ABSTRACT

RECONCEPTUALIZING THE RELATIONSHIP BETWEEN ETHNICITY AND
PUBLIC RECREATION: A PROPOSED MODEL

By

Edwin GOmez

This study considers the relationship between ethnicity and public
recreation settings. A better understanding of the use of public recreation space
by minority group members is needed. A more complex measure of the
relationships between ethnicity and recreation is needed. This study reviews a
selection of the most often cited studies, from the previous 30 years, and
outlines the major assumptions and relationships between variables and
constructs. Although empirical literature has been growing, there have been (1)
methodological inconsistencies, (2) requests for innovative approaches, and (3)
requests for research on ethnic groups other than African Americans.

Through careful analysis, a more coherent and comprehensive picture of
previous research was developed. The end result was the development of a
conceptual model which was operationalized in the study. This model is called
the Ethnicity and Public Recreation Participation (EPRP) Model and consists of
six constructs: (1) acculturation, (2) socioeconomic status, (3) subcultural
identity, (4) perceived benefits of public recreation, (5) perceived discrimination,
and (6) public recreation participation.

Puerto Ricans from Southbridge, Massachusetts represented the ethnic

group under study. A questionnaire was developed to measure the

PLACE IN REIURN Box to remove this checkout from your record.
To AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

020)

P

 

U

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ma mm.“

 

aforementioned constructs. The EPRP Model was tested empirically with
confirmatory factor analysis and path analysis techniques.

The initial model was consistent with the data, but error and path
coefficient diagnostics led to a revised, theoretically consistent model involving
the same constructs, and similar causality. Substantively speaking, there is not
much difference between the proposed EPRP Model and the revised EPRP
Model. They both hypothesize a similar array of relationships.

The researcher de-constructed the model to illustratethe theoretical
consistency between parts of the model and earlier models conceptualized by
previous researchers. The model was found to be theoretically and empirically
sound. The study found that (1) the most important factor was acculturation; (2)
there is a significant spurious relationship between the marginality and ethnicity
factors; and (3) perceived benefits of recreation is a significant construct to be

considered for future studies.

Copyright by
EDWIN GOMEZ
1 999

DEDICATION

This dissertation is dedicated to my parents: Natividad and Julia Gomez. I can
never repay the debt I owe them for their countless sacrifices that allowed me
the opportunities to obtain a graduate degree. Through their unconditional love,
support, and faith in me I made it through. Los quiero con todo mi corazdn y
quiero qua sepan que no son ustedes Ios que deben ser orgulloso de mi, soy yo
el que debo ser orgulloso de tenor padres cdmo ustedes. Gracias por todo.
iBendicidn!

ACKNOWLEDGMENTS

I would like to extend very sincere thanks and appreciation to my
dissertation committee members: Joe Fridgen for accepting to be the chair of
my committee, looking out for my best interests, and serving as a role model;
Michael Chubb for sending me constant updates on my topic throughout the two
years I spent researching the topic; Dan Stynes for getting me on track earty on
with the tough questions regarding my methodology and sample; and John
Schweitzer who was my employer and friend for most of my tenure here, and
who had countless "talks" with me - professional, statistical, or othenrvise.

In addition, I would like to thank five specific individuals that helped
immensely through this process: Ralph Levine who met with me off campus and
after class to discuss my model and causality; Frank Boster who met with me
during the summer, and answered my constant e-mails regarding statistics; Jim
Bristor who challenged my writing skills through his editing, and Omara Rivera -
Vasquez and Julia Gomez who helped immensely with the translation and
formatting of the questionnaires.

I would like to thank the many friends I have made at the Department of
Park, Recreation and Tourism Resources, Urban Affairs Programs, and the
Department of Psychology at Michigan State University. Thanks for the reality
checks and for allowing me to be me. You will all be missed. Lastly, I would like
to thank the Puerto Ricans of Southbridge, MA who gave of their time to

participate in this study. iGracias por todo!

