H ^ ro ra IT! M M m 0) rH cn c (Ti (Ti (7) CTi CTi r-rCO i^ rCTi CTi 0^ C5^ CTi U ro ro u -P CO C -H U) M cN CM 00 in o O O O rH H rH r-^ r-i <-\ r-i cTi CO CO ko in CTi 1 g -rl O rH C 0) g m c !3 ,M is O CO CO o o cr^ cTi • CO vD rH r^ (Ji (Ti CTi (Ti 4H O CO M m ro U -H >H ro CU r-in 00 rH vX) 03 O O O CTi H rH rH in CTi ^' CO rH 0^ CTi en CTi CTi ;t in VD !^ CO VD IX) UD ^X) U) 11 ro -s? in IX) vD iX) IX) vX> O^ CTi o~i I I 0^ en en o H -p m ^ o rH o en rd -p d H rH u d H en -p rS -p tn >1 -p MH •H rH <:

PAGE 186

170 the 80 's when v,'g reach the 1957-68 pairing. The slopes for all of these variables, altho"agh showing some variation, generally demonstrate a gradual pattern of increase throughout the period. This table suggests, therefore, that although expenditures for welfare rose gradually throughout the period there was some slight shifting of the position of states in the period 1955-68. In Table III-ll we present the zero-order correlations between the political variables and these welfare expenditure variables year~by-year Even a cursory examination of this table indicates that there is little here to discuss. The political variables show low and/or variable and unstable relationships to these welfare variables. Certainly no ancourageraent is offered about the explanatory value of reapportion-fi-ient. Substantially the same thing is true for the phange correlations presented in Table III-12. We pass on then to loolc more directly at the pattern of change in Florida as it took place over these years. In general the data presented in Table III-13 can only be described as depressing. In most cases there is a pattern of expenditures shown far below the national mean and frequently, if not consistently, slower in change than the national trend. The one exception to this is the case of ?iid to Families with Dependent Children. As the table shov/s

PAGE 187

171 rH O C^ CO VD O T-i CD C> r-{ '..0 CO cr. roT cs O O O O r-) O K ro O CO 0) 'M -H Cn CD +J fi '^ M ra ro ,.G tJ P-i O G -H C^ ro H Q) tn fT< -H u c P +J !T1 "^ rt -C ro u i^ U C ro ro H CO H }-) +J -P M C ro CL, o u ro QJ tH r i' -p 1 II II 0-) cr^ r-in (Ti r^ cN o iH ro M •q'(N '^ in 'd' CTi O O O O '^^ • O • CN >^ H rH Cn O on CN CM CM CO O "^ U1 in OT o o o o CM in fH nj' rCM n n •sfro m ^ si' -^ VD in CTi LO CO o iH 1-1 I— I CM iH 0^1 I I "sf r^ r^ r^ r^ ^ "m* ro rsi 00 "sf o o vi' rUD "sf O O O O O iH en -^ v^ ^0 ^ CO ijD v£) 1.0 '--O IX) UD cyi CTi CTi CTi CTi cr^ _) _J r-J ,— I 1-H t— ) 00 "* in IX) r^ CO VD I.D UD >i3 W^ VO 0"! CTi Cn <7i CTi O^ iH iH iH iH iH iH 00 ^ in v^ r-~ CO 1.0 1.0 VD VD V.0 IX) CTi CTi C?> CTi 0"i C3> rH r-l rH rH tH iH I I ro "^J* >n U3 i> 00 X> V.0 X) ^ y3 'X) CT'
PAGE 188

172 =X3 H o i H H H CO rH ra rcii ai rH ,Q -H > a) ;> -H -!-J ro r-l 'a o -H tr. O 05 cu g (D o >i rc; K -P -ti •H tn M 5^ C (U o o paw a) GO o w M (D Oi P •4J -P O 'D C U -H -P S CO W (]) -H tn -P C m rd ro ^. ft U CI) U} b CD O Cf] (!) •H tn M ro P.4 O -H. d^ ro t-H 0) ci] tn -H M jz; -u +J ra M fi rd fD o u ft u 0"; VO n ro r-t rH O O CM O n O i-^ ro ^ (N i> .-I O O r-l rH ,H cTi c-1 rn c^ r~O O O O iH O O CN ro Cn] (Ti f^ O O rH rH O CN I t 1 I I I I i I O] O CN liO rH O iH 00 ro ro CN -vf rH (J^ ro CM CN! O rH CO roj 0^ cvj ^ rH rH rH O O O l' III 1 -P H (D >^ cy. ro U5 CO rH 0^ CM r tn CTi CO rH ra CD a (J) a O CO O U •P +J •H 0) M ro ro CM rH CX3 O O I I I I ro -P o Eh Uh O P c O ro CM CM O H 04 H CO o III! o I 0) is ro -P -H ro U CN U) I I O m o •p -P U G ro o ft O u CD OCN CN 00 ro O in .0 .H O O O CM iH '4-)) • IS I of CO CN r^ ro CO CO O O O CN ro "xf CO in -^ in rH O OJ CN CM tH CD ft CO ro| (D !h 13 4-> •H O CO H CN i' r rH i-H I I I I CD w o rH rH cn o ro CO '^l' o o I I I I O CN VD CM m o rH O rH O CO r'i • 00 CTi CTi ID M3 (N O O O O u ro ^ -^ vo rro' CN mh O O CM CM rH O cH ^ o in (T> CM CN m ,H O CM "sr rH rH 1 1 1 1 1 1 1 1 1 U5 <^ rH r-l rH 0^ rCM r-! CN • in U3 !^ CO rH rH O O O O ro vD M rH r^ O r-0 CM rH (M rH rH ro ^^ m (JO rCO \.D vri kD ID IX) yD O^ CTi CTi CTi r~CO U3 VD VjD 1.0 U) UD 0> 0^ O'^ 0^ CTi 0> rH r-4 rH rH rH rH I I I I I ro xj' in vD r~CO U5 1X> (D ^O 'XI VD CJ^ CTi CTi cn cn CTi I — I — i 1 — I I — I : — I rH rH CTi VD m ^ ^ oj rH O ro ro CN I I I I I I ro 'xj" m iX) r~CO VD ID >X) y? VD (D CTi CJ^ C?i CTi CTi CTi I — I iH I — i r-l I — I rH

PAGE 189

1/3 13 O P C -iH -P o rH iH I H Eh M O -PI rol C5 '.0 .d O ro 0) > o 'r-l fC Pa m O (1) u m tn C o (D CD i-H P' i-H M O C -P c o o o ra •H (D-H P ,c; ft M -P -W o m u O c O M fij M 0) ijl -p > G C O iXi O O u rH M •p u c fD w -.—I 03 fO -Nf O O O rH ro r^ vD CN CN O CTi >* in rH CN O O O O rH ID o cTi ^" r^ cTi O iH O rH L-NJ O cTi en o rH ^f in O O O Q iH O a. -H en 5-i O fO >4 H /5 (I! to 'H P fO +J O E-i > -P C^ 0) c; M w c K (U •H 0) u to SH m ro u < Mh rH ft rH 0) tn rH VD • -^t in 15 00 (XI CO rCN TO 1 CM CO in • en cTi u O i>! O CN rH o O c O O O o o o • • • • • ro • t • • • U) • • • f> • c I 1 1 -P 1 1 ro 1 1 1 1 1 (1) •H C5 ft ro U M ft > 5h +J -H 13 a X M rti IW rCO CO "* CN in CSI CM rH CN O rH CO rH CN rH O O O O O rH r^ 00 ro O rH CN IS o en s^ "sf en in f^ rH O O O rH O CN rH O 00 ^ in uD r-^ CO U3 UD iX> UD U3 £i en en iTi en en en I — I I — Is — t I — I I — I I — I ro ;}< in uD r-CO UD UD >£> iX) (£) U) cyi cTi cTi en cTi en I — I I — 1 rH rH I — I I — 00 nI' in lO r^ CO VD VO U3 (JD VD IJD en en c^^ en en o~i i*~i r^ T~~\ rH r~i i~~i ro "q^ in v£) r-00 (^ >X) ^D 1.0 v£) v,D c~i en en cTi en cyi rH rH rH H rH rH

PAGE 190

174 -r-i +J c O U 1 H H H EH fD 10 -r-J -P fd CM o u 5-! d o u C G O rH O 0-} -H "4-1 tnS O 0) Hi rH O T! 5-1 C -P G O U fa •J) rH rQ •rH 5-J > -H O 4J ra ra -H •)-) 1) P4 H 0) in u ,c ex 4J -H u-l CO O 5-1 rH e, O !H 5-{ 0) -P > c o -0 o o ^ 'H u* -P H G rC d) in c; rd u P r. c fo IW !-l m G) •H > +J O M X!) ro CU CM fNI rH •.0 riH r-i ro cri (7. 0") ro rH ro rH o rH O o CM O CN rH o !H rr^.0 ^ '^ in ro CN o m LO ci ro ID o rH [-M CN CN ro r^ P H m c o m -p > O a) c iH rH CD O >'S 0) tr (U 0) > CO O CM (N o o 01 ro rH O O QJ r^ O -P c o a a pq H cn in frH O C (^ O "^ ro O O H CM rH O -P H m -H iH -H ro O -P -H I o p' H sO a) CO iH CTi CTi fvj r-1 CM CM iH rH CM n CN CM CO m <%)' ^ ro CM O c^J CNCTitHCM-^in vD(i)roinf-cM rOCN'^COCMCM rHtHCMCMi— ICM 5-1 ro 0) ro ^ in o r^ CO \S) '.O '•D ^ ^ ^ CTi Oi C^^ C?i CTi C3^ 1 — I ; — I — I r — 1 I — i t — I ro '^ in i>D r~CO VO 'vO VD (£! VO VX) 01 C7> cyi C' CTi CTi I — i 1 — I I — 1 ; — i t — t I ro -ci^ tn >x> r-CO 1J3 UD yD i-D '-0 U? C:^ 01 01 01 O^ CTi rH rH rH rH rH rH ro "^ LO vD r^ ^ VD "X) VD VD ^i) '^ 01 CJI CTi C51 CTv rH I— I [-4 "—1 t"! 01

PAGE 191

% 175 tJ rH (D C (tJ P H GT' Z > -H ^ "H rH fD ro to ~ -d 3 > -rH CN m rH A M ra ffi a fD •P (C 3 !T> > ,G > CQ rH -H CO '•d -i-J CQ :? 3: > CO ^ G T3 -H >i >i 'D -P tn Cl o (1) p -p -H G +> 13 Cn Sh !H rC CO fd CO tn iH C ro (0 -P fD CO -rl ^ m rti C^ 0^ 5 d fD CO OJ > ^ -d tr. CO u CJ >t >i -rl tn fD 3 CD o • jj 4-> fi -P CO 03 -p d (D CO 5 CO ro -P a = W (Q H u -H -P fD fD iH fD • P p CO CO "^ U] rH c n CO CO -P H >i rd rG •H ID fD fD fD tn +J rH g g rG ^ P -P (D -H = tj M %A •p CQ fD tH U a) CD u •P iw >d ^ Oj ^ rC >1 -p CO 0) •1— 1 6 a. 4-> -P '? 4-> fO rG ^ ro rH p U -P -P TS 4-1 g >( >1 C fD CO A (0 P Qj rQ rQ ro z Qa CO .p p C I'D rH 5H ^ ,G r-\ • d c •H ra (D 4J -p d 'd a, CN u rx: G CO -H 'd > 5 rH rH ~ -rl +J fD "d •P U> tH -p -P A G c 4J Q) ra 'd -!-> -p ^ A c fD C ^ 2 ro ^ (D 13 P fD G ,G o Q) ^d tji 'd Ti g rH d o fD m m 3) fi CD Cfi CO c CD -H CQ G rH C c; ro JJ -JJ p p > g r-l H fD rH (U nj tn rC H -H H r-Q -P CQ 1 rC O -H a g g A A -p o s M -H H -P r-i m -H -H 5-4 rC ft fD -P H r-\ m (D rH rH u u •P & 'H H M-i ,Q fD y) a. rC Di S .d P -P -P <: 0^ a, 'd -P CM (D a VI -( P a fD fD tn r-f >i OJ It) rC .5 ,ri iH 3 a m G rQ jj v^ 3: TJ tn CO fD t5^ fO -H. (U -.H fD ,G C E^ !-t •p \A fD :s M -P U • fO -P >( V-! p c C g '•H C rH -H H rj !H G c ^d H c. 14H m ro U O 5H -H ro CO CO G d T! Kl a) -P I-" P 15 M > •rl ^ — t: tn -H G3 rH Wi (D CQ CO fD ft ,G rC -) ^ g g u <-i tn -H CQ -;-' is ra fo o, .G rG M' W cu 5-i u; cn ^ rC (D -P -P CO -P 3 fO to P -P fD G u c ro iH a^ 0) -p '4H O U u ffl m ,c! rc; MH Hh p rC O -rl G o > C ro +J -p u g M > -H u rC p P rg g C) m CQ -p CO U G > tn (!) .-C OJ fC 03 tn tn -H G fD Q +J b-> = > 5 (D fD .5 rC •rH tn .G • -P +J +J :: rG CP "d (D CM iji tn C G CO -H rH -p ,c; 2 rd G a & ,G = ,G +j >-< o •rH -rH u u •p -P -p MH t^ (U d d M M fC -rl m -iH P -p CO m p iw H W S rH o u a, Ph C/i H rH H -d ro 01 (13 5 > o 'd 4H tn -d h the va. "2." •H G fD rH CM M ; 3 • g -P (N fD fD z CO <-{ CO -H p rG tn rH -p ro rH > m A ^ -p -P u UH 'd S G CQ tn U -rl P CO CO •P P CQ ro fO rH rG CQ CQ -H ^ fD tn -P > H rQ +> U fd A 4-' -p tn m G o o ,G -P CO ,G !-l CO -P p G m U r^ H > rG P = tn • ,Q rH rC 4H -P p tn >lrH G -p m -H rl > CO M ft -n,G ft fd -P g ^ >> ID -p G M 'd tn fd G -H ft fD CO CQ U fD rG -P G CO M fD >1 15 rQ > Ti tn -P CD rH .P H fi CO +J u -p MH ^ G H -P n U > -P iH M to ft ^

PAGE 192

176 D !x, w o H CO K Pi P H t5 Csl H M w u H E-i H O (1, fa -P GO D 4-) O xr.\ > -rij Ci I i a u\ O o! t.. '4-! I ??; o H u^ o o a) 0} :3 O >l rd ffi -i-J -M •H tn M !-l G (1) O Q) O, C 5-1 Oi -H ^ P s; W i ,C K 4J 4J U U ti CD O O "^ C M -r-l 4-1 i^ S OT CD m u P O H t-1 O Eh b -"'1 i-l o p^ O P H < C ra 4-i fi S4 (13 ro CO CN rH CM (N O r-\ CO KD r-l O m oj o "^ o r~en o LD in rH O 'vT' O iH 4-l| •H Q) 1-n o C CD CQ +J C CD g > O r-i a E o Q) O P >i o V£i .-I r~CM CM o o o i> en rH <^ in tH O CN o o M H u 4-J I I I in vo (H en o o o -s? tH "si* 00 CO rH 00 O O CO o CO U5 O O -sf rH rH O u ro O in o i-NJ CN H C m 4J CN ro o iH in o o 113 o U3 CN CO iH O CN I I I I Ui G) r-t -H oT (N rcTi ro! H r-i cTi CO 00 iH rH r-H iH O 4->| O iH CM ro O I I -P -H i I or-'* CN "^ iH V0 O O r4 •H O r^4 H UD O 1-4 H O M U3 CN O rH H CT^ iH H rH O 01 iH rH rH 00 O 1-4 CM CO rH CM rrcr. o 00 "* r-4 O H r4 LI") CN E^ 00 OM -^ rH O CNl u -^ in ix) r00 CD VO >.D i.D *^D VD Cj \ CTi en C3> CTi CTi l>l <-H rH iH r4 tH rvj o i-H r^ rH O O "* O 00 4 a • e I I I 'cj* Lo 0^ en rH [-4 rH r~\ !-\ on in oo 00 o rH O rH 00 rH ro r-t^ oo rH O O 01 rH O <=t in i-n r00 U> 5 U) ^.0 1X5 O". en 0^ en en i — ;'H r~^ t — 1 ( — i I I vi' in vT! r-OC' VD VO 'X) VO 'sJD en. CI on cTi o^ I — 1 I — 1 ( — i — i ( —

PAGE 193

177 'Xj CD H -P r; O U ri ,C K 4J -P o d D., r; r:3 a 0) CO 'V re > -r-i -!-' ru I — j CO •H OJ o o >. ^ tC -P -P M 'w C (P O ^S c u O •>-! +J 1-3 is; a G c CD CP -H ijl -p f^ ^^ ra fo u ro U •P C M ro m ro ^ (0 O O Fh U .1.) +J u in m o -H M ro (1, u ro rH 0^1 C7\ t^ O O CM i-i m a^ on CO rH iH O i-H O 03 VD '^ O O rH rH ,-H i,o OJ in -) CN m CM CM 1 1 -! .p -H 0) D X W s r-ro CTi CM O O CM o C\l "m" C-rH O rH CD rH w CM c. '^ "sT CTi ro P r-1 CN rH CM rH 4J -P r-\ 1 ^1 ro -P o E^ roi r-H -P sn CD C ro CD s:, -H U1 CO rH O CD C.) ro -P V) H ro M CD C CD O ^3 VD ^ vD O O^ r^ CM CM C^I rH o r-m ro c?i m rH O CM O o i m CDl 5h cTi CM CM roico O O rH ^H O I (D 15 ro p •iH a U u CD m CD u '^ o r^ rH (N O O I I CO xf M li") rH in CO o o CD O O o o 1 m fCi • 1 • • 1 CD 5H ri +J vo sl^ •H ^ CM ro cr. ^ •-i rH T! o CM rH o CM 1 CD a X H • 1 H 1 3 ^m r-i rH rin O CM rH ^•f CTi 00 LO O O O rH rro in CN rH O O rH O rH CD U ro tn CM r^ rH CM O CM O O rH i 1 1 1 1 li> 1 t 1-H ci OJ r~c?i O O rH O O CO O ^ r-1 rH rH CM O O CO ro rH 00 OO rH rH O O CO 1^ CM Ln ^T) O rH O O O ^ LO uD r00 V.0 'vD U3 IX) V.0 CTi CTi CTi CTi CTi I — I I — F — t — I I — i in vD [^ CO KD '..O VO iJD a^ cTi o> a> r-[ r-H r-4 ,H Nf" ID ^3 r^ CO '^ VO 1.0 >.0 VjD CTi CTi 0> CTi CTi rH rH r-4 rH rH ^ in u) r-CO vi? 'vD W) vo y3 rj-i en iTi ^J\ (J^ r— I rH I — '. rH :: — i