TABLE OF CONTENTS

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

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

CHAPTER 1 .................................................... 1

INTRODUCTION ................................................ 1

The Statement of the Problem ................................. 5

Objectives ................................................ 7

Hypotheses ............................................... 7

Delimitations .............................................. 8

Definitions ................................................ 9

Basic Assumptions ......................................... 10

Limitations ............................................... 11

CHAPTER 2 ................................................... 12

LITERATURE REVIEW .......................................... 12

Development of the Ethnicity and Marginality Theories ............. 12

The Assumption Regarding Residency ......................... 17

Race and Ethnicity: Is a Distinction Needed Between the Two? ...... 19

Inadequate Measures of Race and Ethnicity ..................... 23

Moving FonNard ........................................... 24

Alternative Approaches to Studying Minority Recreation ............ 25

Construct versus a variable approach. ................... 25
Reconsidering past approaches and exploring alternative

theoretical considerations: Pre-cursors to a model ...... 27

CHAPTER 3 ................................................... 45

METHODOLOGY ............................................... 45

Moving Forward: The Development of a Model ................... 45

Questionnaire Design ..................................... 5O

Accultu ration. ...................................... 50

Subcultural identity. .................................. 52

Perceived discrimination. ............................. 56

Perceived benefits of recreation. ........................ 57

Socioeconomic status. ............................... 58

Recreation participation. .............................. 59

Subjects and Sampling Frame ................................ 61

Geographic profile. .................................. 62

Demographic profile. ................................. 62

vii

Population and sample size. ........................... 64

Procedures for high school sample selection. .............. 65

Procedures for adult population sample selection. .......... 66
Selection of subjects: problems and solutions to

creating a list. ................................. 68

Pretesting the Questionnaire ................................. 71

Pretest analysis of scales .............................. 72

Perceived benefits of parks (1A—1G). .................... 73

Public recreation participation (Section 2A-2E). ............ 73

Acculturation scale (Section 3A-3G and 4A-4D). ........... 74

Perceived discrimination (Section 5A-5E). ................ 76

Subcultural identity (Section 6A-6F). ..................... 76

Demographic section of pretest. ........................ 77

Questionnaire distribution. ............................ 78

Collection of Data ........................................ 79

Treatment of Data ......................................... 80

Validity and Reliability ...................................... 81

CHAPTER 4 ................................................... 85

DATA ANALYSIS ............................................... 85

Section One: Descriptive Statistics ............................ 85

Survey Response .................................... 85

Sample and Population Demographics .................... 87

Univariate Analysis .................................... 90

Ancestry. .................................... 90

Acculturation and subcultural identity. .............. 90

Park use. .................................... 92

Data Coding and Clean-up ............................. 95

Section Two: Confirrnatory Factor Analysis ...................... 96

Internal Consistency .................................. 97

Perceived benefits of parks. ...................... 97

Accultu ration. ................................. 99

Perceived discrimination. ....................... 100

Subcultural identity. ........................... 102

Public recreation. ............................. 102

Socioeconomic status. ......................... 103

Test of Parallelism ................................... 103

SES construct as an example of parallelism. ........ 104

Section Three: Path Analysis on EPRP Model .................. 108

Path Analysis Concepts .............................. 109

Over-identification status. ....................... 109

Visualizing the correlation matrix. ................. 110

Local versus global test. ........................ 111

Correlations and path coefficients. ................ 112

The Path Analysis ................................... 113

viii

The Revised Model ................................. 118

Path Analysis for Revised Model ........................ 119
The Final Revised Model ............................. 121
Path Analysis for Final Revised EPRP Model .............. 122
Interpretation of the Model ............................ 124
Examination of Hypotheses ........................... 125
Hypotheses .................................. 125
Subcultural identity and socioeconomic status
on public recreation usage. ................. 126
Perceived discrimination and perceived
benefits on park use. ..................... 126
Subcultural identity and socioeconomic
status relationship. ....................... 127
Acculturation/subcultural identity and
acculturation/socioeconomic status. .......... 127
Subcultural identity and perceived
discrimination. ........................... 127
Summary of Findings ................................ 128
Acculturation, perceived benefits, and perceived
discrimination. ........................... 128
Socioeconomic status and subcultural identity. ....... 130
Public recreation participation. .................... 131
Chapter Summary ................................... 131
CHAPTER 5 .................................................. 133
SUMMARY AND CONCLUSIONS ................................. 133
Discussion .............................................. 134
De—constructing the EPRP Model: Reflections of
previous literature. ............................. 134
Removal of acculturation. ............................ 134
Removal of perceived benefits. ........................ 136
Removal of acculturation and perceived benefits. .......... 137
Conclusions ............................................. 139
Shortcomings and Limitations ............................... 141
Implications ............................................. 142
Conceptual implications. ............................. 142
Practical implications for the field. ...................... 143
Recommendations for Further Study .......................... 146
Procedural recommendations. ........................ 146
Questionnaire recommendations. ...................... 146
Statistical recommendations. ......................... 147
Conceptual recommendations. ........................ 148
Summary ............................................... 149

ix

APPENDIX A ................................................. 150

English and Spanish Questionnaires with Descriptive Statistics .......... 150
Southbridge Public Parks: A Survey of Puerto Rican Residents”
Experiences and Opinions ............................ 151
Parques Publicos de Southbridge: Un Estudio de las Experiencias y
Opiniones de los Residentes Puertorriquefios ............. 158
APPENDIX B ................................................. 165
Map of Parks in Southbridge and Samples of Correspondence ........... 165
Map of Southbridge, MA and its Public Parks ................... 166
Sample Letter for Community Leaders ........................ 167
Sample of Sign-up Sheet for Survey Participation ................ 168
Sample of Speech for Spanish-speaking Parishioners at St. Mary’s
Church During Christmas Eve Mass ..................... 169
Sample Letter for Administration at the High School .............. 170
' Sample Letter for Administration at Head Start Program ........... 171
Sample Pre-survey Letter: English ........................... 172
Sample Pre-survey Letter: Spanish ........................... 173
Sample Pro-survey Autobiography: English .................... 174
Sample Pre-survey Autobiography: Spanish .................... 175
Sample Introduction Letter for Survey: English .................. 176
Sample Introduction Letter for Survey: Spanish ................. 177
Sample Introduction Letter for Follow-up Survey: English .......... 178
Sample Introduction Letter for Follow-up Survey: Spanish ......... 179
APPENDIX C ................................................. 180
Confirmatory Factor Analysis Concepts ............................. 180
CFA Concepts ........................................... 181
The Measurement Model ................................... 182
Internal Consistency Theorem ............................... 184
Multi—factor Model ........................................ 185
Parallelism Theorem ...................................... 186
Confirmatory Versus Exploratory Factor Analysis ................ 187
APPENDIX D ................................................. 189
Data Analysis Matrix and Chart ................................... 189
Correlation Matrix of 22 Items Used in
Final Confinnatory Factor Analysis ...................... 190
Local and Global Tests for Three EPRP Models ................. 191
REFERENCES ................................................ 192
References ............................................. 193

Table 1.

Table 2.

Table 3.
Table 4.
Table 5.
Table 6.
Table 7.
Table 8.

Table 9.

Table 10.

Table 11.

Table 12.
Table 13.
Table 14.
Table 15.
Table 16.
Table 17.
Table 18.
Table 19.

Table 20.

LIST OF TABLES

ConceptsNariables Addressed in the Literature Regarding

Ethnicity/Race and Public Recreation Participation ..... 26
Direct and Indirect Effects of Ethnicity

Constructs on Recreation Participation .............. 49
The Acculturation Scale ............................... 53
Subcultural Identity Scale .............................. 55
Perceived Discrimination. .............................. 57
Scoring of Park Use for Each Park ....................... 61
Mailed Surveys and Response .......................... 86
Age and Gender Breakdown in Census and Sample ......... _ 88

Education for Persons 25 and Over of Latino

Origin in Census and Sample ..................... 89
ANOVA of Years in Puerto Rico by Age Cohort ............. 91
Mean Difference in Perceived Benefits

of Parks Between Students and Adults .............. 93
Assessing Errors for BEN Construct ...................... 98
Descriptive Statistics For Items Used in the CFA ........... 101
Item by Factor and Factor by Factor Correlation Matrix ...... 105
Parallelism Analysis ................................. 107
Reliability Analysis .................................. 108
Obtained Correlation Matrix for Path Analysis (N=318) ...... 113
Matrices for Original EPRP Model from Program Path ....... 116
Matrices for EPRP Model from Program Path (revised) ...... 120

Matrices for EPRP Model from Program Path (final version) . . 123

xi

Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.
Figure 6.
Figure 7.
Figure 8.