PAGE 194

(D P C -H +J C o rH I H H H <\> rH Eh -H i-J !h j P^ m D M O o -H w i-ITI C tt) IH 3 (0 -P > ro O rH O tn -H O Q) rH O TJ !H c: 4J o o ro a K -fH -P ro c H o M -P c o o ai-H JJ -H O M O o u U (D P > C O o o u •H Q) to ..a M -P o c u MH a) o > o in CO 1,0 in rO tH O rH rH (Tl CN CO O rH O rH O I I i I I i u "* in ^D r-03 '^f' in •vS rCD nj 5 VD 3 KD 3 IJD (JD Xi U5 CTi (T> CTi a> (Ti 0-i CTi o^ c^ CTi >H rH rH rH rH rH rH s-^i rH r-l rH ^ in vri f^ 00 ijD u3 (.D m:> CTi C3^ (Ti 0^ (Ti iH rH rH r — i rH ^ in ^£1 r^ CD <£) ^D UD i£) 'JD (J> C^G^ O'l CTi I — I rH rH rH i-H

PAGE 195

179 3 -P ft; o u CN i-H 1 H H HH 0) iH ,Q TO •H > CO -H >-) o 05 > M-l o O O to U O fl) C hI CD r-\ O Xi M M G CTi -P ra C C O u u a) (D-H H ^ ft -P +J -H }_i ^ pL, o M o d ^ d H O U !-! CD iji d COO fD V rH O M O U C -P m m H o -P M OJ CO O) rH iH O .H r-H iH ro ro CN c m o csi 0) u a to •H CO CN CD c CO c ra S o 0)1 4->| r 1 I I (U O O CN CO CN CD O O M iH O O H W (ti 3 4J P -H frt •O 0) a X p w c CD u S-l rH rH • rH CD CN CN r• iH M OD cn r. "vjrH rH O M-) O rH o H rH r-{ rH O • • • ^ • • • • e^ • • • 1 CU 1 I 1 1 1 IS rol ro 4J .r4 a 0} (0 H U a •4J u -H ffl Ti CM C X H 0) M nj CH r-H CX3 r-\ H CTi iH rr-CO rH cn r-~ rH ^ CD OO O CN CN o r-\ CN o o Q) rH rH CO O rH • • • • g; • • • • • 'M ^ in CD p^ CO ^ in CO rCO ro CD ID CD CD CD CD ^D IX) CD CO a) CTi C5^ CTi (Ti CTi C?i CJ^ a^ 0> CT. ix rH rH rH <-\ rH i-H rH r-i rH r-i ^ in CD r03 CD CD >D CD CD 0^ cTi cn CTi cr\ rH tH rH rH rH "sj* tn CO r00 CO CD CD CO CD a> o^ (j^ a> rH r-H rH i-H rH

PAGE 196

180 10 fn h'3 rH I • CM o r*ro P ra = r-i & ~ > rH CO P rQ -r-i rH ro P ft! CO = ^ P > 44 P -H Ui rC A ro rH !>-l ra QJ c ro 44 (U 44 rH ftJ ra Z rJ tn > rC5 CO > > Cfi H •H 13 44 IP H r-j ro ro CO tyi m ^ a > CO A C 13 a> rd H >i >^ ra 4-) tTi ^ 4-> i/J (ij 44 P CD -r^ c: P +j ro tn !H 5-4 rG CO 13 CO Cji 13 •J a ro ro P ro CO -H 5 ^ ^ ro a 0-. 5 C ro CO 44 d CD (D A 13 Cn C/l Di M 3 \j >i >^ CD H CO ra P l)H -r4 P ^-^ t A 44 44 C 44 CO ra P to -P ra ^ (1) 0) -,4 H tJ. ro C)5 ;5 CO 44 CO ro > C LQ n 5h S-4 -H 44 ro ro ro CO ro rH re r3 0) CO 5 .H CQ 4-) in s (U ~ > u p o 44 CO ro A r-l ^ ai i m 13 ,3: Ci, ^C! rC CD >1 4-J CO A^ +> i >i C !D CD A 44 A en P il) 0) 2 A rQ ro ft to 4-J ? C r~i rH 5^ r^ 1 — i • 13 HH p r^ •H ro fC (D -P -P 13 •T3 a CN U A C CO A •H S > S 1 — i rH r -H 44 tC3 13 4-1 -P en o CD CD 4-) +4 A C CQ C P ra fD ^ •p 4-J rC rt CD ro C ro d p n: (1) CD P P ro G ^ CQ C.) g rH -r4 ro A H :-i 0) rC o-^ •H H O -H rQ 44 CO > I ^ H U s g r^ rc; 44 cu Q P P -H A H -p f.O K! •H H M A ft ro 44 44 = &1 H c m (D rH rH M p 44 ft p • H IPI "3 it! 0) CD CD D^ A Di ro A rH r-N P 4-5 -U ^ Oa Oi 13 +J ft = A O -r< m P P Oa ro ro tji 4H 44 rH >i rH ro A /5 rQ rH p c ip C .-Q -!! .Q :j 13 &> to ro tji P !T> ra "H 1 CD = CD CD CD -H ra ^ G >( rH d Eh 5H & iH c p. A rC +4 CO CO 15 u ro 44 ra -H O 44 44 ft to ro A -H > CO r-i oi u CD ro 15 p 44 CJ P ra +1 l>1 M f4 C C O ft B ai f^ iH r-i •H O CD P r; C 13 n A ft ^ CJ '44 li ro p p -H ra 4-1 ra -p u Cfl Cfl CD c 'C 13 g tf' W 44 44 -P :s p > -H 13 >i t; '-M V. 0) ro fO ro ro p o 13 ft ra 44 r-i Hi ra M CD CD P c: > tP -H d P nj CO i::A ID 0) CQ UJ ro ft .a 'd tn ro :C •r-j A n g u v-\ iJI H LO d H ft r -P Isro ro (U CD Qa rd ^ P ro CD Uj • Jh o:) CO A A ro 4-) 44 CO CO 4-1 H rO K) 44 44 ro e: P C p nj A nj 04 0) CD 0} +4 MH P 4J !h m rC ,£^ 1+4 M4 P ^ o H C d CO U C ro 44 44 O CJ g P > p ra >( S -H u A p p 13 5 rQ g i — i m to CD CD 44 to u C > Cji > Q) ro CD ra ro Xj^ b-i -r-t c ro o 13 Q > f? r •^ S ro ra > ri^ -H tfl tn 44 r! • 4J 44 44 ~ A ra r-\ 0) Q) S C-J Qi tn C c:; CO •H tH 44 pi H ^ .C, r-5 ;: d C (D CD CD 15 A = rC ,d CO +J A-i u .,..! r4 U o 44 44 44 MH 44 P Q -d 13 P p, ro •H 44 MH K'j ;3 CD CD 4-> CO IIH P IP IJH A d H m K rH a rj Oi Oi en H rH IH 13 H 44 nj •H w ro ra rQ ra > cu ,.a p o nd CD HH tn -p 13 ro rC^ the va '"2." -H C ro Z • rH CN •n party were c

PAGE 197

181 Table 111-13 .S OF WELFA?S EXPENDITURES IN FLORIDA AS RELATED TO TREl-IDS OF CHANGE FOR KUL STATES Year M,ea n Predi .cted Actual Flo rida Uiierripl oyroei --enefits $28.55 $27. 69 1964 ^ '•> -~, 03 1965 33. 65 28. 02 27. 39 1965 36. 17 29. 68 28. 01 1967 37, 55 29. 28 28. 39 1968 39. 52 30. 71 31. 81 59, o,Ld 93 Aae As, s_.i,stance. 49. 72 1964 "49" 31 1965 59, 13 49 84 51. 30 1966 64. 03 57 20 49. 15 1967 58, 21 53 .82 50. 10 1968 66. 60 48 .08 48 95 Aid to Fa mill 3 0.02 es YL .t"h Depai ident Children 15 .54 196416" .55 1965 30 88 15 .75 15 13 1966 33 03 16 .05 15 30 1967 36 28 21 .02 15 30 1968 3S 25 14 .88 22 .02 1964 71 Ai .81 §. to th 3 Blind 60.80 59 .74 19S5 72 .71 60 .48 61 .18 1966 78 .50 57 .5'4 63 .00 1967 81 .36 65 .97 63 .90 1968 82 .23 63 .95 54 .25 Abl to the Parma 59.04 nently < ^j.nd To tall V Disabled 56 .48 1954 55 .40 1965 GO .88 5S .46 57 .21 1965 69 .47 55 .25 59 .50 1967 75 .12 55 .07 50 .90 1968 76 .20 61 .82 51 .25 Gen eral Assistai ice 1964 21 .84 1965 22 .57 1966 23 .38 1957 24 .04 1958 25 .05

PAGE 198

1 po Table III~13 Coritinuod Predicted Actual Florid? Per Capita Welfare ;iixperiditu e 1964 $25.15 $17.37 $13.02 1965 25„99 19.70 18.98 1966 29.29 21.01 19/34 1967 33.00 21.53 21.47 1968 38-62 25,89 21.80 Wel_f,3ra Exoenditujie as, .a Per, Cent, of ,i0.tal Expenditure 1964" ~ 11.03 '" 11, .00 10.50 1965 11.91 9.55 9.60 1966 10.83 9.40 9.50 1967 1 5 1 9.48 9 40 1968 11.01 9.93 8.90 fhe value predicted by the regression equaition for that year was $14.88. The actual value V7af5 ?22.02. We hava no means of connecting this ch.anga clearly and unambiguously with reapportioniTient. So, we must leave it at this indeterminate position. H eal th' and Hospitals. Another ..service v;hich v/e associate with i^rban needs is the ar,ea of health and hospitals. The aggregation of people in cities raa]<;es public health facilities feasible and, at the same time, concentrates need, VJe examined this ^irea, therefore, to see if a reapportioned legislature v7ould spend rriore in this area. The specific variables examined in this analysis were per capita expenditures for health and hospitals, expenditures for health and hospitals a.s a, per cent of personal income, and expenditure for health and hospitals as a per cent of total g e n e r a 1 e x p e n.d i tu r e

PAGE 199

183 Table III-14 cresents the intercorrelations of tliasa variables over tne years of our analysis. The consistent • high correlations indieate little change in state ran'kings on these variables. The slope measures indicate a slow gradual rise throughout. Tables III~15 and III-15 which present correlations of the political variables with absolute variables year-by-year and partial correlations to analyze change show no significant relationship. Nationwide, there is no relationship between apportionment or any of our political variables and the level of expenditures or the level of change in these expenditures. In Table III-17 we look at Florida's change pattern in relationship to the other states. We found no signs of jumps in these expenditures for the state. The state was frequently above the mean on this type of expenditure (this is a rarity which might be explained by the larger aged population in the state) and changed at about the same pace as other states. £?i£lll^_Ys.' '^^-s final policy area which we wish to examine is highway policy. As we indicated earlier we have no clear hypothesis in mind with regard to this expenditure variable. When highway policy has been mentioned by reapportionment reforHiers it generally has been in the context of distribution of funds to local government3---a

PAGE 200

184 ;-i o I H a) G (13 PM H 4J H (0 rH a 'd u 0) o t • fi s fli w rH iH O O Id 4J -H IH ,c ci 4-> C rH OJ +J CO i>o j: ^1 ca ro &i c: 01 C''. 01 CO E-< O M rH fr; U • • • • (D J K ro A C/j f!) CT P C'^ K fd Q O s:; -^ m -tH <3 w iH fX, ro d i3 -H d JJ &i ri F-' --! W (D CM '.-M -r; IT) O H ^-C fc4 ^d D O O O i-H M u !-"4 O C K C d VI 1) U i-H rH \-i tH iH • H pi 04 ra G ra >1 r > HH cTi 00 CO 'X' m p CTi cr. 0^ 0-1 ai P in !-! S-l s^ o -H (U c m (U 'H 1 1 H p M-4 fd < ^ in w? [~CO P CO Vi3 MD iTi U5 5 M 5-t 1 i 1 I J-i ru ro ^' ir) U3 r~ •r-i a) UD UD i£i vO vD P3 1> 0^ 01 CTi C?i C3^ 0^ r~l rH l-^ iH rH

PAGE 201

;/. 18 3 CO m < a H S?, 5^ > ffi Eh O K E-i fj H 1 H O f=4 H O^ H CO rH H P ^ S E-i ra H H E^ H n w CO (^ x; P O H EH < O U CO m 1 4-1 m (!^ cr\ 0:1 4-) en tT' I-l -H < 1) ^ H >i Pa 4-> CO ^4 O O K U4 O 4-> fd P. -.4 b-^ b d; P-. g 0} iK +J 4-1 •H ST. 5-1 M G oj a> 0^ -H. 4-* t^ S'^ Oj Tji O cr* O ^ rH O CD r-J 0-1 O rH O Cn! rH O r-i ^-i 00 O CT: C3 o m m CN O r-i CM O H fl t t i t I i I I I U} O u a; CD O ra •H CQ 3 O CD J4 > O •<-' P4 4-1 D< 4-) +j •H tri O CD C U •H 4-) ron3 4-! S4 CO CO O M CO lH t4 &i CD 44 C 15 M m O ra ,c; I-:! Pt! O H C.Q 00000 O-l H 1 1 i M i 1 ro iX! sn KD CM rH O H rH O M-il r^ m i£) 1-4 rH n 0-) cNj iH o r^j 01 c U }-! o "=t f^ o OJ raj cN O rH C O .0 I ! Gl •H a CO o III fi CO O vO U) CO ro fc! O OJ O rH rH O I I I 0<1 LD CO rH !> O rH O rH O rH ro
PAGE 202

186 o •H .p c LO I— i I !--! H H aj rH (Ij EC H -p J-! (C o O d. u en m 5-!o CI) > o o o r-1 M f-ii & in "sf (Ti to r-H "sf O O :H !--) r-i r-f O u ml -H G Ml I fa > O I s •H rG CJ fO 4J -P -Hi 0) P^ O 5-1 M O 01 =Q' u") 'xi' C r -I JH M-lj iH O.j O -.-I O f-i u m CD fji G Oi C O U -P rCJ I'd V w 4-1 O M 4J c; O O C (0 tD -H 4-> S-i Pi O o r— 1 H H ro c o 0) o 4->! 0) O! !Ml| 'I' Ph! ral r~1.D CO CN iH CM 00 O m r-l B • • i i i I I -H T! O O iH ro -P o O c o O r-J Cn! CTi .-H O rH O 00 rH O O to iH fd +J H Oj tn o Tjj (til A +J ro CD M O IW o I I rH O CM O ^1 ro (!) M P H c a rCN] o LO H rH O rH rH Till ro p (ti :S fH (d 0) K M-i (D 5h 13 4.J •H HD C m I I I I I nS* vD in iH r-~ si' O O O O eH CN I 1 I i 00 -Nf tn uD rQO U3 i^) U) kD U: kD CJ> CTi (Ti CTi O^ CTi 1 — 1 t— i [ — 1 rH 1 — i I — I o") '^ LO v^ rCO U3 US i-D IX! ^D VO 0^ C^ 0> CTi (Ji OA I — 1 E — i I — I I — If — I I — ro vf' in U3 rCO 1.0 VD >X> 3 i-D ^O 0-1 CTi Cy\ CTi CJ^ (Ti H rH iH iH rH rH .Q z ro • •H IN M z ra > CO ro m 5 •H rC 0) 4-> 'p rH i ro Q > M 13 0) 4-) fi ro 4-> iH w •H >i tn 4J i 0) 4J tf •H M •n 4-> ro g IH H ro -o' X rH % (D iH rfl = CO 0) 4J 3 ro H u ro o > s 0) 0) rC Q P Q) Ti rd 4-1 c 01 m -H H CO ro CO ro m ro

PAGE 203

187 s 03 s '^r> • P • 2) rsi rH g IN = fd P ro = bi r > rH in 5-4 .r4 rH CD fti p to = n3 p > 4-1 P to rH rC rd I'D r-i BJ P 4J Si 4-1 rH ro CO > 05 rH •H 'd 4-1 m •H nS ra CO &i ^ • >1 > CQ A ra 4-' in rd iH -;J ! >i -H to fD M <4H -H •> -P jj G 4-> m ro P CO T3 a) 0) -H -rl tn ro CO 5 CO 4J m C z en m 5h M •H •p ro ro ft K! fO d 13 P O CO CO 15 r-t rH O r. n to CO P to to y) rC ^ -H ra ro ro (T3 -H P rd rH E g r^15 JJ +J tP 2 ^ = 'U M M 4J CO ro rd , JJ CQ ^ 4-1 Cm 4J P 5 4J ro rC 4J rti u ,_i P 5h 4-* 4J rd 4J 'G Q) >. >1 c ro CO rd 4J ,Q 05 Q) :3 ;5 r-H rH :: -H -P ro T3 4-) +1 CD •P 4J rc: c CO c fO 'O -P P /5 ^ c fd c rd d 14 P 54 ro C rd CO V 'O tr> TJ t3 Q) 0) g r~~i u o ro P d Q) G Q) 0! CO to c ra -H CO C ip M in P. ra 4J P P P > g rH H ra rd H H tn,c; -H •H -H rQ 4-' CO > 1 H g S rc; ^. 4J Q P U -H rd H m •H -r^ M rC 0^ ro 4J 4-1 = tr> H w CD rH rH M !h 4J a. 5H • H m tfi CD :5 Cu Cu ^d 4J a, = rd Q) U5 b P b a rti ro tri MH 4J H ro rC ,Q rQ rH P fi m C ,fl ^ TJ tn to rd tn P tP ra CD r 0) C) H fd rd d >1rH d B rH C fi rQ .j:: rC 4-' tj:i CO 3 U (Xi +J rd -H O r +J 4-) Ci^ CO ro rd •H > CO U rd ^ M 4-1 M -P >i U !-; sn fi u Cu c rH -iH -rl u u C d T3 •[—1 rd Oj fi ^-4 iH ra U M •H fS 4-1 o o CO CO c 'd TJ g m m 4-> -P V 5:. u > -H 'd >i c 0) ra (U CD fd u t:! Oi ra 4J n3 ^4 (P CD p c > &1 -H d !H w CD Q CO to o fd c^rd 15 tn (tj iH rC F^ g u rH tn -H CO d H Oi P ^: ra nj (1) a fi rd M fd CO Vi m CQ .^ rC nj 4-t -P CO CO (13 crj -P 4-> fd sn; iH d u ro rd Oj O G) 0) 4J ip 'H 4-1 CO rC ^ m UH p fi, rb r^ d d CO d nj -P P o 5h > 54 ro >i •r^ O .c P P 'TS ^ rQ CO 02 Q 4-i CO o c > cy > UH 4-1 u 0) tJ xi u U ra •H 4-1 p CD 4J CO 4-1 p m (p pC d ^ r-H u o Oi 04 CO H rH H 13 H 4J O ^ 03 > -P u TJ (D <4H tp 1X5 A fT3 > H C ro rH CM n party were c