Figure 9.

Figure 10.
Figure 11.
Figure 12.
Figure 13.
Figure 14.
Figure 15.
Figure 16.
Figure 17.
Figure 19.
Figure 20.

Figure 21.

LIST OF FIGURES

Lindsay & Ogle’s (1972) Conceptual Model ................ 28
Craig’s (1972) Conceptual Model ........................ 30
Washbume’s (1978) Conceptual Model ................... 31
Klobus-Edwards' (1981) Conceptual Model ................ 32
Stamps 8. Stamps’ (1985) Conceptual Model ............... 34
Hutchison’s (1987) Conceptual Model .................... 35
West’s (1989) Conceptual Model ........................ 37

A Model of the Factors Affecting Ethnic Leisure Participation . . . 39

Conceptual Model of Explanations of Ethnic Variation

in Recreational Behavior ......................... 42
The Ethnicity and Public Recreation Participation Model ...... 45
The EPRP Model (with non-constrained paths) ............ 111
The EPRP Model (revised) ............................ 119
The EPRP Model (final version) ........................ 122
The EPRP Model (final version) ........................ 124
The EPRP Model without Acculturation .................. 135
The EPRP Model without Perceived Recreation Benefits ..... 137
The EPRP Model without ACC and BEN ................. 138
One Factor Measurement Model ........................ 182
Multi-Factor Measurement Model ....................... 185
EFA Model ........................................ 188

xii

Chapter 1
INTRODUCTION

Over the past twenty years, research regarding minorities and public
recreation participation has predominantly centered around two theories: (1)
ethnicity theory, and (2) marginality theory. The ethnicity theory simply states
that minorities do not participate in ”mainstream" recreation activities due to
cultural differences. The marginality theory states that participation in recreation
activities by minorities is limited due to socioeconomic factors. A

Although studies regarding ethnicity, socioeconomic behavior, and
recreation have existed prior to 1978 (Burch, 1969; Hendee, 1969; Craig, 1972;
Lindsay & Ogle, 1972; Schumen 8. Gmenbeng, 1972), it was Washbume’s
(1978) seminal study that initiated the current paradigm regarding the ethnicity
and marginality theories’ approach to studying ethnic/racial phenomenon with
respect to recreation and recreation settings. Since then, problems have
occurred in operationalizing major concepts, and consequently, in measuring the
two theories. I

A decade after the marginality/ethnicity framework was developed, West
(1989) made a significant conceptual contribution to the empirical literature by
noting that recreation researchers neglected to explore the problem of interracial
relations. Discrimination (whether real or perceived) became a probable
explanation for the under representation of minority groups in outdoor recreation
settings. The inclusion of this concept creates an even more complex and

sensitive framework for the study of minorities and recreation behavior.

In the 1990s, the empirical literature regarding ethnicity/race and
recreation behavior has been growing in importance for several practical and
empirical reasons. One practical reason is changing population demographics
(Dwyer 8. Gobster, 1992). Ethnic group members have increased tremendously
as a percentage of the population in the United States (US). A culturally
(ethnically and racially) diverse population is projected to increase. Dwyer
(1994) reported that between 1990 and 2025, the US. population is expected to
increase by 50 million and that 81% of the increase will be in minority groups.
According to the US. Bureau of the Census, by the year 2010, Hispanics are
projected to be the dominant ethnic group, surpassing Blacks.

As a result of demographic changes, a related reason for the study of
ethnicity and recreation is that recreation providers will have tremendous
challenges ahead in terms of service delivery, policy-making, and identifying
participation patterns of "non-traditional” users. A better understanding of the
use of public recreation space by minority group members is needed. Blacks and
Hispanics, the two largest groups, tend to concentrate in urban centers. This
concentration of minorities in cities could mean declining use of state and
national parks and added stress to urban park and recreation resources.

In addition to the practical reasons for the significance of a study on
minorities and their recreation behavior, there are several empirical reasons to
conduct such studies. First, the relationship between one's ethnicity/race and
park use is unclear from previous literature and requires further exploration

and/or reconceptualization. Second, conceptually, recreation researchers have

had difficulty in defining and operationalizing key concepts when researching
ethnicity and recreation. Third, the nature of the interaction between ethnicity
and marginality requires further exploration. Current understanding of the two
concepts and their combined impact on recreation behavior has not been
specifically explored in the design of previous studies.

Fourth, and most important (conceptually), Floyd, Gramman and Saenz
(1993) and West (1989) noted that marginality and ethnicity theories have
suggested that a perception of discrimination influences levels of recreation
participation. However, "lack of attention paid to this issue in the recreation
research literature continues despite current policy debates regarding structural
discrimination and the spate of empirical research on discrimination in the social
sciences at large“ (Floyd, et al., 1993). Research is needed to explain the
impact discrimination has on use of recreation areas.

Although the empirical literature has been growing, several concerns
have been identified: (1) methodological inconsistencies, (2) requests for
innovative approaches, and (3) requests for research on minority groups other
than African Americans (whom are the dominant minority group used for
comparison). In response to the first concern, this study reviews a selection of
previous studies and outlines the major assumptions and relationships between
various variables and constmcts. Through careful analysis, a more coherent
and comprehensive picture of previous research will be developed.

An innovative approach to research involving minorities and recreation

behavior is the measurement of one of the independent variables - subcultural

identity. This approach differs from previous studies because it is not simply
identifying persons as Black, White, or Hispanic, but rather it asks the extent to
which one feels they are affiliated as a member of a minority group. The
constmct is a more productive measure of ethnicity.

This reconceptualization is important for two reasons. First, it is more
consistent with Washbume's original conception of the ethnicity theory. Second,
it focuses less on the issue of whether the minority group is ethnic or racial, and
focuses more on the issue of minority versus majority. The subcultural identity
dimension measures one’s identity as a minority group member, no matter the
racial or ethnic makeup. This notion detects the identification with the minority
culture versus the majority culture.

Although this is different from the traditional sociological conception of
race and ethnicity, it is consistent with everyday usage. For example, when
majority people discriminate against a minority group member, they often do not
distinguish between whether the minority is a racial minority or an ethnic
minority.

The literature review provides an additional innovation - a basis for the
development of a conceptual model that considers the interaction of several
ideas regarding ethnicity and recreation behavior. After reviewing past studies,
the author has incorporated several conceptual constnicts into a model
illustrating the interactions between the different constructs. This model

consolidates previous research into a ”snapshot" of what the empirical literature

considers to be the most critical dimensions of ethnicity, recreation behavior.
Park use is used as a proxy for recreation behavior in this study.

This study also considers the request for research on minority groups
other than Blacks. The Hispanic pepulation is the second largest minority
population in the United States. Of these, Mexicans have been looked at in
particular in recreation studies because they are the largest component of the
Hispanic minority group. Previous studies that researched Hispanics are often
concentrated in the Midwest, South or West.

This study differs from previous studies in that the comparison ethnic
group consists of Puerto Ricans living in the Northeast. In addition, most of the
Puerto Ricans in Southbridge are immigrants (respondents emigrated from
Puerto Rico) or first generation Americans (parents from Puerto Rico, but
respondents born in continental US). This allows for the inclusion of
assimilation/acculturation measures in the analysis of recreation behavior.

The Statement of the Problem

The purpose of this study is to identify the factors (based on empirical
literature and logic) which are most important in explaining the relationship
between one’s (sub)cultural identity and its effect on one’s recreation behavior
as represented by park use (dependent variable). These (independent) factors
include the following: (1) acculturation, (2) socioeconomic status, (3) subcultural
identity (minority versus majority), (4) perceived park benefits, (5) perceived

discrimination.

Presently the research question is no longer whether ethnicity
(subcultural identity) is a stronger predictor of recreation behavior than
marginality (socioeconomic status), or vice versa. The issue is unresolved, but
research is leaning towards ethnicity as the greater predictor of recreation
behavior. As such, this study considers three research questions.