PAGE 204

188 i H bi W 4-t 4-' •^ m U3 rC^J OD !>• m CN r— 1 in ^ iH CTi ) i-C 4J -P w rt! P K 4J ^-i K m CO ro rH (71 c f-l S CO rC CU p^ < H 0) r-l P > >-p G .-H ro! rol U5 p r-t^ .P r-: en ffl to m ^ i< CM uJ-l rJ tQ fC CO i o TO El H H Di -P > S raTj M cn cn H t-f o ..-^ 01 iD ro CN O sj< r-H "* • ri-H m • rH (N ro H H o .&4 Di O 0) -p s: Oj p ro U-: tN O H O -P -H ai r^i ro -P -H r-i O rH O Q) Q) 1 1 { 1 T-i CH li.'i jp 0-4 ro ,: P p. a /5 H i>.i (D P P4 U ? to TO fO ;:?^ h^! -P r-l -P C-H M kI M ffi H w in C'3 O O -H t:! jg 'X 'U CI c^ 'G TJ B B o 13 q c W t21 M r^ C! r-, fd fti ffi O E fo o >5 U U-1 CO o W ."0 CI ID CN CM ^ rH '^ U3 rC: o CO r^ rH CO o a CO -P ffl -P c > Vi rT3 o o O O cn iH CN o o •P rH CD rH rH O O rH CO •p i 1 r r 1 r 1 l' l' l' l' H tq (d fC rSn -H CD e^ Oi Q PI P^ tj K K < D — ra 1-1 w CJ S^ V4 f^ --C s r-r^. H C-Q pi MH -P rH O rH O cn rH cn cn O W r-^ O O O O 1 l' 1 1 1 1 i (0 Ch u p '^ in ^ rCO -^ in VD r00 <* in CD t^ CO (T! kO 1.0 i£i vo v.0 O li) W3 1.D VD wo CD VD CD CD K? cn (Ti CTi CT; CTi CTi CJ^ CTi CJl CTi 0-^ a> (Ti CJ> (Ti !>1 rH !— i rH 1—1 rH rH rH rH iH r-J. iH rH rH r-H .H

PAGE 205

189 tJ 0) rH G fl a ,Q = -H ro f-' (1) ra • 0) iB ^a (I) P H CM to •H Ki-n e r-< P = ra -P Vi O -P its ^4 M -H > CO to fu a 13 13 n3 ro ra pj, M J-1 W C tfi ^ ? 0) nj Q) -H IH > rH r^ a ^ 0) a -M ^H O O tH ro C .... \ I \ m O Q G O w rH (N rH rH rH rH ro i r z O +J CTi -P rH u ffi c 4J (D -H = Ti (D C rH U (!) TJ o 0) •H T3 C u Q) -n O T} W fO u ^ ra u ffl (D (D M •p g 0) ri 1 — 1 p; ^ P4 0) rS (D c fi. rC +J ft C ro rH P -H fD D') ^ c M -l-J -P -H CD 01 (fl rH d M ^ K m P (U ra T! u > ro u-i m ro rd rC: tt) Aj M 5^1 r>j LO CD • CO CO fO 00 KD rH <^ CN "vj< 0) t! tP o r-l o C tH CN rH rH rH rH. rH n CN O O to CO Q) iH iH ra d r-i C U-i • . ro • • • • • r-i H J-! i 1 1 -P 1 i i ro 1 1 ra Gn ^ H !^^ JJ !-i H -P -H to H 4-A Cn q -J ID o rH to OJ (D COO -P & to UH ^ S rd to 13 -H u ro W a CO o ra CD u § t 13 •P Di (d .c; r-i 'S D fU a H (1) O X fO P OJ rH fl W fi' rC 5 ^ o -P p ft3 r-l -p ra +J >l -P m rH e MH C rH H ro H c P-! oi ffl a; fti O rC N ffi T! = C (D j-i -H p H • fd P c| ^C vj IP • 0) rH to (D ,-1 f.t-j ifi (!) 4H 0) ^ = 1 ^ O o n CO 3 LO CTi CTi O P CM (N O n CN !lj -p :3 !^ m rj J-; r-l O H c\ o -P .H C-J OJ CN ::3 -P •H rH CN CN C^ 00 .-1 -H ro rd H p fd u > &< (U CO LQ 4-i CD H > 1 • M • .J-it It? 1* l' r i' 1 C ra -P a. CD P c (1) D X > P Demo the -H U w a (U TJ fd CN ^ rc; = ra P c jj tn (D fS 4H -H f3 Q H CO 13 rH (d CO rd fit (d u =:} LT) kT) rCD -^ LO k£) rCO "^ in ix> r00 i m vD U) ^0 UD kJD 1..0 i>0 U3 ID i£) IX) 1X1 <.0 U3 >Xl m s i a) CTi en CTi CTi CTi fd -f^ 1 '^ !->. nH r-l H rH rH ^1 (-1 rH rH rH rH rH rH '.H "^ •p

PAGE 206

190 OJ (!) '; P p r-i c/ii tH r-S fO ro CO CI) = r-l 'f^ -H P C it! •P '^J 4-j 0) ro 51 tn > .C rH > r-l H X) 4J ^->. iH CO 0) OJ 0)

CO rG C HD rd ^ >1 ^1 ro 44 tn CD Q) P 4-J i 4-1 0) M C !4 u M rC^ CO 'T3 CO tn 'D p -d ro nj 44 ra 0) CO -H ^ CD cu a 0 ro CO CO (1) G CO iT' ><. >i CD Q) -H CO ro M Q) -H 4-1 44 CI J-) CO ra P &T CO O OJ H H tn ro CO 15 CO 44 p w 01 CO M M -H 44 ro ro ro ro p p m CO 5 .-4 a '!— 1 m CO +J CO CO A rC •H fd m Q) ro ro CU -r4 rG ro g S rC & -P 4J tn +4 5 5-1 S4 03 44 CO fCj (D 0) 0) Qj Q) M a) P r-l A 0) ;5 0. ^ 0) >i -p C/J 4-J b Q^ 44 44 ^ +4 ro rG (U 54 ro tj iH p u P P (!) ^ P> >1 >i c: ro CO rG p CO p (D o ^ rQ ro = a, CO P a) c ,c; ^ H • T! M-! G 0) •H +j +J TT) T3 a CN D fi G OT ^ JJ r-( r-\ z rH P ro '"d 44 c Q .p P ^ G CO (1) 4J +) rd rC p; (D ro C ro (U ;G ^ CJ o P u (d G .G to p> ^3 'O Q (D g .H U V ro p G IvO ri) cu m CO iM fC3 -H w G 44 M H 4-) 4-> r> p > g rH •H ro ,G H (D 1 •H ^^ H .Q P CO > H g E r^ 4J (D Q P M -H ,G H •H •rt M rC P-4 ro P +4 : tJl H rH H M M JJ CU (U D-i P. • (D fl.) P. rd Oi ro ^ rH (D Ck 0. 'd P Q) Qa = ,G rH p a, (C3 <0 ro Cn 44 4J ^ ,Q r-i p c CU G (D 'a iji CO ro bT P tn Eh z Q) (D a) •H ro ^ G >i ^ G rQ ,Q ,c ^ 44 CO 01 > ro 44 ro -H 4-i +4 Da CO ra rG r4 > CO 0) ro 5 M P CJ U M S4 P. c o Q) 04 •H, u CD M s:: G 'd •r-l ^G a "W >4-i ro u M •H ro p CO M (!) C CD -d 'd g cc U) 44 44 -U ^ M > -H f-d >i !ti fS ro fC 54 OJ 'd a ro 44 ',D CD P C > tn H G 54 (3j CO CO fO CU^ -d &T fO Ci CJ i-H tn Q •H CO G H P-i fd (5 GJ a .c; rG M ro CO m CO ,-c; .-^ 03 -|4 p CO CO 4_) 44 ro fi U G M CC3 ^ 0) 4J (D ^ !-! 44 ^ ,d M-i IJ-I P .C ^ -rH G u ro >i rC::. P 3 t3 0) ^ 0-5 m 03 a) +4 CO d > Cn > P3 ro &i en r4 C ra ^d ^ <: (U ro :5 A -H tn CD en 4-) 4-> 44 +4 = rG ro rH Cn tP C G CO •H (D rH P 44 r4 G ro In •n party were c

PAGE 207

191 Table 111-17 CHAHGE3 IN LEVELS OF HEA.LTH AND HOSPITAL EXPENDITURES IN FLORIDA AS RELATED TO TREMjS OF CtlANGE FOR ALL STATES Year Mean Predicted Actual Florida Per Ca_2_i ta Hea lth and Hospital Expendi ture s 1954 $11^36 $ 8.12 $ 8.41 1955' '^ 15.01 11.69 13.67 1956 15.45 15.01 14.75 1967 13.07 16.41 16.08 1968 20.80 18.67 17.84 Ex penditures for Health and Hospitals As a. Per Cent of Total Expenditure 1964 6.05 7.73 7.30 1955 6.62 6.56 6.90 1956 6.00 6.81 7.20 1967 5.93 6.97 7.10 1958 5.83 7.18 7.30 Expend itures for Hea lth and Hosp it als a^ a. Per Cent of Personal incoine. 1964 1.03 1.47 1.50 1965 1.06 1.54 1.50 1965 1.07 1.50 1.49 1967 1.08 1.46 1.63 1968* *Data not available. subject with which we will deal later. The particular variables included here are highway expenditure as a per cent of general expenditure, highway expenditure as a per cent of personal income, and per capita highway expenditure, The results of the regression analysis, presented in Table III-IB, indicate that during this period there was little changing of position among the states from year to year and that there was a general"-although irregular — upward rise in expenditures. Our correlation analysis,

PAGE 208

192 Ta.ble III-18 'ORRSLATION AND SLOPE COEFFICIENTS FOR ADJACENT PAIRS OF Y3AFS OF HIGHlfAY M12ASURES Fairs of. Per Capita, Highway Expend i tur e, Yea rs 2I slope 1963-64 .97 1.19 1954-55 .98 1,25 1965-66 .98 .85 1966-67 .93 1.36 1967-68 .97 .78 Pairs o.f ^.ghda^^dBxpanditure a^ a. Per Cent of P.er.sonal Ye ars In come r slope 1963-54 .95 1.08 1964-65 .98 1.16 1965-56 .97 .85 1966-67 .96 1.02 1967-68 .98 .91 P,a ir s ijf ii.2XiJ'i3_iI Expenditure a.s _a Per C ent of. To tal Years Expen ditur e r slope 1953-64 .93 1.09 1964-65 .94 .88 1965-65 .97 .83 1956-67 .92 .92 1967-68 .94 .79 *A11 fifty states included in these calculations, shov/n in Tables 111-19 and 1X1-20, shows no relationship of appor bionnient cr any of our political variables to this type of expenditure on a nationwide bases. The analysis of Florida's position in the residual tables of the regression analysis (see Table III-21) suggests a number of interesting conclusions. Most notable, perhaps, is the fact that unlike most of our other variables the

PAGE 209

CO W hJ § H Pi <; 1--. tJ <^ u H (Ti R r-i H 1 fA H o H c^ H f? CD ly iH Kl ^ ro P' ph te 03 en s; O M Ph < l-i g r^ o u CO--I i Di -P D P E-^ O (-I 'P Q ty o i ft 4-1 '^' 4J I fe Q CD 0) rH Q ra •H M ro > (D > -p ro rH m PI -H -CT. o t-^ o u-i 'i> r-o~l CM 00 ro m O "^ ITl KO CTl H -u +.J .~l C T—, f^ r-i r-i rj O r->. CM 0^1 CM CM O rH rH C^J CN 5-i C 0.5 1 s p. g a < tn r3 4-1 OJ >i ,a u M -P -P CTi ro u^ f^ '^ on C) IJD rH r^T (Ti CM !> r3 O -xf 00 00 in rH -H tn O rH r-i i-l O CN g rH rH O O rH O +J rH O H rH rH O M 5-i C -H >..>• i A^ > U JJ ffi 4J -P ro CTi '^ O C?i VX> irj > (D nj nj 0} f^'! S: ^ P p rd .s C r^. bi tr E ro o P -H -rH Cfi 0! fcj r-~ 0^. CM o li) vo iH CTi CO CM CO rH Cvl W ^ r^ CM 1^ VD CTi (D -P C ::i 5-; ro Ul O O O 00 O rH O rH O rH rH O Ilia ^ k • rH O O CM rH rH 1,1. 5h 1 1 1 H • • • • • 111 II III II ro ^ CD ip m i-:i cu u ft ffl 0) J^ u c •H G^ ^ •p ra >-< -rl -H (i) CO V,0 CT^ rH O in ID T^ O CO 00 "^ (T^ 00 '0 CO rrH CO rH CO q -P -P ro M G r-< C O ro O rH 0\! rH O rH O O C^J o o o o o \ \ i ft III' II VJ • • • • ft 1 1 1 1 ri" nj !K^ X O A4 O H M G ro ro H CO -^^ M r-l tN PO ^ OJ O CN Ca >* O CTi -nC ro VO CO "sf r^ rH P -P ^i G ro rH i—t O 00 C) ("n O O O CM CM r-H CM 00 CM CM CM CM ^ u u n ^ i-n vD rCD 00 '^-in VD [-cx) 00 'd' in vD rCO ro ivO U5 VD VD V£) vo S-O vX) U3 kO UD UD VD V.0 (-O VD ^O VD .^ CTi ff-i 0^ C5^ CTi CT"! cr> (Ti cTi CTi cr. O^ cTi 01 a^ H iH rH rH rH rH rH r-l r-i r-i v-i ^-i rH rH rH rH rH rH r-H

PAGE 210

194 13 C M ai O a t! -rH fO ^ n M 03 Q) •H M 5-J +) P M s::^ -P ro M (Ci CM 0) rH > U5 IH -r-i o (1) m H^ U c -G (D rH C P (0 M u b-i 4J d c o c: ro CO H (D-H 1 +J ^ D. M -P -H ro .C fii 4-) W ^ C H rH c u CD !h OJ tn 4-) > fi c:: n o o .a a U rH M A 4J (X c -H ,C3 a (D CQ ^ U c; -p OS G Cfi l-H ^4 H cu +J > 'H ra O PM •>-0 <}' UD CD 00 ro O O O rH O .-i to > ij -H LO CO o o i' r (.0 "xf CnJ CO C\ CN O O O O i-i iH III' I o 01 M &, 01 4J u (D IT) ^ PM CN rH • • tc! CO 113! m > 5? •g •rl W u 5 o o CO in rH rH i I +1 a r3 •P O M of c; o u PM tol m m > I •H a ?4 O IP o ro ^ ID n rH rH O O rH O 1 I I OJ CTi LD iH O CM I I I CM O iH OO CD U M a 4J p r! H Ki •d G c CD PD nH CO rH CM rH -sf X 00 CO iH ro 00 rX CM 'X) rH IT) CM 00 rH O O O OJ OJ H O O • O o • rH • P 01 c?i cTi cr^ rH >H .H tH rH rH 0-) "sh !n vo r~ CO (.0 i,D vO 'vO "vD vn CTi (T'. CT" C71 CTi (Ti r-H i-H rH fH M rH ^ z m -H CM ^ = ro > CO ro CO ^ H ,^ CD -P 3 rH •. n3 CU > iH p CD +J ,^ ro U r-i m ^ •H >i tn -P cu -H 7-i ^ 0) •n ^ ro -p ti G ro -H 'd CO H •P CU ro ,£1 cu CO CO c CD ro A CJ -P -H rH in .Q 3 Oj >1 CD +J pe; •H 5h CU fi n -p ro B m H ro 'O : rH • fl) rH rC CO CD p 13 ro rH M fd o > o h CD ri^ P P Q) nrJ rc; cu -p c JITI ip •H H CQ ro to ra CO ra S

PAGE 211

1 OK 95 Ci 13 0 = ^i t! 1-3 o ra m I—' flj g 4J iN a ,G :U P ro ro = tp r 4-) -:4 > .-H CC rH •iH H ^ ro 10 U5 z 13 OJ > 0) -H ;3 0} OJ = rG .i;: DT rH ro (Ji d) -P ,G 4-> a .H > cc r-H -H ^ 13 4J m fC rc] 00 rH ^ • > M ra G 13 .G rG >i >i ro > tn 4J 4-> (U P +J Q) -H G M tn u M rG w i >i 0) CD 13 CO ro 5-1 -H • ft -p -p a 4-> Cj ro ;3 CQ Cfi '0 (!) -H •H en ra r; 5: CO 4-1 pi Cfi s = CO CO !J ^-i •H 4J b ra ro o ra (D 13 p (fl CO -rH ;5 H ,G rH c n CO Cfi 4-1 GO Cfi z .c; ^ H ru ra CD CO ra •H rG ro r-i s g ,G (D -P 4J 'CTi 4J ? 13 M "H (D +J CO rd 0) (D (1) (U 0) 5-1 c;j 4J rH rQ M-i '^:s 5 Oa ^ r^ 0) >i ra to 4-1 O 6 a, p -P 5 4-1 :s rG u (d iH M +J ^ 4J •X3 ai >l >1 G ro CD rG 4-> CO Q) 3 > iH rH -H m ro 13 4-) 4J (1) 2 -P : 4J ,G G CO r^ ra T5 4J -p rC rC C • ra CD ra rG M (D (IJ C^O J-i rG C rG en 4-1 u '•C Cn TJ 13 (D OJ e = 4-> ra G C) 0) Q) 10 m c CO G Mh M (Ti d fd -P -P :3 ;3 CD g G -H ro rG H iH bT rC H H •H U CD ra 4-1 CO > 1 •H U g g r^:: rC -tJ r-H Q M -H ,G O H Ui •H •H U ((3 -H ra 4-1 4J = tn H Cfl (D rH rH M M > CD rH a, 5-1 • H fO m (U Q) a rC rQ ro ,Q rH 4J -P 5 a, Ch CD +J P Qj = rG o; 00 O P PJ b 0-, It ^a a< tJ> >+H 4-1 rH (1j r^; rQ ,Q vH 4J IH CD G ^ is 13 Di ro bi &1 ra Ci z (D (0 CD 'C rG G >i rH G Eh .H a ,.q rQ rG rC -P Cfi ra o ro P ra -H P -P 0^ G ro rG •H > ro VH CD bi 5 CO 4-> U 54 -P >i M M G ss O -H ra a. c i-H •H H U m ;-! '^ C 13 n rG &I f: 1^1 ^;-i ro en -H 2 +J o CO t/O ra r:: u 13 13 g w m -P -P -P !-! o H 13 >i r^ 0) (T; nj f" (Ij U (U Q) G 13 Qj ra 4-1 ro ^H CD CD 3 U > U -H G M U-j C.) CD CD W CO O (P O CD a ^ 13 rp ro H rc: g g :5 tJi > -H Cfi G H Qj J-> ^ ro fO (D (D rG M ro CO U M t;) rd ^. ra C 4J bi m Cfi ro w -P -P m C 54 G ?4 ra rG a, (D (!) (D -P rH Q) CD !H 4J w r^ ^ U-i IH 3 a i ,G G G Cfi r^ rti +J -p O -i-i !-l > !h ro >i 't-i u rC ro ;3 O 15 rQ w K (D -P o in > tn ^ CJ iU (C Q-i bT -H ^ c H 13 tJI = 5 5; nj ra :5 u -H tn in 4J c 1 iJ -P ;3 13 rG ro "H fC CM tn tn fi C Cfi CO G) G 4-1 4J rH ^ C d = -H 4-1 (D a) -P -H ^4 H "H IH ^ G tg rH c; U PU Oi en 5 H c H 13 H 44 rQ > o 13 ai u-i b-i CD CO 1 rC ro > H G ra rH n party were c