The first research question addresses which of the two factors
(marginality or ethnicity) is the better predictor of recreation behavior. The
second research question addresses the combined importance of ethnicity and
marginality factors on recreation behavior (park use). The third research
question considers the relationships between acculturation, perceived benefits,
and perceived discrimination on park use.

The development of a model will illustrate the relationships between the
previously identified factors. These relationships are based on prior empirical
results. The purpose of the model is to simplify the literature and provide a
basis for conceptually thinking about minority issues related to recreation. The
model is not exhaustive, rather, it reflects the important "currents" in the
literature. For example, there could be other factors that could influence park
use, such as distance to the resource, cost of transportation, age, and family life
cycle stage. However, exploration of these and other factors are not possible

due to time, financial, and practical constraints.

Obiggiveg

1.

Determine the relative influence of subcultural identity and socioeconomic
status on park use.

Determine the relative influence of acculturation, socioeconomic status,
subcultural identity, benefits sought and perceived discrimination on park
use.

Develop a model to illustrate the relationships between the different
factors influencing park use.

Hymtheses

1.

Subcultural identity [SID] and socioeconomic status [SES] are significant
predictors of public recreation participation [REC]. Subcultural identity
will be a stronger predictor of park use than socioeconomic status.

Ho: rsroIRrsc = 0. l'srssrRrsc = 0
H1: rsrorrzrsc * 0- rsrssrrzec * 0
H23 l'srornrsc > rSESIREC

There is a significant relationship between perceived discrimination [DIS]
and public recreation participation, and perceived benefits of recreation
[BEN] and public recreation participation.

H03 rDIS/REC = 0. raswnec = 0
H13 Isle/Rec 9‘ 0. rsrssrRrsc i" 0

There is a significant relationship between the subcultural identity factor
[SID] and socioeconomic status [SES].

I"Io: rsro/srss = 0
H13 rsrorses * 0

There is a significant negative relationship between acculturation [ACC]
and subcultural identity. There is a significant positive relationship
between acculturation and socioeconomic status.

Ho: rACCISID Z 0- I'rtccrsrss 5 0
H13 rACCISID < 0. rAce/ses > 0

5. There is a positive relationship between subcultural identity and
perceived discrimination.

Ho: r00 5 0
H1: rCD > 0
6. There is a negative relationship between perceived discrimination [DIS]

and public recreation participation [REC].

Ho: rmc z 0
H13 romeo ‘ 0

7. The path model hypothesis:

Ho: the data are consistent with the model (Error = 0).

H,: the data are inconsistent with the model (Error :0).
Delimitations

There are several things that could be included in this study regarding
minorities and recreation behavior. For example, one could include data
regarding life-cycle, the presence of children, neighborhood ethnic/racial mix, or
sense of belonging to a community. Structural factors such as the impact that
schools have on minorities and recreation behavior could also be considered.
Anthropological factors such as the historical development of the ethnic make up
of Southbridge and intercultural relations would also be noteworthy.

While some of the factors previously mentioned will be considered in the
literature review, as well as in the background of the city, it will be contextual in
nature. To study all possible and relevant factors would be both time-consuming
and costly. This study is not meant to be comprehensive, but rather it converges

upon the key variables outlined in the literature review.

The study is further delimited to one city: Southbridge, Massachusetts.

All cities have a very unique sociocultural history. As such, comparisons will

vary from city to city. Although distinct findings will vary by city, a clearer

understanding can be made possible regarding the Puerto Rican population
because of the close ties that Southbridge Puerto Ricans maintain with the
native population in Puerto Rico.

Definitions

Cultural Acculturation: the process whereby ethnic and racial minority groups
retain their own cultural norms while adopting aspects of the dominant
culture.

Cultural Assimilation: the process whereby ethnic and racial minority groups
absorb the cultural and behavioral patterns of the dominant group
(Gordon, 1 964).

Cultural Identification: the extent to which one identifies with the minority
(subcultural) or majority (mainstream) culture.

Ethnic Minorities: minorities who are discriminated against by the dominant
group because of cultural differences (language, customs, traditions).

Mainstreamed Values: values that are consistent with norms established by the
majority group. When one is "mainstreamed" it connotes a weaker
affiliation with a minority culture and a stronger affiliation with the majority
culture. Mainstreaming is the process by which one is integrated into the

wider socioculture.

Majority Group: defined as the dominant group in a cultural setting that sets the
cultural norm(s) for their given society. In the United States, traditionally,
this dominant group has been Whites - males in particular.

Minority Group: defined as the ethnic or racial groups which ascribe to
subcultural norms within the wider socioculture.

Racial Minorities: minorities who are discriminated against by the dominant
group because of their phenotype (physical features).

Subculture; the use of subculture in this study connotes a minority culture which
is different from the dominant culture. It is not meant to imply that a
minority culture has less value than the dominant culture.

Basic Assumptions
It is assumed that individuals have basic values that are particular to their

cultural environment. The family is the strongest influence on one’s value

system and cultural norms. This value system guides behavior, overall, and
recreation behavior, in particular.

There are no private parks in Southbridge, MA. Although public parks are
located around neighborhood areas, they are also located off of major city
streets. There are also no restrictions to public park use by residents in other
parts of the city. As such, it is assumed that all Southbridge residents have
access to all Southbridge public parks.

It is assumed that there is a dominant paradigm for recreation in the

United States. This paradigm is reinforced by the White majority. The stronger

1O

one identifies with the dominant culture the more they will participate in public
recreation activities and settings.
Limfis’ tions

The Puerto Rican respondents in this study cannot be considered
representative of other Hispanic groups and should not be generalized to the
entire Hispanic population. Southbridge Puerto Ricans can be generalized to
other Puerto Rican populations in the US. mainland and the island of Puerto
Rico. Although not representative of the entire Hispanic population, Puerto
Ricans can be compared to other Hispanic groups to illustrate differences and

similarities between the many Hispanic group members.

11

Chapter 2
LITERATURE REVIEW

Washbume's 1978 seminal study initiated the current paradigm regarding
the ethnicity and marginality theories’ approach to studying ethnic/racial
phenomenon with respect to recreation settings. Since then, problems have
occurred in operationalizing major concepts, and consequently, in measuring the
two theories. As such, this chapter provides a context for the development of
both the marginality and ethnicity theories. This chapter also provides a critique
of the literature regarding ethnicity/race and recreation and alternative F
approaches for studying the topic.
Development of the Ethnicmr' and Marginal'sy Theories

Washbume (1978) provided recreation researchers with the conceptual
definitions for ethnicity and marginality (within the context of Black
underparticipation in wildland recreation). “The marginality perspective
[suggests] that Blacks do not participate because of poverty and various
consequences of socioeconomic discrimination ..." (Washbume, 1978, p. 176).
As an alternative explanation to the marginality perspective, Washbume
maintained " that leisure patterns of Blacks are based on their subcultural
style, or ethnicity (Md, p. 177).

Note that the original definition of ethnicity referred to "subcultural style."
Consequent studies treated ethnicity as synonymous with race or ethnic group
designation. This is a deviation from Washbume’s initial conception of ethnicity

as subcultural style, and has different meanings in the wider social science

12

literature (West, 1989). The implication is subtle. The focus should not be on
the ethnicity of particular persons, but rather the identification that people have
towards a culture which is different from the dominant culture.

Sex, age, education, and income were the variables used to measure
socioeconomic status in Washbume’s study. Blacks and Whites were
compared. However, the ethnicity perspective was not examined, only the
marginality perspective. Washbume did not consider researching the ethnicity
perspective on the grounds that there was inadequate data suitable "for a
realistic examination of ethnicity as a valid perspective" (1978, p. 185).

Washbume found that Blacks are constrained by marginality factors.
However, he also mentioned that leisure differences between Blacks and Whites
" are not solely attributable to socioeconomic factors, " and later forecasted
"declining utility for the marginality perspective” (1978, p. 179). The implications
of his findings were ambiguous at best. While firmly establishing a logical
argument for the ethnicity and marginality