PAGE 212

196 -ritP .1.' ..P n'; oD <^ O lH c-i '.n in rH ro CN in u") r!-j ::: "^ r-J rH rH [H t-J03 O O O '^^ r'H o O H Q) • • • • • V • • • • 1 1 1 i I I 1 < Q) CO >> ,C3 m < >1rC| JH -P 4-! CO o] CN o; O CM rH CO rH O CN CN CTi "sf in CO tH Cn o O .H rH O o n o o rH m CNI o O O '& § M M G 9 • • • • • H 'U 0) 1 1 1 1 Id) till PH P a t: ^ 5-1 Eh CU -H -P OJ P iH P S CO £ 4i S p •H .< W — (U V 1 m PMTJ m H s W x! (D r3 CD D H^ pqro >i ,d rH !i< ft) d m 4J .p a)\ fO M rt: W .H •H tn u o '^' rj CTi 01 a (\ -^ (T) en o CM rH rrH cr> H CJ !-) M C 3 rH O O O rH O (M rH O O rH OJ CM O O O tf c 0) 4J • • • Oj . • ro < i<;-H :5 C M H 1 1 U 1 1 +J 1 1 1 > f>H m ff) a) -H -P h3 S CO ^ i-q Oj .H (!) .^ CO J-l ,a & ^1 m o u p fd ro OJ X ol CM H O -P H CQ w 1 Eh H D-! M 3 p +1 H H > fO S3 > c d H H^ y 4J > K fOTj OS Q) H O f^fl W 3: O • V.0 rH • u "* • CO 0^ • u in • cN o • P-i Ptn 0) S-) -H tr. & o o O rH O O rH rH Q "i S 0) +J C tJ • t • iH • • • • • M rH t?; (^ Q) Oi M ro H 1 1 1 1 cu 1 1 ^ CO O 1 +J CU !Q ^ tJ-i p^ Ph ru cq Em r--^ ro p P. O E^ ^ -P rH ra ro! rtt ^. t5 M CO P H S •H (U Cfi CQ l-^!4-i & M 'B m uZ! Q) ^ fU CO iJ U h^l C ol a! CD tg O td ro (J !H U O K 14-1 en ID M! U3 ro cTi CO ;3 [--• CD cn 10 CD 13 O O OJ CTi rH H E-< S-i -H tn (Ul O O rH O H +j o o o o O 4-> O H O O rH P-i S (ij -P C Pjj • € • -H • • • • • -H rt O-P 5 U fO 1 1 TJ 1 1 'd 1 1 1 H-l U fd ro r^:; rfi i^ S 1-:! Oi U (1) 0) 2 a 0. P< '0 X M o gl fi M H u •rH ^.Q D ro 1—1 > > J CO tfi o rO "^ rH nj ^tID t-\ CO -fi ro in O CO rH ^ fti < tJi -H M o O r-4 O O S o o o o O ^ O rH O rH O H &i fi 4-) -P • • rC • • • r-G E-i (0 !-l ^~i 1 tr 1 iJ r \ i 1 Pi rC ro -H -H < U D^ U K K Ck ro r-H tfi 'Cj' iTi CN "^ CD CM CN CO t^ CTi CO m CO OJ "^d* -H S-J O rH O O rH rH O rH O O ro ro O rH rH -P -P f • • • • • • • • • • M P 1 1 1 i 1 t 1 fO O P^ U u ^ IT) CO rCO ^ in CD r~CO ^ in KD t^ CO 'V UD '-D CO CO CD CO CD CD CO CO CO CO CO VD CD (U C7^ CTi (Ti CTi CTi en CTi CTi >H [ rH T-i r-^ r-{ ^ rH rH rH ,H rA tH rH rH rH rH

PAGE 213

197 o -H o CM H H H 43 CO to Q) o s rd Vi ft! > O '-i-i o w H Di o o CD <-'. C ?3 O fO iji -;-) rt id o o J CO I n 1 m |!> IN 11 .3 1!^ ir, U-! "cjo in .-J. O C.I H • I ^ u fo m p^ G C -H -.-I 0-' .-P m c; c o It. u u p., CO u fi K -SI I o c > j U u (a I r-l. O "sf o H H (4 a M m p* M o| P: C 0) 5-1 P 4J •H C 0) a X) b ml 01 O in rH (^ U! iH H O 4-) C \D cu en u in o tH CM iH O rH 5h rC .g Oi s:; r-i ^1 c; Q> U) fi U c P. ra C S-l 1 0) P4 fcl VI 3 4J ^ 1 -P > J-1 'u o P-! > >j n3 n3 >=* CM ^ r-~ m ,"?: u~i in CM vD ro :? CM r^ rr-l cTv O O O iH O rC O rH 1-4 rH rH rC in IX> r-CO 'ti' in VD rCO ^ in vj,-) r-00 2 vO t,D <-0 1-0 vD KO D u? U3 U3 KD (D C7^ cr\ o^ a^ 0-1 C7^ CTi (Ti CTi CTi CTi CTi o^ CTi (y\ '^ rH iH r-l r-i rH r-l 1— 1 rH rH H rH rH r— I r-l iH ^0 (U o c I—! tn .Q = H rd • CO H CM m s-l = ra ITJ > CO CO (0 rd tc ^ > •H XI di 0) -p b^ rH c k 03 (d G) > rC 5-1 u ;:3 . rH •r4 >1 = b-1 -P rH (D H = T! rH 54 iH -d Q) m ^ fd u 4J ^ p C .-D rH 54 H ro T3 > "? CO rH 4-> fd T) fti ,c! (D rs in CO CO fi c c m (D fti &> rC rC u •H o 4-: H to iH CO 44 ^ fCi CO p a m >1 Q ra rd 4J 0^ •? •H 5-1 Q) rH d rCl •n 4-1 54 m +J >i e 44 d rH H a CD O Ti ; d rH • ra 54 0) rH CO r^:! = •H ^ •P 5 to d fd H U g (D ,a c 4-1 d • T3 (0 CN rc: (i) rd = +j G o bi iH H a H Cfl "Zi r-\ fO CO A3 m ro > CO rd rd 15 -P

PAGE 214

198 • iNJ p G) S CVJ ~ rH p ro 3 r (d rH cn U tH C11 > (Si P = p > OJ P P ro H (0 s tC3 A & •P iH nj P Sn > -P /:; cn > t-i •r4 44 U-l rH ai M t:! tji • > > CO ^ (ij •d A >t >i ro 4-1 fi H +4 +j 4-J Q) t^ a '^ M 54 I—* CO '0 r-i tr. d ro fO 4-1 ro a) C) -r^ "fi rC: a. P. ri 5 cn C;} fO r.'j 4-1 c Gi >1 >i 0) 0) -r4 ro M ^A r^ 1 • •p 4-1 C 44 CP cn P CO 0) i 4-1 CO ,d +4 k5 Q P-, 4-> OJ ^ OJ tij P fd CD i-i p 54 P A ,C1 +J P >i >i fi ro w 44 ^ 44 rQ m a (D (D ^ fit 13 : &. jj -rH ,C rC r-i • aj hands ands m 54 i-i 4-1 +J 'd 'd Da CN} U fi A d r-f rH z •H 4-1 44 n P > g iH 1-4 ro rC H H •H •H g -r4 0) ,Q -P cn P> H g e rC A P n •-! Vi ^4 .a H .-r-j H u .i:^ a fd 4-! 44 = tn cH tH !-l M 4-1 (ij Qa u • (1) 0) nj Pa rC K ro ^ ,H H 4J +J ^ Q< Oa Ti 4J Qa = ,£3 rQ a P f^ ft! Q) ro tp 1)4 44 ro ^ ,Q rH P fi m C iri n3 !J| o cn ro tn P in ::: :: CU a: 0) -r4 ro rC d >1rH i.-^ ^ ^ .C -P cn cn 5 ro +4 ro -H 4J JJ D4 en !T3 rC -H > CO 0) pj 5 M 44 u M u 5-1 ^. Qa Q r-i •H U -M 15 >i ro fD !T3 ro M 0) o -d a. ro 44 0) P r: > b"i ri C 54 Q) CO Di fU Da A ^J !j> ro g fS U rH bi H c;: si; •H Q-A fO 13 e m C u C 54 fd A Q) w 4-1 U4 M 44 rfn ^ t •+4 P r£^ ^ -r4 ri c; cn 4J -P u 54 > 54 ro >i x; .—( P Ti ^ ^ CO m tn > fD f!3 bi tr> •H s n3 ^ 15 (D 113 •^ rc: •H b^ lj> 44 P 4-> 44 = ^ nj rH tJ' tTi C C en H (U rH 44 -p H c c: O 0) ^: rc; : x; .C CO H H o O 44 44 44 4-i 44 54 -rs tj i4 M nj 0) -r4 4-1 Q Q) 4-1 w 1+4 P M-4 44 rC c o u C-, P^ m H H H 'd H 44 u '-d o; m &i J.! -a a .^ ro > -P ro -n party were c

PAGE 215

199 Table III~21 CHANGES IN LEVELS 0? HIGHWAY EXPENDITURES IN FLORIDA AS RELATED TO TRENDS OF CHANGE FOR ALL STATES Year Mean Predicted Actual Florida £.§£. .?3.:^&it^ Expe nditu re for Hi ghiva ys 1964 ?65.25 $47.15 $47.59 1965 70.47 48.36 48.86 1956 71.01 52.70 44.49 1967 77.99 41.82 47.84 1968 78.11 54.57 43.84 Hi cfhwa y Expenditure a_s a. Per Cent of Pers onal Income 1964 2.94 2.37 2.27 1965 3.03 2.26 2.21 1966 2.85 2.15 1.88 1967 • 2.83 1.84 1.86 1968 2.71 1.84 1.58 iilslMax Fxpe nditure as, .a. Per Cent of Total Expenditure 1954 29.07 30.28 27.70 1965 25.90 24.70 24.80 1966 23.90 22.99 22.00 1967 22.71 20.95 21.00 1968 20.95 19.59 18.00 national mean and the values for Florida are declining throughout this period for highway expenditures as a, per cent of personal income. This suggests that highway expenditure is losing position in state budgets both in terms of its claim on resources (personal income) and in terms of its priority position in the budget (per cent of total expenditure) This fact becomes more interesting, however, when we note that the mean for per capita highv/ay expenditure rises gradually over this same period of years. This strongly suggests that road-building is being increasingly federalized.

PAGE 216

200 We should also note that Florida has tended to lag behind other states in the rate of change on highway expenditures. Or, putting it another way, Florida dropped off more rapidly than other states. This perhaps indicates that Florida grasped the notion of tapping the federal till earlier and more completely than some other states. Florida's extensive utilization of the interstate highway system in the last decade tends to be congruent with this interpretation. This means that highway building in Florida has increasingly become tied to federal decisions. Apportion| ( ment has had little effect and is li'kely to have little \ under these circumstances. { SumiTiary In the preceding analysis, we have approached the relationship of apportionment to state expenditures j I from three directions. First, we examined absolute levels | I of expenditures over a period of years in relationship to several political variables (including apportionment). i Second, v;e examaned change in expenditures in relationship to political variables through the use of partial correlations. Finally, we looked at the pattern of change in 1 expenditures for Florida in relationship to the change of other states to see if we could identify any extraordinary changes in expenditure which could be related, at least in a loose chronological sense, to the reapportionment of the

PAGE 217

} 201 legislature. Approaches one arA t\vo failed to produce any interesting relationships. In general, it seems safe to say that our analysis was a further confirimation of the findings of several other researchers. That is, we would have to conclude, with thein, that party and apportionment show no systematic statistical relationship with level of 8 expenditure or change m expenditure. We want, nevertheless, to re-emphasize one of Sharkansky's conclusions on this matter. He says that the inability of political scientists to produce significant correlations between political variables and expenditure variables should not be taken to "signify a lack of political influence over state budget-iriakers Findings only show that certain readily measured political characteristics do not vary from state to 9 state in the same manner as expenditure." xn other words, the fact that a political variable does not systematically covary with expenditure variables does not mean that it is of no importance whatsoever. In the language of factor analysis, any distribution contains common variance and specific variance. That is, part of the variation is, attributable to general factors which are related to a given type of dependent variable across time and measurement situations. Another part of the variation, however, is attributable to factors operating in particular tim.es and

PAGE 218

202 in particular rrLeasureiT.ent situations. What Sharkansky is saying, then, is that correlation analysis of the type we have presented, and that he and others have presented, demons trates that political factors are not part of the common variance. That is not to say, however, that political variables are unimportant, although they cannot (as presently defined and iCieasured) enter into general laws of state expenditures. Our third spproach (the comparison of Florida's actual y and eKpacted values on our expenditure variables) examines the relationship of apportionment to expenditures within the concept of specific variance. We looked, in other words, at Florida's pattern of expenditures as compared with other states and tried to relate it to the historical fact of reapportionment. Our findings in this part of the analysis were indeterminate. For two education variables and one v/elfare variable, v/e found Florida spurting significantly ahead of the values predicted by the regression equation. These spurts occur in the period after reapportionment. However, in the case of education there is another event h which competes with apportionment as an explanatory factor (i.e., the threatened teacher's strike of 1968). We have no real explanation for the welfare variable. On the basis of the regression analysis we are left with airbiguity which we were unable to resolve.

PAGE 219

203 Looking bacl-: to the quotation from Malcolm JeWcO-I with v/hich v/e began this chapter, we note another type of perceived consequence of malapportiorrmeiit Jewell suggests that local governments suffered at the hands of a m,.alapportioned, rural-dominated legislature. Am.erican local governments are in many ways dependent upon the states, Jewell mentions two types of dependency: (l) the need for authorization to act — symbolized in the hom.e rule struggle — and (2) the fiscal dependence of local governments on the state.. Jewell suggests that the state-local legal dependency m.ight be painful for local governments in a malapportioned legislature because of the unv/illingness of a rural dominated legislature to ma'ke broad grants of power to local governir.ents to meet the burgeoning demands which result from t7,rbanization. O'ewell also points out that the state-local fiscal dependency can be painful in a state with a malapportioned legislature because of inequities in fiscal resources granted to urban and rural areas. Implicit in this and in the writings of apportionment reformers v/as the hypothesis that local governments would get more equitable treatment from a fairly apportioned legislature than they did from malapportioned legislatures. We want to examine this exoectation in the case of Florida, to see if it has come

PAGE 220

204 true. We look first at state-local fiscal relationships and than tarn to legal relationships. State-Local Fiscal Relat ionshi ps Because of resource limitations, it was not possible to collect and type the amounts of data which would be required to carry out a comparative analysis of state-local fiscal relationships in the age of the reapportionment revolution. We pursued instead a rather simple and descriptive approach to the problera. In the Florida case we looked at several statelocal fiscal relationships to see if notable changes occurred in the period after reapportionment. Recognizing that this type of analysis is frequently ambiguous and indeterminate, v/e thought the question of state-local fiscal relationships important enough to warrant exploring it — even with limited methods Another important reason for giving attention to this question is the fact that is was regarded as being the very heart of the reapportionment problem in the Florida case. For example, an editor of one of the state's more respected and influential newspapers put this problem at the very forefront of his case against the m.alapportioned legislature. He stated the case as follo%\?s:

PAGE 221

205 Florida's "Pork Chop Gsng" is a group of 22 state Senators who dictate what lav/s are passed or not passed by the BS-iaeiaber Senate and the 95-member House of Representatives in the state legislature. These 22 rural county ^Senators represent only 18 per cent of Florida's population, vh-O pay only 14 per cent of the State's taxes and get back 27 per cent of the state's benefits Stated another way, the 685,000 people in Florida's 40 small, rural "Pork Chop" counties ^^aid state taxes of only $9 5,2 50,79 2 in 19 59, and got back in state benefits $175,689,828.1-'12 That editor, along with other writers, stressed the connection between the malapportioned legislature and statelocal fiscal relationships. We felt, therefore, that we should at least explore it. In order to carry out this exploration of fiscal relationships, we need first to specify the area of study further. We conceived of three important types of relationships. First, the state can share the fiscal burdens of particular programs with a local government. Second, the state can share its resources with local governments by refunding, partially or completely, revenues collected from statewide taxes. Third, the state can strongly affect the fiscal vitality of local governments through control over their powers to tax and spend. In our discussion of the fiscal relationships between the state of Florida and its local governments, we discuss all three types. As examples

PAGE 222

205 of state-local sharing of program burdens we discuss education and road-building. As examples of unearmarT<:ed distribution of state-collected funds, we discuss the racetrack funds and cigarette taxes. Finally as an example of .state impact on local fiscal powers we discuss the ad valorem tax and its regulation by the state. In all cases, however, we want to emphasize that these relationships affect the local governments differentially. We shall be looking particularly, therefore, for changes which show some greater state interest in providing urban areas with the fiscal capacities they need under conditions of growing service demands. State -local program sharing — roads The state of Florida began its involvement in road-building in 1931. This move to shoulder a local government financial burden was part of a larger program t& bail out the counties in this period of economic distress. The state levied a seven cent tax on gasoline and ei-^rmarked almost all of the revenue (that derived from the fifth and sixth cent of the tax) was designated as the second gas tax and earmarked for distribution to the counties. In 1942 this arrangement was incorporated into the state constitution. The distribution formula, as provided initially by statute and later by constitutional amendment, allocated one-third of the funds on the basis of

PAGE 223

207 population, one-third on tha basis of county area, and onethird on the basis of the value of the roads provided by the county when they were tiirned over to the state in 1931. In 1949 the state passed a sales tax and the seventh cent of gas tax had its earmark transferred from education to roadbuilding with, the saF.e distribution formula provided as the 15 second gas tax. This gasoline tax revenue became part of the reapportionment crusade because the -formula for allocating these funds for county road-building apparently worked against those rapidly urbanizing areas of the state. Particularly that part of the form.ula pegged to the state of roadbuilding in 1931 seemed unlikely to serve the interests of the most urban areas. The actual impact of the distribution formula and its relationship to county collections and population is demonstrated for the 10 most populous and the 10 least populous counties in Table III-22, By comparing columns I and II of the table it is possible to get an idea, of the relationship betv/een county population and county collections of the gasoline tax. This indicates a close positive, relationship between population and collections. For all 67 counties the correlation is ,,99. The relationship between population and distribution of gas tax revenues can be seen by comparing columns I and III. This indicates

PAGE 224

!a o H t— p-i [-( CO H -. Q H •0) 6 CJ -W ci Li b Lf> cc^ '.0 ui r^ p-; CM r>j CTi CTi in CO en fO "^ r-l iH X t-; en Q cr-. rH in r^ fM rOT Ofocor^-cnmconiOrH fO CM I— i CMCNCNicNCNf^!n-=l'ror>j cNOiix>crirH[H"Nfcrir-cN Bh ro rH rH rH -H rH rH CM rH rs! ^ 0) u fl3 w 1 03 rH iT5 m -P c ^j 1 £-! r-j 4J M-i r--ViDvX)U?iHU5r--Or-i(NCN •v}'C7^cM0^lnlnncr,mlnln niHoif^LDncNvD'^cTiro '^LO'^"=d''*int^cM'^incri CO in ^ -^ n rO <;}< 00 CJ rH r-i OOOOOOOOOOsf c (C ^ 0) SH u 03 !-! m .1) a P. rH CO ra c; -p M EH i 1 '' i ^ H il) 1 r. 1 to N H '-U P .. fC rofOrHcricncocDinu-)in sf pH in r--! u3 sf" rH i-n CO fX) 'd' r-r-~ rHrHrHOOOOOOO CO JJ P • • • • • • \ • n D.i 0^ 0^ r-C^ LD =:? ro (>-) CM OOOOOOOOCO H r^ u CM M d) +> CM -P rCi tn P ^ U -P (D 5h ca rd ro C (0 4-> fu ^ Q) -H T5 J -H M fU -H -P >i J rH a' rH CO u cu 4-* Ht; >i ^ rH M 0) CQ ^ !h 1 !-! IH C a)rDrHrj;5Cg^ro>0 Q) C b-1 :3 -H 15 U rt! a' Cj 3 ^d ,Q r-l lO rH r-i £h r^ ro -H ro ,ii X nj rH UH ^ E-^ fD D -H -H, M M ro CO M ra M C rH m -H r^ -H fU tH 1 u QQiXjCMCQOCMpjMCQ ca &j D fc :e Q a ^q m s

PAGE 225

209 a positive correlation (,94 for all 67 counties) but also shows a redistributive effect from more populous to less .. populous counties. Whereas t"n.e ten most populous counties accounted for better tlian 70 per cent of the state's population and paid about tvjo-thirds of the gasoline taxes collected, they received only about 40 per cent of the funds distributed. By contrast the ten least populous counties made up lass than one per cent of the state's population, paid in a little more than one per cent of the gasoline taxes collected, but received almost 5 per cent of the funds distrib'uted. While the percentages may appear miniscule, this means that these smallest counties paid in about $1.8 iTiiliion in gasoline taxes and received about $3 million in disbursements. A feel for the redistributive effort of the formult'i. can be derived from reading down column IV which expresses the 'amounts received as a percentage of the amounts collected. It should be noted that the size of the ratios is somewhat exaggerated because they are based on total collections while the state returns only the revenue from three of the seven cents to counties. Nevertheless, the relationships shown by these ratios accurately display the relative tax or subsidy imposed on each county. A comparison of the 21.5 per cent ratio for the most populous Dade county v/ith the 421.1 per cent ratio for the least populous Liberty

PAGE 226

county is rather startling. For aJ-1 counties there is a raoderata, negative corre].ation (-.41) between this ratio .' and population, deraonstrating that the less populous counties enjoyed either sharply reduced real tax rates, or even subsidies, while the real tax rate was much higher for more populous counties. It was this disparity between amounts collected and amounts dispersed to the counties which attracted the interest of reapportionm.ent reform.ers. The distribuion formula for the gasoline tax has been changed in the period since reapportionment. The 1967 session of the legislature revised the distribution formula for the funds derived from the seventh cent. The new formula provided that the distribiation should be threefourths on the basis of collections and one-fourth on the basis of county area. Moreover, in" pursuance of the ,new state constitution of 1969, thd formula for the second gas tax was m.odified such that after January 31, 1969, the distribution was made on the basis of a form.ula giving onefourth weight to popuiation, one-fourth to area, and onehalf to collections "^^ Table III-23 shows the relationship botweon population, collections, and distributions for the ten most populous and ten least populous counties in fiscal year 19G3-70. We have divided the distribution into the seventh cent and second gas tax distributions because the

PAGE 227

O H E-4 C' O I--: 1^ pi. C' fr-: VD CO '.:7! f. Q ro ^ M < K W— ( y) S 13 P O O H U Eh O M W P i-:i O 1-1 ^A O D en U fn CM O i ^^^ H fH tH H 00 M < (U a M !-l l-l ,^ : 1 fi! B-i : m a a o H OS QJ Si ,.H 4J (W .Q H M 4-> .U i5 in (D a •H U P U~) r-CM C7^ 00 m CO CM in uo i^cM-NfvDOO'^inootn in n •H CM rH r-.H CO CM iH CNI in CTi ooO'Tj'cocNOOinrHcn VD V IT) CM ITs r-'^ o u"i cy, r^] 0^, O^DOlnc ro ID ^ in U3 ^ ^ CO CO CM in' ooooooooo O 00 ill CD 0) .—1 in u o C4 }^ *-< X 0) •,u ra JP^ r^ C-i X (' ro 'W r-i to r^ US CM CM oo ro 0^ :H -o UO coincM<=trHcriocDcri r^ CO 4-i fU OD ^ tH ro r-j 00 o CO "X! i-O o rHOO'^OO'^CM'^'^rH CM CO a O rn • = C f^) kO u-| uo vD "si^ -nT cv! 00 r^j yo ooooooooo O 00 iH u I— i in nj 3 4J +' 1 1) (!) I A. 1 '^^ Eh ^ ., 1 1 y. i ra n 'll-l B-i w 1 c K! tn [>. CO o uo iTi 00 r^ o r~rCO^'^OCOrHt^OUO tn f^ -!.;r^ -H LO r(Tl rH ID VD i^D CM UD (D CO O^. <-{r-\r-{r-^C>v-iO o o 13 -U g b y: CO U-) rr-^ in 00 00 CM inOOOOOOOOO o rH CJ t — i rH VD fd rH M 4J r-i 1 C'^ E-! o >4-i ,JJ •M r-rrH o IX) CO r~r-00 r^ t^ iHrHOcnco^oininin •^ ^ ^-^ -;-) CO :--{ rr^CM iH rH 00 00 o rMrHrHOOOOOO o r-S ro • • • • 9 • i ^ G rH CO o 3 ^ P4 O) Pa CQ U urn ro 4-) 0) •-a ro Q) TJ -H .^ C -H ro !h >i -H (D 3 >; M rH ^ m !^ ^ <; r::i rH rH (u +J CO 5h J-) ro r-\ rH Ki tn ro & Eh CO^JrHCUrHMQJrC >)Eh (D ^ ro rH g c > ,-^ ra O /J C J3 -H tn CD "Ti U ro O P rs c > rH rH ro
PAGE 228

212 "formulas are new different. These two should be ccmpared v.'ith the, distribution Golamn of the previous table. The table indicates that while there is still 3 strong redistributional effect, froiri big to small counties, the effect has bean softaned. Whereas the correlation between population and the distribution~to-cc'llections ratio was -.41 in 1963-54, it is reduced to -.36 for 1959-70. Columris III and IV of the table indicate that there is some slight increase in the correspondence between county population and county receipts from the gasoline tax. The 67-county correlation betv/een population and distribution of the second gas tax is .99, and the correlation between population and distribution of the seventh cent revenue is .95. This compares with the .94 correlatiqn between population and distribution in 1953--64. In percentage terms the table shows that the 10 counties in which 70 per cent of the state's population resides now receive slightly better than 55 per cent of the gas tax revenues distributed v/hereas they received only a little better than 40 per cent in fiscal 1953-54. In coiriparison, the 10 least populous counties (accounting for less than 1 per cent of the state's population) received a little over three per cent of the gasoline tax funds in 1969-70 where they had received almost 5 'per cent in fiscol 1963-54,.

PAGE 229

In dollar terms, the tan smallest counties received a. return of just over $3 rolllion for collecticns of about $2.2 iTtillion in 1969-70 v/hereas they had received a return of about $3 million on collections of $1,8 raillion in 1963-64. We would conclude that the legislature has modified the operation of the gasoline tax in such a way as to shift funds in the direction of the more populous counties v^ith thesir transportation needs while still effecting substantial subsidization of the poorer, less populous counties. State-lGcal, program sha r inq---educa tion Still another area cf public policy in which financing is shared between the state and local governments in Florida is education. It_ is of iiaportance to our analysis since, as noted in the Jev7ell quotation, it frequently has been associated with the reapportionment controversy. In Florida, the state became involved in financing education with the passage of the Minimum Foundation Program by the legislature in 1947. By passing this act, the state recognized that financial resources v;are unequally distributed among the counties faced with functing educational programs. The Minimum Foundation Program was designed to correct this problem. The exact formula for distributing state funds through the.MFP has been tampered with several times by the legislature and has always been rather complicated. Kover hheless,

PAGE 230

214 it alwc-ays had as its basis the relationship between demand for school services (as measured by attendance) and the ability to pay (measured by amount of revenue produced by a given millage rate assessed against non-exempt real property) Of interest to us in the context of the reapportionment controversy is the relationship between population and this distribution of funds through the MFP Table III24 shows the relationship for the ten most and ten least populous counties in fiscal years 1963-54 and 1967-68. We must note immediately the fact that there is a close correspondence between population and the distribution of these funds at both points in time. The correlation between population and percentage of i^'IF? funds is .99 for both time points. Thus> there has been no marked revision in the formula fox' distributing ^IFP funds in our study 18 period. But, there never was any serious discrimination against urban areas in the v/ay the funds were distributed. The important, changes in education which have occurred during pur. study period concerned levels of educational spending rather than distribution formulas for the funds dispensed throxigh the MFP The story revolves around the special education session of the legislature in 1963. Under threat of a teacher walkout (which eventually materialized any^Afay) the legislature raised the sales tax, increased its

PAGE 231

'^^I '*!' Table I I I -2 4 RELATIONSHIP BETWEEN STATE AID TO EDUCATION AI-
PAGE 232

216 We noted tho impact of this special session in our analysis of the regression squations when we saw Florida's change spurting ahead in 1968. Wa repeat here that we find it difficult tc attribute these changes solelYf or even primarily to the fact of reapportionment because of the pressure applied by the Florida Education Association during this period. The 1969 session, v/hile lacking the glamour and drama of the special session, was regarded by the Education Department's specialists as showing some important advances. Moreover, the specialists felt that the legislators demonstrated greater sophistication in dealing with 21 educational questions. In assessing the v/ork of the reapportioned legislature in the area of educational funding we find a mixed picture. We cannot overlook the fact that at the sam.e time the state increased its share of the funding for education, it simultaneously linaited the ability of local school districts to tax for the purposes of improving their schools. The legislature imposed a ten-mill limit on the tax rates the local county schools boards could impose to raise funds for current operations. Caught between what it perceived as a ^ nascent taxpayer's revolt and an impending teacher's walkout, the legislature gave with one hand while taking away with the other. The ten -mill limit, first imposed by

PAGE 233

217 3tat-ute, v/as v'ritten into the constitution v.-hich, the legis22 lature proposed in special session in 1968. On balance, however, it seeir.s that the legislature v/as somewhat more attuned to need for better quality education which goes along with urban society. Refu nds to local governments of s tate col lee ted taxes— cigarette tax_ and race track tax. Both the tax on cigarettes and the tax on pari-mutuel betting at race tracks have always bean shared Vvith local government units. Of the two, the di.stribution of race track funds is the one which has been associated with the reapportionment controversy in the state. For example^, H. D. Price points out v/ith dismay that race track funds "are distributed among the sixty-seven counties on an equal basisi" That this is discriminatory against large, populous counties in his mind is implied when he points out that "the race-track fund alone can account for the bigger part of a small county's general and school funds. "'""^ TVie other major tax returned in an unrestricted manner to local governmants is the cigarette tax. The cigarette tax has not had the same stigma attached to it in the minds of the reformers because it has been traditionally distributed to the cities on the ba.sis of collections. However, cigarette taxes collected in unincorporated areas went to the state.

PAGE 234

218 x.Of these t^^G funds, only tlve distribution of the cigarette tax fund iias been changed in the period since reapportioniTierit, Since the popula.tion concsntration still varies markedly froni county to covinty this means that the small counties a.re still receiving preferential treatment in the distributioh of the race track funds. The change in the cigarette tax carr.e out of the special education session in 1958 v;hen the legislature increased the total tax to 15 cents per pack. .For municipalities the tax conbinues to be returned on the basis of collections. The important change is that the county governments are now given a share in the fund (although it is a much smaller share). While the inclusion of the counties for the first time is clearly not of benefit to urban counties exclusively, it nevertheless v/ould be of assistance to those counties faced with urban type demands in unincorporated areas. §.t3te i_nf IjiStlQS. on .lo.ca]_ fiscal v i t a, 1 i t y -thg. t.en-IL4Jii limi.t. This final type of state-local fiscal relationship which we examine is not derived from the state and local funding of programs, but rather frora the state's ability to regulate and limit the fiscal powers of the local governments The constitution of the state provides local governments (cities, counties, school districts) v/ith authorization to

PAGE 235

219 levy an ad valorem or real property tax and such othar taxes as the legislature shall authorise by law,''' For these go'^'ernments the property tax is a fiscal mainstay. The most important legislative development in the period since reapportionment, concerning the property tax, suggests the involveinant of an urban interest, but not of the sort envisioned by most reformers According to the Florida Municipal League, when the legislature met in 1957, the first session after re-iapportionment, it seemed to have two comiiiitments----"no new taxes" and "ad valorem tax r-elief." Accordingly, the legislature im.posed a lim.it of 10 mills as the m^aximum rate a county or municipality could impose for its operating budget, In the 1958 education special session the limit v/as extended to county school boards. In the constitutional session, the limit was applied to all three as a part of the new constitution. 29 In evaluating this move "in relationship to reapportionment we should take three things into consideration. First, we shc'dld recognize that the imposition of this limit probably represents the implementation of a kind, of urba:n J interest (albeit not'the type envisioned by most reapportionm.ent reformers).. Wha.t the legislature seemed to be responding, to was the pinched pocketbooks of homeowners who • felc that real property had reached its limit as a source

PAGE 236

220 of tevenuQ for local governinent. In this case the urban d'A-eller a,nd. property ovv'ner as ta.xpayer seemed to override the urban dweller as service consumer. A second factor to be considered in evaluating the tax limit is the fact that it probably should be vievj-ad as part of the process which had been going on for several years aiiTied at bringing equity to the system of property taxation. This battle featured the homeowner and the large landholder as the protagonists in a battle waged in the state court systems. As the battle for fair Valuations had been won in the courts, it became apparent that this was not necessarily going to ease the tax burden on the homeowner. What is noteworthy from our viev^7point is that the process could be finished off in the more urban oriented legislature v/hich had begun in the courts. The third factor about the legislature's limitation of local fiscal power to be noted is that the legislature has bsenM.:!nv,-illing to see its at'tempts to unburden homeowners bring fiscal disaster to cities. Acting under the constitutional provisions that it would set the date for the rollback of the millage, the legislature has continued to extend the date so that cities which were above 10 mills in 30 1957 have still not been forced to reduce their mdllage. Thus, although the legislature seem.ed to respond to sviburj^:)a,n ~

PAGE 237

221 interest at tlia expense of the cities, it has been unv/illing to hold the feet of city governments to the fiscal fires so, far S-gini-nary. In this chapter we have pursued several analytical approaches to relating state fiscal policy to the factor of legislative apportionmentIn general our findings have been negative. In the compative correlation analysis, we were unable to show any high or consistent correlations bet^veen apportionment and the expenditure m^easures we examined. In the latter part of the chapter v;e looked at state-local fiscal relationships in Florida in order to explore the idea that reapportionment might have changed the state's policies in ways beneficial to local governraents. Our findings here are mixed. We saw that the gasoline tax distribution formula has been changed in such a, way as to leave more road funds available to populous counties. The state share of educational financing has increased, but as we noted there was strong pressure from the teacher's association and we would be hard pressed to defend the thesis that reapportionment had caused this change. Finally we noted that the state has hemmed in local government financing with its ten-mill ad valorem tax rate limit. In fiscal affairs, then, it seems that reapportionment has not had the effect of opening the fiscal floodgates in the manner that some thought that it might „

PAGE 238

222 Footnotes to Chapter III Malcolm E, Jevyell, "Political Patterns in Apportionrnent, in Malcolra E. Jewell (ed.), ^he Politi cs, of Reg ]?P.'^'''^ [-. iqni Tien t (Nev7 York: Atherton Press, 1962) pp. IBIS, For an alternative list sea Gordon E. Baker, Rura l vs. Ii£]?3Jl Political Power. (Garden City, N, u'.: Doubleday, 1955), pp. 19-26. "See especially Thomas Dye, Poli tics, Econ omics., ,and the Public (Chicago: Rand, McNally, 1956) and Ira Sharkansky, Spending j.n the Ameri can States (Chicago: Rand, McNally, 1968) This method is basically the same as that used by SharkanskVo For e>:a.iTiple, sea Spe nding in the American S^tates., pp,. 48-52. '^This is what Sharkansky does in part of his analysis. Ses Spending in the American States, pp. 45-47. ""Observations on the Measurement of Change, m Edgar F. Borgatta. (ed.), Socioloq-jcal Met hodo logy: 1969 (San Francisco: J'ossey-Bass, Inc., 1968), pp. 113-133. Spendi ng in tlie. Am eric an States, pp. 39-43. ''Tlie question might be raised as to why we make no reference to tests of significance. The answer is that we are dealing with the entire population, not with a sample. Significance tests are appropriate measures when the question is whether generaliz;jtion of a statistic from a sample to a population is safe. That is not the question here. The more appropriate method for judging the imiportance of coefficients in this kind of situation is by squaring them to get the coefficient of determiination. The coefficient of determination, expresses the per cent of the variance in the dependent variable explained by the independent variable (s). In the specific case under discussion, coefficients range from -.14 to -.42. We are talking, then, about explaining from 2 to 18 per cent of the variance — a modest amount at best I

PAGE 239

• ':"' 223 See for example), Herbert Jacob, ''The Conseq-aences of Malapportionment, Social £orce3., ZLIII (December, 1964) 255-261; Thomas Dye, "Malspportionaient and Public Policy in the States," Jou rnal of Politics, XXVII (August, 1965), 586-601; -Richard E, Dav/son and James A. Robinson, "Inter-Party Competition, Economic Variables and Welfare Policies in the American States," Journal of Politics XXV (May, 1953), 265-289; Richard I. Hofferbert, "The Relation Between Public Policy and Some Structural and Environmental Variables in the American States," American Pol itical Science Review, i_iX (March, 1965), 73-82; Thomas Dye, Politics, Economics and t he Publ ic ; and Ir?, Sharkansky, Spending ijn, the Airier ica.n States .. Spend in.g' in .the. Americ an States p. 64. "^For a thorough discussion of the legal and theoretical bases of state-local relations, see Anvv-ar Syed, The Political Theory o.f A-merican Local Government (New York: Random House, 1956). See especia.lly chapter three. "•'-V„ M. Newton, Jr., Crus ade for. Democracy (Ames, lov/a : Iowa State University Press, 1951), p. 122, Mr. Nev/ton was managing editor of the Tamipa Tribune when he wrote this. ^'"''See for example H, D, Price, "Florida: Politics and the 'Pork Choppers,'" in Malcolm E. Jewell (ed.), T_he Reapportion ment Revolution, pp. 81-97. ""Arthur -i-i. Cunkle, Financing Florida Government: 1969 (Tallahassee: State of Florida, 1969), p. 21. '''^'ZI-.QXJAa. Statutes 208.11; Florida Constitution of .1885, Article IX, Sec. 16. -^^Cunkls, p. 22. "^Florida Starutes, 203.44 and 208.111. ''-"^Florida Statutes, 2 36. "The 19 70 session of the legislature revised the distribution formula for the MFP, but the change has no clear urban-rural relationshio. ^^Cunkle, pp. 55-59.

PAGE 240

^-'Eldridge R. Collins and Howard vTay Friedman, "The 1969 Legislature," F 1 o r i d a S ch o o 1 s XII (September-Cctober, 1959), 22-23. 21. Ibid. 22 Florida Cons titu tion of 1969 Art. VII, Sec. 9. The Florida courts have subsequently ruled this provision in violation of the United States constitution as it applies to the school system. "•"Price, p. 89. '^ 'Ii§K§. .of. Florida 68-30. ^-Florida Co nsti tution of 1969, Art. VII, Sec. 9. ^^^Goraon T, Butler and Ralph A, Marsicano, "Legislation Affecting Florida Municipalities, Flo rida Municipa l Record. XLI (August, 1967), 3. ^"^Laws of Florida, 67-395 and 67-396. CuTikle, p, 23 '^^^lloiMa Constitution of 1969. Art. VII, Sec. 9. ^'-'See for example, Laws of F lorida 69-278 and SB 155 for bhe 1970 session.

PAGE 241

Chapter IV POLICY OUTPUT ANALYSIS — II In the last chapter we examined the relationship of fiscol matters to reapportionment. We saw that, at least as far as levels of expenditure were concerned, no real changes could be unequivocally related to reapportionment. Yet, it would be wrong to conclude on the basis of that analysis that reapportionment is unrelated to matters of policy in the states. The consequences which have commonly been associated with malapportionment are not exhausted by the concerns of fiscal matters. Neither are the significant dim.Gnsions of legislative output limited to those requiring expenditure of funds. Caudde and McCrone note that: Many significant policy dim.ensions are not susceptible to measurement in terms of expenditures. The anactm.ent of civil rights legislation in the American sta.tes is one such significant dimension. Anti-discrimination legislation in education, employinent, public accomodations, and housing can not be measured normally by the amount of money expended. Instead, these policies are, by their very nature, qualitative. In accordance with this viewpoint, then, we think it im.portant to examine some of these nonexpenditure policy output areas in relationship to apportionment. 225

PAGE 242

226 The value of examining apportionment in relationship to nonexpenditure policy outputs is further underscored by the findings of Walker in his study of the innovative tendencies of American states. While other political characteristics tended not to have independent explanatory value in relationship to his innovation scores, malapportionment did. As Walker puts it: Earlier studies, using expenditures as a measure of policy outcomes have consistently found that apportionment has little importance as an explanatory variable. Our findings indicate that apportionment does make a difference where innovation is concerned. Although the other political factors do not have great independent impact on innovation, the clear implication ... is that those states which grant their urban areas full representation in the legislature seam to adopt new ideas more rapidly, on the average, than states which discriminate against their cities. This kind of finding makes it clear that apportionment cannot be dismissed as politically unimportant (i.e., unrelated to significant dimensions of policy) and that this kind of analysis needs to be continued. In this chapter, therefore, we attempt to get at the matter of nonexpenditure policy outputs in the Florida legislature in the light of legislative reapportionment. We want to ask whether such nonexpenditure policy outputs have undergone any significant changes as a result of reapportionment.

PAGE 243

227 Ms,tl?^dol.o,g_^^l — Accepting the notion that qualitative outputs should be exarained in .relationship to political variables is much easier than actually carrying out such an analysis. The research problem amounts to asking ourselves, "What has the legislature done in the period since reapportionment that it would not or could not have done before?" The most satisfying answers to this question would come from a comparative study involving change-overtime analysis for several states. Limited financial resources ruled out the possibility of this kind of analysis, so we settled for a simpler method. Our approach was to pose the research question (through a mail questionnaire) to two groups of political leaders and rely upon their judgment as an approximate answer. Where possible we will corroborate or qualify opinions with other evidence. We sent a mail questionnaire to all incumbent members of the legislature in 1969 and to the administrative assistant to the chief executive officer (mayor or city manager) of 105 Florida cities. On the one hand, we wanted to get soma estimate of the impact of reapportionment on nonexpenditure policy outputs from the legislators themselves. On the other hand we wanted to sample the opinions of officials directly concerned with a type of output (miunicipal regulation) which has traditionally been associated with

PAGE 244

2 2 3 reapportionment. Malcolm Jewell, in the quotation with which we initiated our discussion of policy and apportion-nient in the last chapter, mentioned such things as "home rule legislation and laws to deal with the specific problems of the larger cities .' as examples of output having consequences for the city which bear an apparent relationship to malapportionment. We thought it important, therefore, to find out whether these city officials (as important consumers of nonexpenditure policies) perceived any significant changes in the post-reapportionment period. Using both groups (i.e., legislators and city officials) also offered us some opportunities to examine differences in the perceptions relating to the effect. of reapportionment. Fin dinqs---Of the 105 questionnaires sent to cities, 52 v;ere returned to us completed. Of the 119 members of the House of Representatives 36 completed questionnaires and returned them, while 10 of the 48 Senators returned the questionnaires completed. The coiTibined legislative sample came to 45 respondents out of a possible 167. We decided that these two groups were large enough to provide a reliable basis for cur analysis. We shall report the findings from the legislative group first, and then turn to the cities. One of our questions directed to legislators was designed to get a general estimate of the impact of

PAGE 245

'29 reapportionraant. on .Legislative opera-cion;.: 1 ,•=, f1 ... 5 The first pari of the question asked each ?i.egislator to choose a statement approximating his feelings about the importance of reapportionment. He could indicate (1) that there had been "many important changes" as a result of reapportionment, (2) that reapportionm>ent had made some difference, but mostly the legislature operated as before, or (3) that reapportionment had made no difference at all in legislative operations. Table IV-1 DEGREE AtTD TYPES OF CflZlNGS IN LEGISLATIVE OPERATIONS PERCEIVED BY FLORIDA LEGISLATORS AS RESULTS OF REAPPORTIONMENT Degree of Change — ..,. .. ,_. N Per Cent Many important changes 38 82.5 Some change, but mostly same as before 7 15.2 No difference at all 1 2.2 Types of Change N Per Cent Legislation passed 13 28.3 Style of legislative operations 21 45.7 Urban representation 13 28.3 No change specified 8 17.4 (N245) As seen in Table IV-1 the Florida, legislators believe in overv;heIm.ing proportions that reapportionment has resulted

PAGE 246

in change in legislative operations. Cver 80 per cent of fhem indicated that reapportionrnant had raade many important changes in legislative operations. Of interest equal to the overall judgment of reapportionment effects was the indication by the legislators of the wajys, in v,'hich they felt legislative operations had been altered. We asked those who had indicated some perceived change to specify some changes that, they saw as resulting from reapportiomnent. The responses to this question fell into three categories (see Table IV-l) Some (about 28 per cent} pointed to specific pieces of legislation passed as evidence of change. Forty-five per cent listed what we have called style changes. That is, they indicated that the quality of legislators had improved, that the legislature V7as more efficient or effective in its operations, or in some other way indicated improvement in the way the legislature v;ent about its business. Another group (28.3 per cent) described the perceived change in terms of some aspect of increased ur-ban representation. They either cited increased nurribers of urban legislators or increased awareness of urban interests. Perhaps the most striking feature of the descriptions of change given is the frequency with v;hich stylistic rather than substantive axam^ples were given. Legislators did

PAGE 247

2 3 J indeed perceive change as resulting from reapportiomnent But, when pressed to specify the nature of that change at least 45 per cent of theiri turned to stylistic matters. Perhaps the wording of the question triggered this pattern of response since it asXed about "changes in the legislature," and v/e, therefore, do not wish to build too large a case on the finding. Nevertheless, when we consider the fact that as many as 28 per cent of our respondents did not perceive the question to iraply only stylistic raatters the point does not appear insignificant. There seems to be general concurrence v/ith the legislators' belief that the style of legislative operations had significantly improved in the post-reapportionment period. In February, 1968 the legislature commissioned the Center for State Legislative Research and Service of the Eagleton Institute of Politics to malce a study and recommendations for reform. The report notes that by the time the study got underway reform had already begun. The study details important institutional reforms accomplished during our study period: Among the many substantial institutional changes effected in 1967-59 were the abolition of the Legislative Council, the creation of the Joint Legislative Management Committee, and the establishment of permanent standing committees with interim responsibilities. A new Code of Ethics for members of the legislature was adopted, as well as new regulations governing the activities of lobbyists. Statutory

PAGE 248

f 232 revision, bill drafting and other legal services were transferred from the Attorney General's office to the Legislative Reference Bureau. The office of the .Legislative Auditor v/as created and, in addition to conducting traditional coitipliance audits, was empowered to conduct performance audits of executive programs. A legislative intern program for graduate students in political science, law, and related fields was instituted. The use of electronic data processing during the session was expanded. The use of fiscal notes wa.s begun in the House of Representatives, and by a new bill printing procedure was adopted by both houses. Above all else, the legislature supervised a substantial re'v'ision of Florida's Constitution which included, among' other significant features, provision for annual sessions of the legislature, the exclusion of limitations on legislative compensation, and a direction for reorganization of the sprawling executive structure into twenty-five or fewer departments. As if this v;ere not enough, the legislature also approved staff assistance for individual members of the House and Senate, started pre-session orientation conferences, and conducted a budgetary school for legislators in late IQSS."^ This is quite an impressive list of stylistic changes in the Florida legislature. To this we can add several additional pieces of testimony to the revamping of legislative procedure. In June, 1970 the Council of State Governments recognized tlie Florida legislature as the nation's most progressive during the preceding two years. At the same time it awarded the Louis Brownlow Memorial Fund prize to the Speaker of the 1970 session of the Florida. House for his article outlining the legislature's modernization efforts." In February, 1971 the Citizens' Conference on State Legislatures released a report in which it ranked the Florida legislature the fourth

PAGE 249

best overall." Finally, it should be noted that a st\3dy of expenditures for legislative operations showed that Florida ranked fifth in the nation on the basis of average 10 expenditure per legislator in the fiscal years 1965 and 195 7. All of this evidence is corroborative of the general legislative perception of important stylistic changes in the operation of the Florida legislature during the period since reapportionment. Our questionnaire offered legislative respondents two opportunities to indicate the kinds of policy achievements which they regarded as noteworthy in the first two sessions after reapportionment. In the first instance, we simply asked the legislators to list what they considered to be the most important bills passed in each of those sessions. T'-vo pieces of legislative action received almost unanimous nomination: constitutional revision and executive reorganization. Ho other measures received more than a few nominations Consti tational revision was accomplished by the legislature in a special session in the summer of 1968. Constitutional revision had been a lingering issue in Florida politics for some time. The first steps towards proposing a document came in the 1965 session of the legislature when a constitutional revision coinmittee was

PAGE 250

appointed. Constitutional Revision certainly v;ould have come to the legislative floor without reapportionment. In fact, as we noted above, the legislature was already in a special session to deal with constitutional revision in 1957 when it was declared unconstitutional and ultimately reconstituted. After the interruption of the 1957 session, it was not until the summer of 1968 that the legislature could get to the matter again. In a nine day session it worked out a proposal xvhich it submitted to the people in November, 1968, The constitution took effect in January, 1969. It was the constitutional revision which provided the impetus for executive reorganization. The constitution provided (Art. IV, Sect. 5) that the executive had to be reorganized into not more than twenty-five departments. Thus, in the 1969 session of the legislature an important item of business was the implementation of this proposal. Since the legislators had chosen these two enactments as their outstanding accomplishments of the post reapportionment period, we wanted to know if they thought reapportionment had pla.yed a role in their passage. As it happened, our questionnaire included a structured question concerning these measures in which respondents were asked to rank-order several factors in terras of their importance 12 in influencing 'che. legislative action in these two areas.

PAGE 251

^SO Table IV-2 LEGISLATIVE PERCEPTION OF FACTORS IKFLQENCING PASSAGE OF CONSTITtFTIONAL REVISION AND EXECUTIWJ REORG?iN"ZATIOS Tvo -Party Governor s Passed in Other Ra nk Reappor •tioniTiant S y s tera Leadership Other St-ates Reasons 1 2 3 4 N 28 11 6 % 58.2 24.4 13.3 0.0 0.0 10 10 6 % 24,4 36.6 24.4 14.6 0.0 N % 1 2.8 1 2.8 11 30.6 16 44.4 7 19.4 N % 0.0 6 20.7 11 37.9 10 34.5 2 6.9 N % 9 45.0 7 35.0 4 20.0 0.0 0.0 (N) 45 100.0 41 100.0 35 100.0 29 100.0 20 100.0 As can be seen in Table IV-2, there are two indications that the legislators saw reapportioninent as being important in these irieasnres. First, it was included in more of the responses than any of the other measures, vrnile many of the legislators chose to include only two or three factors in their rankings, alm.ost all included reapportionment. Second, it had the highest mean ranking and was given a first place ranking by 68 per cent of those ranking it. Of the other factors, only "other reasons" had a large percentage v/ho ranked it in first place. In that case the sraail N suggests a small group of legislators who felt strongly about some particular factor, but whose enthusiasm was not widely shared among the legislators. We should also note that there was no consensus among the legislators on the "other reasons" listed.

PAGE 252

236 'I'hs secorici type of opportunity given legislators to 'i indicate policy achievenaents xn the post-reapportioninent period asked for a specific typa of policy~--legisla tion affecting the cities. Since reformers had emphasized that raalappor't-ionment deprived cities of authority and power they needed to meet their needs, we thought it worthwhile to ask legislators v;hat they had done concerning urban problems. As can be seen from Table IV-3 the legislators overwhelmingly believe: that they have been more attentive to the problems of the cities since reapportionment than before. Moreover, their rankings of possible factors influencing urban legislation indicates that they think reapportionment vvas the most, important factor in producing this effect. Of those specifying particular types of things done for cities in the first two sessions after reapportionment, the responses were divided between authority grants and fiscal aid. From the data presented in the previous chapter and the data to be presented about city. perceptions of legislative action, we might inject the notion that the fiscal aid was more illusory than substantia.!. The responses from the legislators indicate that they perceived the f^ct of reapportionment to be significant. Their overall judgmtents expressed it;' moreover, they related reappcrtiomnent to significant policy outputs in the

PAGE 253

Table rv~3 LEGISLATIVE PERCS??TION OF POST-REAPPORTIONM.ENT RESPONSE TO UKS^JNI PROBLEMS AI
PAGE 254

138 TABLE IV-4 CITY ADyilNISTRATOR PERCEPTION OF POST-REAPPORTIONMEIsrr RESPONSE TO U^REAl? PROBLEMS AND FACTORS AFFECTING RESPONSE Over a 1 1 Judqinen t or" Legislative Response "In general, do you think the legislature in the 1967 and 1959 sessions was more or less attentive to the needs and problems of the cities than it was in the 1965 and earlier sessions?" N % More attentive About the same as beforeLess attentive 8 17 25 50 16.0 34.0 50.0 .00.0 Factors Influential in Legislature's Dealings with the City Two-party Governor s Action in Other Rank Reapportionment System Leadership Other States Reasons N /o N % N % N % N % 1 21 53.8 8 21.1 2 6,1 18 75.0 2 14 35.9 16 42.1 5 15.1 3 12,5 3 3 7.7 11 28.9 15 45.5 3 12.5 4 1 2.6 3 7.9 11 33.3 0.0 39 100,0 38 100.0 33 100,0 24 100.0 Areas in which the Legislature Failed to Meet City Needs N % Authority gra/nts Fiscal assistance Other 6 24.0 24 96.0 2 8.0 (N=25) to our questionnaire. As can be seen in Table IV-4, some startling differences appear. Whereas the legislators gave

PAGE 255

O-Jf themselves credit for being more attentive to urban needs, many city administrators saw them as baing the same as before and a majority saw them as being even less attentive than before reapportionment. City administrators did give reapportionment an important place in affecting the outcome of urban legislation, but this finding is difficult to interpret in relationship to their negative evaluation of legislative efforts in the urban area. One clue might be derived from comments written in the margins of a few of the questionnaires. These comments indicated that the legislature was suburban rather than urban in its orientation. The reconciliation between the findings may be in the fact that city administrators perceive reapportionment to have produced many suburban legislators who fail to understand urban problems. While we have no basis from our data to estimate the generality of this feeling that legislators v;ere suburban-oriented, one Florida newspaper ran a story which indicated that this point of view had wide currency — and that it had some truth to it. The newspaper studied the delegations from populous Dade, Broward, and Palm Beach ^ counties and found "that most of the lawmakers do live in green, tranquil, upper-income subdivisions some distance. '' from the turmoil and clutter of the urban core." Nevertheless, the author of the story was unwilling to come to the

PAGE 256

240 quick conclusion that there was a line of causality betv;een this fact of suburban residence ?ind any perceived failure of -, 13 legislators to meet urban needs Another aspect of the city administrators' -parceptions v/hich deserves mention concerns the perceived areas of legislative failure. The striking thing here is that virtually all of those city admihistrators v7ho indicated that the legislature had been less attentive since reapportionment cited some form, of failure in the area of state fiscal assistance to the cities. The specific coiiaments written in 1 /I indicate that the ten mill limit on ad valorem, taxes"" was a major blovv' to the morale of urban administrators. The negative evaluation gf legislative performance seems to be fairly directly related to this action, which cities viewed as impairing their ability to raise funds to meet their own needs We have some other evidence which suggests that the fiscal limits imposed on the cities so overshadowed other actions of the legislature with regard to cities as to becloud city administrators' perceptions of legislative actions. We can, for example, get a third estimate of the degree of legislative attentiveness by looking at the numHoer of urban oriented bills introduced into the legislature in sessions befcrG and after reapportionmerit

PAGE 257

*.Table IV-S SUmmRY OF FLORIDA LEAGUE OF MUNICIPALITIES LEGISLATIVE ACTIVITIES, 1953-1969* Per Cent of bills selected by title fcr full reading Per Cent of bills deemed iraportant enough to follow thrcugh legislature Per Cent of bills significant enough to be mailed to cities Total number of bills introduced in both houses 1965 7,5 3.5 1.5 3958 7.3 3.5 2.2 4435 1967 11.9 4.8 4.2 5181 1959 13.8 7.4 3.6 4784 *So-arGet Derived from legislative reports of Florida Kunieipal League. See the following citations in Florid a Mimiciual Reaord, XXXVII (July, 1963), 16; XXXIX (July, 1965), 5; "XLI'*'! August, 1967), 8; XLIII (July, 1969), 3. Table IV--4 suggests that there was an increase in legislative attention to urban concerns during the two sessions after reapportionment. Using the Florida Municipal's League's' concern as the criterion for urban subject matter, we can see 3 qrov/ing percentage of the bills dealing with urban affairs from one session to the next with the break between the percentages for 1965 and those for 1957 demonstrating a si2:able jump. On all three measures of League concern

PAGE 258

(urban relatadnesc) fchiE jump pattern occurs. This suggests that city administrators' conclusiori of lack of attentiveness involves soi'ne selective perceptual processes. Nevertheless., the argument might be raised that although raore urban oriented bills were introduced after legislative reapportionment that gloom among city admitiistrators might be justified in terms of the outcomes. In other v/ords, more bills do not necessarily m^ean better times for cities. Fare ag.-3in v;e v/iil rely on the Florida Municipal League's legislative reports to try to clarify the relationship of the reapportioned legislature to urban policy. The Florida Municipal League's evaluation of the 1963 session of the legislature was m.ildly favorable. The most imc-crtant piece of legislation affecting urban affairs to emerge from that session was a bill increasing the cigarette tax and providing that revenue generated within corporate limits from the additional tax would be returned to the cities. The rem.ainder would go into the general revenue fund. Two non-revenue measures which the League had hoped to gat passed received a chilly reception from the legislature. These included a measure providing for easier c 'annexation procedures and a proposed constitutional amendment allov.'ing cities and counties to unify certain functions. In fact, with regard to the annexation procedures the League

PAGE 259

243 leaders reported chat "... it took the full efforts of tne Leacrae to retain what had bean secured in the last session.""" Thus, although the League was pleased with the financial assistance expended through the cigarette taxes, it v>3s disappointed in the legislature's unwillingness to extend greater authority to municipalities to solve their ov/n problems. The Leagu.e of Municipalities found the 1965 session of the legislature no friendlier towards its program to broaden municipal authority. It tried once more to gat relaxed annexation measures passed. Concerning this the leaders reported: To state that the Coiiimittee on General Legislation was not particularly interested in the passage of. any legislation affecting annexation would be a gross understatement. After more than one hour of hearing the annexation proposals, the comn-iittee rapidly voted all of them down in defeat; Thus rebuffed, the League turned primarily to a defensive strategy in the 1965 session. The leaders reported that "probably [their] best legislative success was in opposition to harmful to municipalities and amending some legislation to relieve the full impact on the cities.' Clearly, the League did not see the 1965 session as being receptive to urban oriented legislation. In the 1957 session of the legislature financial questions were dominant as they had been in 1963 The

PAGE 260

244 Municitm"! Lea.gue leaders felt that the two slogans best catohing tne^ rri6od, of the legislature both in its reg-ular r .and special se,3sions were "no new taxes'' and "ad valorem t3x relief." -he iraposition of the ten-inill limit in response to the tax relief imperative has already been •nentioned. This was itiade more demoralizing to city officials when an increase in cigarette taxes which would have provided soiae replacement. monies was vetoed by the governor. And, to vaake matters worse, the legislature did not deal with several non-financial proposals affecting urban life. Among those listed by Lea>gue officials as left undone at the end of the 1967 session were proposals for comprehensive planning and zoning authority in counties, authorization for urban 18 renewal projects, and home rule for municipalities. Thus, the first session of the legislature after reapportionment left city interests disappointed and demoralized because of its determination to reduce property taxation and the failure to provide replaceiaent money. This seems to be the thing which is picked up in our questionnaire. City administrators perceived the legislators after reapportionment as not being any more attentive to their problems than before, and they overwhelmingly cited fiscal failures. The actions of the 1967 session seem to have set the responses to our questionnaire.

PAGE 261

245 The Municipal League liarl a. m.uc'h more favorable evaluation of the 1959 sassion of the legislature. The dominant theme o,f their report on this session vjas concerned with grantin':? of authority to Tmanicipa.lities Ainong the grants v/ere home rule, urban renevval authority (with referendum) authorization of sale of industrial bonds, authority for unification of tax a:ssessing and collacting procedures for cities and counties, and simplified procedures for amending 'city charters. It was also noted that for the first tiiiie a bill for what the Leag^ie considered more realistic annexation procedures managed, to get reported out of committee in theHouse. At the same time that substantial grants of power were being made to the cities only limited attention was given to fiscal needs. The legislature provided no new source of revenue for cities in the 1969 session, but did postpone the date for the imposition of the ten-mill limit. In general, while no fiscal gains 'were made we would have to conclude that the 1969 session of the legislature showed considerable concern for municipal problems and extended grants of power to assist municipalities in solving those problems. We would re-emphasize the fact that the failure to raise taxes and replace money lost to ad valorem tax relief seem.ed to be the fact about the post-reapportionment legislative

PAGE 262

246 sessions whicn. stuck in the heads of our city respondents. The fiscal failure of the legislature seamed to blot out all of the authority grants made in the 1969 session v;ith the result that legislators received sn unfavorable evaluation from city .administrators. We are saying, in short, that it would be misleading to take this negative evaluation at f^ce value. 'There is one additional bit of evidence which supports our contention that the legislature made substantial delega' tions of authority to the cities after reapportionment although .^it did not give the desired financial help. This concerns the use of local 'legislation to deal with municipal problems rathar, than allowing local governments to handle their own 'concerns within legislativa guidelines. The use of local legislation was quite extensive in Florida in earlier periods. Vve asked our city respondents to indicate, for the period before reapportionment and for the period after reapportionment, whether the legislature had dealt with their problems mostly by general law or by local law. Only a little over 2 2 per cent of the respondents indicated that the legislature dealt ^rfith most of their problems by general law in 1955 and earlier, while about 49 per cent of them indicated that most of their problems had been deslt v;ith by general law in 1957 and after. This shift in the direction of greater use of

PAGE 263

247 general legislation rapresented a geri63ral effort on the part of the legislature to extend greater power to cities to control their ov/n de3uinies--to solve their own problems. SuininarY — -We riava attempted to at least catch the flavor of nonexpenditare policy changes in the reapportioned Florida, legislature by. use of questionnaire data. Data from the legislative questionnaire indicated that legislators thought reapportionment was instrumental in the passage of what they considered the two most important pieces of legislation in the post-reapportionrnent period — constitutional revision and executive reorganization. Legislators also saw themselves as being mora. responsive in an area of policy traditionally associated with the reapportionment controversy, i.e., urban policy, as a result of reapportionment. Finally legislators saw reapporticniTient as resulting in improvements in the way the legislature goes about its business — a judgment which we saw to be shared by less biased groups. Our city questionnaire offered an interesting contrast since city administrators were less eiithusiastic about the reapportioned legislature. A majority of the city administrators judged the legislature as being less attentive in the post-reapportionrnent period. We noted, however, that this negative evaluation seemed to rely exclusively on the fiscal failures of the legislature and to ignore the

PAGE 264

J.'-i'0 substa,nt.ial gains made by cities in authority to act. The municipal pressure group, the Florida Municipal League, seemed to be quite pleased with these authority grants made by the 1959 session of the legislature. Overall, it would seem, that our two chapters on policy analysis accomplish two things. First we indicate while the fiscal floodgates have not been opened in Florida by reapportionment (as they have not in other states) some re-weighting of interasts can occur within, constraints of existing revanue3--e. g. shifting second-gas tax money in urban directions. Second, we give some evidence that reapportionment in Florida was felt by some members of the states' political elite to have made possible changes in nonexpenditure policy areas which were of significance.

PAGE 265

249 Footnotes to Chapter IV -Dona.ld J. McCrone and Charles F. Cnudde, "On Measuring Public Policy," in Robert E. Crev/, Jr. (edj, State Poli tics; Readings on Political Be havior (Belraont, Calif,: Wadsworth, 1968), pp 523--530, The quotation is from pp. 523-524. ^Jaclc L. Walker, "The Diffusion of Innovations Among the American States," Am eri can Political Sci ence Rev.iew, LXIII (September, 1959), 880-899. The quotation here is on p. 887. ^The list was composed of all cities listed with a population of 7000 or over in Officials o^f Florida Municipalities (Jac]^sonville, Fla.: Florida League of Municipalities, 1970) The figure 7000, while admittedly arbitrary seemed to represent a m.inimal level of urbanization and served to produce a sample near our goal of 100 cities. '^Malcolm E. Jev^ell, "Political Patterns in Apportionment," in Malcolm S. Jewell (ed.) The_ P.ol,itlcs .of Reapportionment (Nevv' York: Atherton, Press, 1962), p. 18. ^The exact wording of the question was as follows: "A lot has been written about what would happen to state legislatures if they were reapportioned on the basis of population. In general, which of the follov/ing statements comes closest to your position regarding the effect of reapportionm-ent on the Florida legislature? Reapportionment has made many important changes in the v/ay the Florida legislature operates. ^Reapportionment has made some difference, but mostly the Florida legislature operates the same as before. Reapportionment has made no difference at all in the way the Florida legislature operates." ^C. Lynwood Smith, Jr., gtrencfthening the Florida Le_cj;islature (New Brunswick, N. J.: Eagleton Institute of Politics, 1959), p. v. •7 Ibid. pp. 3-5. ^§.t^ Petersburg Times, June 25, 1970, p. 1-B. The article referred to in the news story is Fred Schultz "Legislative Modernization — the Florida Experiment," Sjtate Gov ern m en t, XLII (Autumn, 1959), 256-259.

PAGE 266

^Michael Richardson, "Legislature of Florida 4th in Nation," S_t^ Peter sbi^rjo; TijTie,3 (February 4, 1971), p. 1~A. •^ ^National Municipal Laague, State h^qislatxxTes, Pr o^ r es s Re por t er. See Table 2, supplement to July-A-'jgust, 19S9, issue. ^^Por a brief coverage of the baclcground of constitutional revision and the nature of the nev^r document see Maiiining J, Dauer, et al., Shoul d Flo rida Adi3.2.t the Proposed ^^58 Co nst itution? (Gainesville: University of Florida Public Administration Clearing Service, 1958) Studies in Public AdiTdnistration ^o. 31. l^Tv^Q exact vjording of the question is as follows: "T-rfo things which many people regard as significant actions of recent legislative sessions are executive reorganization and constitutional revision. Please rank the following factors in the order of their significance in influencing the legislature's action in these two areas. (Let "1" represent the most significant factor.) Legislative reapportionment Grov/th of the tv/o party, system ^Leadership of the governor Fact that other states recently took similar action ^^_Other (Please specify) ^^Juanita Greene, "Did Reapportionment Help Cities," ThPMiami Heral_d (January 3, 1971), 1~C. ^^^The ten-mill limit is discussed above in chapter four. ^"^This summary of the 1963 session draws upon Gordon T, Butler and Ralph Marsicano, "Report of the 1953 Legislature as Affecting Municipalities, ". Florida Miinic_ipal Record, XIKXVII (July, 1963), 3, 15-17, 24. The quotation is from p 1 5 ^^Gordon T. Butler and Ralph Marsicano, "Report of the 1965 Legislature as Affecting Florida Municipalities," ZiQiJ.^ ilillli'S.iP-^ ^5^5i2£^' ^^'^^^ (July, 1965), 3, 5, 15, 23, 27, 31. The quotation is on p. 3. -'-''Ibid., po. 5 and 15. \ \ \

PAGE 267

251 1 P. Gordon T. Butler and Ralph A. Marsicano, "Legislation Affecting Florida Municipalities," Florida Municipal Reg_qrd._y XLI (August, 19G7), 3-4, 8-9. l^Gordon T. Butler and Ralph A, Marsicano, "Report of the 1969 Legislature As Affecting Municipalities," Florida Municipal Record, XLIII (July, 1969), 3, 5-7.

PAGE 268

Chapter V SUJyvIARY AND CONCLUSIONS In this chapter we want to do three things. First, vie. shall review briefly what we have done and what our research findings were. Second, v/e want to state some conclusions concerning the relationship of reapportionment to policy outcomes in the state of Florida. Third, we want to relate what we have done to the body of existing literature in policy outcomes analysis, drawing out some of the implications and suggesting areas for future research. .§iii™MJ_Y. — ^"^'Q began by exploring the nature of cleavage in the Florida legislature to lay a foundation for policy analysis. Wa anticipated that reapportionment might be m.ost likely to affect policy outcomes where a pattern of urbanrural cleavage characterized the political divisions of the state. Using roll call analysis techniques we discovered that urban-rural cleavage was important to voting patterns for three of the four sessions v.'e analyzed. The urbanrural cleavage pattern was overshadowed by the rise of partisan cleavage in the last two sessions of the legislature. We also examined the coinmittae assignments of 252

PAGE 269

^D:^ legislators for possible rela.tionships re urban-rural cleavage. We could not show any evidence of discrimination .. against urban legislators either in the nuraber of co.Tiinittee atisignments received or in access to influential coramittee assignments. We noted, hov/ever, that urban-rural differences did have an impact on committees simply because rural members were over-represented. Having established that urban-rural differences had j. characteristically been of importance in the Florida legislature we v;ent on to examine policy outputs. If urban-rural differences did matter, then reapportionment's alteration of the urban-rural balance might produce a change in policy outcomes as the minority became the majority. We used a two-pronged approach to relating policy-outcom.a to reapportionment. One approach related expenditure policies during the period of our study to reapportionment and several other political variables. The part of the analysis dealing with levels of expenditure supported earlier conclusions that reapportio2im.ent was not likely to make miich difference. However, we also included in our expenditure analysis an examination of the distribution of funds within tv/o state programs among the more and less urbanized countiesWe found in one which reform.ers had always believed related to reapDor tionm.ant — the distribution of road funds — that there

PAGE 270

v7as 3 significant change, in the distribution fo.rrnula in the post-reapportionraent period. Significantly, the shift was favorable to the urban counties. In the other program we examined~"-education — we found no real relationship to urbanx'ural differences in the distribution of funds, either before or after reapportioninent The other prong of our output analysis was concerned v.'ith nonexpenditure types of policies. ^ In this section we relied upon perceptions of governmental officials (gathered by questionnaire)as an estimate of the impact of reapportionment on policy outputs. We sought the perceptions of legislators as producers of these policies and of city officials as subject toan area of pol icy-male ing commonly thought to be associated with the urban-rural cleavage. -The legislators believed (1) that much had occurred in the legislature as a result of reapportionment, (2) that they were more attentive to the needs of the urban areas in the postreapportionment period, and (3) that reapportionment v/as significantly related to the two major legislative achievements of the post-reapportionment period—constitutional revision and executive reorganization. The municipal officials, on the other hand, felt that legislators were less ar.tentive to their needs in the post-reapporrionment period, 'This negative view of the legislature apparently

PAGE 271

. '. #> derived frora the legislature's failure to live up to the expectation of rcamicipal officials in rcioney matters. City officials did indicate that their problems v;ere handled more by general legislation in the post-reapportionment period. This response, along with the Florida League of Municipalities indications of the legislature's scticns v/are such as to significantly broaden the authority of cities to deal with their ov^^n problems, suggests a less entangling relationship between the legislatare and the cities in the post reapportionment period. Conclusions — ^With this review of findings before us, v';e need to see v/hat major conclusions we should draw. As we saw in chapter one, the question of the political significance of reapportionment has frequently been called into existence. We need, therefore, to address ourselves to tha.t question in the light of our findings reported here. We cannot conclude that reapportionment is of no significance to policy outcomes. While our analysis of levels of expenditure offered no reversals of earlier findings, we did find change in the within-state distr ibu tion of some funds (road funds) Moreover, legislative perceptions and Florida Municipal League opinion support the idea that the legislature moved significantly to change the state's relationship to the cities in the post-reapportionment period— i.e., to

PAGE 272

qiva theiTi greater rights of local self-government than before reapportioniTient. Rather than suggesting that reapportionment is of no 'political significance, our findings suggest that change does occur — but not in amounts of money spent. I mpl ications and Suggestions f.o.r. Fur ther Re s ea r ch-W j th this conclusion of the report of our research, ws can now turn to the task of relating it to the other research in policy-outcomes now being carried out. The relationships and implications which we v/ould emphasize are related to comments on this body of literature made by Jacob and Vines. Reflecting on the developments in the field of comparative state politics which had taken place between the first and second editions of their textbook, Jacob and Vines found one major source of anxiety. Their concern was with the emergence of what they called a, new orthodoxy that asserted that political processes had little power to explain the level of expenditures and outputs of state governments." They continue with the com,m.ent that, "It is only a slight parody of these findings to assert that politics makes little difference in the governance of the Ainerican states." Jacob and Vines characterize this conception of state politics as "outrageous." They suggest several reasons for its emergence: "inadequate measurement of concepts, the utilization of inappropriate techniques of analysis, and

PAGE 273

overiy narrow definitions of what constitutes politically relevant outputs and outcomes.' We wa.nt to examine two of these deficiencies which seem relevant to what we have attempted here. The first of the criticisms which seems germa.ne to our concarns involves inappropriate use of analytic models. In explaining this criticism, Jacob and Vine note that: The statistical models v/hichhave been used by most investigators presume that the variables that are entered into regression equations are conceptually related. That is generally not true of current research. We relate 1970 data on income with 1970 data on expenditures even though we ]<:now that there is a. lag between the demands generated by the inflation or depression of 1970 and the political decisions made on account of those conditions. Few scholars would deny the existence of such time l£!.gs But we do not know their exact dimensions and therefore are hampered in utilizing our hunches about them in formal analysis. We do not know whether it is more appropriate to enter 1957, 1968, or 1969 indicators of demand into a regression equation seeking to explain 1970 levels of expenditures. In short; the problem is that most contemporary policyoutcomss research fails to satisfactorily incorporate a. time dimension into the analytical scheme. Most previous research in the area of reapportionment has been siobject to this time dimension criticism. With few exceptions policy output studies have depended upon cross-sectional correlational designs to study politics and

PAGE 274

'"* policy,. They study the relationship between various kinds of variables and policy outcomes by studying the covariation of these variables across units of analysis (the states). Variability across units is substituted for variability across time. The negative conclusions concerning the potential effects of reapportionment came from studies which really demonstrated only that, at any given point in time, there was no independent relationship between various kinds of expenditures and the degree of malapportionment which characterized the state's legislature. Implicit is the assumption that comparing the expenditure patterns of states vmich are equitably apportioned with the patterns of those v/hich are less equitably v^pportioned is equivalent to comparing expenditure patterns of states before and after reapportionment. That is a sizeable assumption and one that we v/ould certainly challenge. Looking at states which have gone through the process of reapportionment, with its likely disruption of existing leadership groups, is likely to be quite different from comparing more or less malapportioned states. Although our study does not show any marked change in expenditure patterns, we might argue that taking a longer look to allow the changes to take hold might produce m.ore. We would certainly concur in Jacob and Vines assertion that time lags in political change need to be explored m.ore

PAGE 275

259 thoroughly. This is true not only in the study of reapportionment, but in tho study of otlier political varia,bles as well. The other of Jacob and Vines' criticisms of policy output research which seems to be related to our efforts deals with the overly narrow conceptualizations of outputs. CoiTimenting on this problem, they call attention to the fact that "most analysis has concentrated on explaining the level of expenditure in the fifty states and accounting for differences in those levels," They feel that there are at least two things wrong with this over-emphasis on levels of expenditure. First, important aspects of expenditure other than absolute l evels of expenditure are neglected. To illustrate the point they comment that: Ti'in.iich counties get highway funds is perhaps more controversial than how much money shall be put in the highway f und r which comjnunities will benefit from state expenditures in education is more controversial than the total amount the state spends on education. Such distributional questions have alv/ays been the center of the political scientist's attention. They must be reasserted in the study of state politics.-^ In addition to the aspects of expenditure which have been neglected, output analysis has overly narrowed the range of analysis by neglecting nonexpenditure measures. Jacob and Vines note that "... the output of the states in regulating private life through statutes on divorce, abortion, occupational

PAGE 276

260 licensing, and the work of tha courts is not captixred by expenditure data but these activities a,re an iraportant part of the states' political activities."^ We agree v/ith Jacob and Vines in this criticism and have iriade some small steps towards overcoming it in this study. We incorporated analysis of distribution of state jiunds along with ana,lysis of levels of expenditure. Significantly v.'a found a relationship betv/een reapportionment and the distribution of state road funds among the counties of the sta.te of Florida, We also took some small steps towards examining nonexpenditure types of policies. While our method~--depending upon a sampling of elite perceptions~-has its shortcomings we were able to give at least some attention to nonexpenditure policy by resorting to this simple method. It revealed that members of the political elites reaardad reappor tionmsnt as being significantly related to certain policy advances made in the post-reapportionment periods In looking towards the future of policy output research, we would emphasize the need to push further in the directions v;hich we have begun here. We need to genuinely examine change over time rather than facilely assuming that inferences about change can be made from the study of statistical covariation across units. Sha.rka.nsky has m.ade

PAGE 277

efforts in this direction, yet his research falls subject to I the second problem. That is, he uses only rr.easures of expenditures for his output variables. The efforts of t Walker and McCrone and Cnudde need to be emulated in the area of policy outputs. The difficult task of formulating measures of nonexpenditure outputs is of crucial importance. We share the view of Jacob and Vines that census data will no longer do. They remark that: Cens\as data, because they were readily available, have been used for the preliminary analyses of the last decade. These indicators have now been exploited to their full extent. Having shown that they can use sophisticated data, analysis techniques, the students of state politics must now obtain data more_ relevant to the concepts they are using. They must collect data that are not routinely solicited by government agencies as well as recover data that are collected but not universally published. Now that these easily quantifiable data, have been fully utilized, the process of index construction to get at the more difficult concepts m.ust be called into play. While the task is hard, it is essential to any fuller understanding of the daterrninants of public policy.

PAGE 278

/D-d Footnotes to Chapter V -^Herbert Jacob and Kenneth N. Vines (eds.), liO-LlfeLos. i.il the smexican Stae.s, 2nd Edition (Boston: Little, Brown, 1971) The coinments hei:e are drawn from the "Epilogue, pp. 556-562 : — 2 Ibid. p. 559 "Ibid. p. 560. ^IJ?M., pp. 550-561. ^Ibid. p. 559.

PAGE 279

BIBLIOGPAPffiT X Books and Pamphlets Anderson, Lee F, et al.. Le^is lax JA^e, Roll CaJ^l a.rT^YMs. Evanston, 111.: NorthweBtern University Press, 1966. Baker, Gordon E. RuxaJY.s Urban Poli tical PowerGarden City, N. J.: Doubleday, 1955. Blalock, Hubert M., Jr. gjjlJjLaJ^ IgJ.g.I.glig.g.. ^'^ No nexperime ntal Research. Chapel Hill: university of North Caroline Press, 1954. Social Stati stics New York: McGraw-Hill, j.y 60 Brady, David and Edmonds Douglas. The Effects c^ Mala^sor~ tionment on ?o2^.icv Output in the Am eri can States Iowa Cxt^'r""Universxty of Iowa Laboratory for Political Research,1966. Bnch3r,aR, William. lieqislaMve ZM^isanship: The Deviant Cas^e of. Calif orrila. Berkeley: University of California Preiss, 1S63. Cunkle, Arthur L. ZiiLancijig: Fj^QiLi-dia, G^^^^ 196^. Tallahassee: State of Florida, 1969. Dauer, Mannina J,, et al. jahc^vad Fj^rida Mc^^t jAe Proposed .196;"' Constitution? Gainesville: University of Florida fSli'c'ASininistration Clearing Service, 1968. Studies in Public Administration No. 31. David,. Paul T and Sisenberg, RalphDevaluation, of the UjjDan and S.uJ2urban Voiis. Bureau of Public Administration," University of Virginia, 1961. Dixon. W. J., ed. M-Sag-^Lic^i C^ompute^ Berkeley: University of California Press, 1968. 263

PAGE 280

264 Dve, Thomas R, Po litics,. Ec-pnoniics, and, the Pub .lie. ChicagoRand McHally, 1966. Fenton, John H. |'lilwe.s;'c Pol itics. Nev/ York: Holt, Rinehart & Winston.1SS6. Harlan, Harry H, yiodern Factor Analysis. Rev, Ed., Chicago; Univei'sity of Chicago Press, 1957. Havard, Williaiti C, and Beth, Loren P. Represent ative Gover nment .and Reaj^sioxtixsnient: A Case ,Stuc[y. of. Fl_qr.ida. Gainesville: Public Administration Clearing Service of the University of Florida, 1950. Studies in Administration No. 20. The Politics, of Mis-Representation. Baton Rouge: Louisiana State University Press, 1962. International Business Machines. Systera/36 Scientific Subroutine s P ack a ge White Plains, N. Y. : IBM Technical Publications Department, 1958. Kev, V. 0„, Jr. Southern Pol itics Nev; York: Alfred A.' Knopf, 1949. Lockard, Duane. New England State Politics. Chicago: Henry Regnery, 19 59. MacRae, Duncan, Jr. issjies 3_nd Pa rties ijn Legislative. Voting; MsVhg,ds o,f Statistical Analysis. New York: Harper & Row, 1970. Morris, Allen. Tjne Flor id a Handbo ok. Tallahassee: Peninsular Publishing Co., 1969. Newton, V, M„, Jx Crus ade for De mocracy Ames, Iowa: Iowa State University Press ^ 1961. Ofjic.i_al,s o,f F_loX,ij5.2 Mun i c i pa 1 i t i e s Jacksonville: Florida League of Municipalities, 1970. Salisbury, Robert H. Missouri Politics and State Political Systems Bureau of Government Research, University of Missouri, 1959. Sharkansky, Ira. Spending in the American States. Chicago: Rand McNally, 1968.

PAGE 281

SiTiitn,, C. Lynwc'Od, or, Stre'n:':j'l:heriincf the .Flj2.ili.2J. Jd§.Sl_§.l.^ti'J?.S.' New Brunswick, N, J„ : Eaglaton Instihuce of Foiitic;s, 1959. Syed, Anwar. Zlie Polit i'ca l T'neorv. jof Arri€:;ri_caD_ I^ocal Govern," nvant. New York: Random House, 1966. v^^irt, Fredrick M. and Hav/ley, Willis D,, eds. The Search for C_o;Timun.itv Pow er. Englewood Cliffs: PrentiGe-Hall, iS6S II. Articles Bohrnstedt, George W, "Observations on the Measurement of Change," in Edgar F. Borgatta, ad. Soci.o.lo.gi.c.al Methqdplpqv: .1569. San Francisco: Jossey-3ass, Inc., 1968. pp. 113-133. Butler, Gordon T. and Marsicano, Ralph. "Legislation Affecting Florida Municipalities," Florida M unici pal Reco rd, ^CLI (August, 1967), 3-4, 8-9. "Report of the 1963 Legislatnf-e as Affecting Municipalities," Flori.da Mun icip al Record., XXXVII Y (uuly, 1963), 3, 15-17 "Report of the 1965 Legislature as A^ffecting Municipalities, Flor_ida M^inicical. Record, XXXIX (July, 1965), "3, 5, 15, 23, 27, 31. ___. "Report of the 1969 Legislature as Affecting Municipalities," F_l or..i djs. M^jjiLCJjoaJ,. j^^ XLIII (July, 1969), 3, 5-7, Cattell, Rayxnond B. "The Meaning and Strategic Use of Factor Analysis," in Rayinond B, Cattell, ed. Handbook o.f :!ili~ t iOiJ^Xi-?..^ Experiment 3 1 Psychology. Chica.go: Rand McNally, 1966. Pp. 174-243. Cnudde, Charles F„ and McCrone, Donald J. "Party Competition and Welfare Policies in the American States, Arnerican Poli tical Science Review, LXIII (September, 1969) 85SCollins, Eldridge R. and Friedrnan, Howard Jay. "The 1953 Legislat^are, F.lorida Scho ols XII (SepterrLber-October 19S9), 22-23.

PAGE 282

266 / "^^ Daiaer, Kanning J. "Florida Reapportionraent B-gsines.s ajii^i.' '^£2:^2^i£ ^22J^lSJlSl3BS.i ^^^ (Marc>., 1967), 8-14. and Kelsay, Robert G. "Ux-ireprersentative States," National Mx^nicipal Re view XLIV (Deceinber, 1955) 571-575'. Dawson, Richard and Robinson, James. "Inter-Party Competition, Economic Variables, and Welfare Policies in American States," Journal of Politics, XXV (May, 1963), 265-239. Derge, David. "Katropolitan and Outstate Alignitients in Illinois and Missouri Legislative Delegations, Am erican Political Science Re view LII (DeceTxiber, 195S) 1051fo 6 5 Dye, Thomas R. "Comparison of Constituency Influences in the Upper and Lower Chambers of the State Legislature," V^estern Politi.cal Quarterljy, XIV (June, 1951), 473-480. 'Malapportionment and Public Policy in the States," Jo-arnal of Politi cs, XXVII (August, 19S5) 585-601. and D3;'son, James W. "Conmunication to the Eiditor, American Polit ical Science R eview LXIII (June, 1959) r 528-329. Kulau, Heinz. "The Ecological Basis of Party Systems: the Case of Ohio, Midwest Journal _of Political Science I (August, 1957)"', 125-135. Fenton, John H. and Charaberlayne, Donald W. "The Literature Dealing v>'ith the Relationship Between Political Processes, Socio-Economic Conditions, and Public Policies in the American States: a Bibliographic Essay," Polity-, I (Spring, 19S9) 388-394. Flinn, Thomas A. "Party Responsibility in the States: Som.e Causal Factors, AmsJS'i.-Ca.n Po litic al Sc^pLgmig. EiiyJ^mi, LVIII (March, 1964), 50-71. "The Outline of Ohio Politics," Western P ojA:ti-'--a 1 5^a3-;t9.^=.y ^^^^ (September, 1950), 702 --72 1

PAGE 283

26 7 Friedman, Robert S, "The Urban-Rural Conflict Revisited," •E@.3jL9rn, PoiJ^tical Qi^arterlv, .XIV (June', 1961) 481-495. Fry, Brian R. and Winters, Richard F„ The Politics of Redistribution, '' ii?iisr_icaTi pQlitical, S.cience Revi ew, ^ LXIV (June, 1970), 50S~522. Frye, Barbara. "Coalition Flexes Muscle," Tampa Tribune, (April 14, 1970) 6-A. Greene, O'uanita "Did Reapportiorniiient Help Cities," The Miilffil Herald, (January 3, 1971), l-C. Gruirm, John G. "A Factor Analysis of Legislative Behavior," ilidrwest. Journal .of Pol.iti.ca.l. Science, VII (Noveirbsr, 1963), 335-356. "The Means of Measuring Conflict and Cohesion in the Legislature," S g u t hw e ste r n Social .Sciences Quarterly^ XL IV (March, 1964), 377-383. Hawkins, Brett '^7. "Consequences of Reapportionment in Georgia," in Richard 1. Hofferbert and Ira SharkansTcy, sds. State and Ilirban Politics : Rs?_dij1.3s. .ijl ,Q.Q31S.3.5L?Jt.il'l:§. P.yt0^ic PoldcxBoston: Little, Brown, 1971. Rp. 273298." "' Hofferbert, Richard I, "The Relation 3etVi?een Public Policy and Some Structural Environmental Variables in the American States,." Aiaer jean Pol itical Scie nce Review, LX (March, 1966), 73-32. Hofstetter,. C. Richard. "Malapportionment and Roll-Call Voting in Indians," Journal of Pol.itics, XJiXIII (February, 1571) 92-111'. Jacob, Herbert. "The Consequences of ?-ia.lapportionraent : a Note of Caution," S-ocij3_l Foirces XLIII (December, 1964), 2 56-261. _. and Lipsky, .Michael. "Outputs, Structure, and Pcv;er: An Assessment of Change in the Study of State and Local Politics, Journ^Ll of Folit.ics, XXX (May, ^ 196B) 510-553.

PAGE 284

^ '-J ^> Jewell, M^ilGolm E. "Political Patt.erns in Apportionment," ^ in Malcolra E. Jewell, ed. Th e Politics of. .^laSiSSXtisa.-inentSew York: Atherton Press, 1962. Pp. 1-48 McCrone, Donald J, and Cnudde, Charles F. "On M-easuring Public Policy," in Robert E, Crew, Jr., ea. S_ta;te ^ Politics: Readings on P olitical Behavior Belmont, California: Wadsworth, 1958. Pp. 523-550. .HocRae, Duncan, Fr. "The Relationship Between Roll Call Votes and Constituencies in the Massachusetts House of Representatives, AjBeri.ca_n Political Sc ience Re view XLVI (Deceiuber, 1962), 1046-1055. National Municipal League, State Legislatures. .Pr.o.gj;.eg.s ^ Rep ort er (July-August, 1969. Parker, Daisy and Waldby, H. Odell. "Twenty Years of Struggle for Legislative Apportionment in Florida," £.l5Li:4a. State Univers ity Governme nta l Research Bulletin.. Price, H. D. "Florida: Politics and the 'Pork Choppers,'" s in felcolm E. Jewell, ed. The RLea]2portj-_onjiient Reyolution. New York: Atherton Press, 1962. Pp. 81-9 7. Pulsipher, Allan G. and Weatherby, James L., Jr. "Malapportio-Uaent Party Competition, and the Functional Distribution of Governmental Expenditures' .Amer.ic.an. ^ Pol.iti,CLaI £c.ience. .R'Bview, LXII (December, 1968), 12071219. Richardscn, Michael, "Legislature of Florida 4th in Nation," tu'-' P.cter5b urq Times, (February 4, 1971) 1-A. isfe..?Leter3j.ju_ia ^LiSfi-S.' (^''^ne 25, 1970). Schubert, Glendcn and Press, Charles. "Measuring Malapportionment, Americaji Poll tic. 3.1 Sci ence Re2.Li.§!il/ LVIII (June, 1964), 302-327. Schults, Fred. "Legislative Modernization — the Florida Exoeriment, £tj3jfce, Gojierjimeji^ XLII (Autumn, 1969), 256-259 X Sharkansky, Ira and Hofferbert, Richard I. "Dimensions of State Politics, Economics, and Public Policy," AmeJliiiaJl Political Science Rev iev/ LXIII (September, 1969), 867879.

PAGE 285

259 Shall, Charles VJ. "Political and Partisan Implications of State Legislative Apportioninent, Law and C_qnteraporary. Prpbl^eriisJ XVII (Spring, 1952) 417-439 WalKer, Jack L. "The Diffusion of Innovations Among the American States, Ame rican Political Science Pv.eviaw, "^ LXIII (September, 1969) 880-899, Wigains, Charles. "Party Politics in the Iowa Legislature," Midwest ylournal of Po litic al Science, XI (February, 1967), 8S--97. ^^^ • Legal .Mat eri als Flcrida. Cons ti tut ioii of 18 85. S_eJlivda. Cons tit u,t ion of^ 1959., Flor ida S_tl.ty.1l§.§.ligws of. -Flo rid a. IVM3_nu s cr ip,t s Daaer, Manning J. "Florida is Different; So Is Her Politics Different. y. Court Cases EsJsai: iL. Ciir.r.,United States Supreme Court, 369 U. S. 186 (1962)

PAGE 286

B 10 GRAPHICAI, S :^CfiTCH John Richard Todd v/as born December 13, 1939, at Freeport, Texas. In June, 1957, he was graduated from New Iberia High School, New Iberia, Louisiana. In June, 1961, he received the degree of Bachelor of Arts from the University of Southwestern Louisiana, In Septeirber, 1961, he enrolled in the Graduate School of the University of Florida. He worTced as a graduate assistant in the Department of Political Science until April, 1964. In June, 1954, he received the degree of Master of Arts with a major in Political Science. From September, 1964, through June, 1966, Mr. Todd served as an Instructor of Social Studies at Central Florida Junior College. In September, 1956, he resumed work at the University of Florida tov/ard the degree of Doctor of Philosophy. From Septeiriber, 1968, t'nrough August, 1971, he v;a.s an Instructor of Social Sciences at the University of Florida. Mr. Todd is currently an Assistant Professor of Political Science at North Texas State University. John Richard Todd is married to the former Lannis Patricia Fuip and is the father of three children. He is a m.airber of Pi Sigma Alpha, Phi Alpha Theta, Phi Kappa Phi, and Pi Kappa Delta honorary fraternities. He is also a merrber of the Southern Political Science Association.

PAGE 287

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. "^^^t/^ Manning -f/^Dauer, Chairman Professor of Political Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Alfred B. Clubok Associate Professor of PoliMcjal Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Marvin E, Shaw Professor of Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. /\^€^y^c^-^ /Ar/Tt^^^^^iJ^ i^^ Thomas A. Henderson Assistant Professor of Political Science

PAGE 288

-jSS \. -J^--. I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. DaVid P. Conradt Assistant Professor of Political Science This dissertation was submitted to the Department of Political Science in the College of Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December, 1971 Assistant Dean, Graduate School


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E29TFESZ9_0R2SGM INGEST_TIME 2014-11-14T19:57:25Z PACKAGE AA00026391_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES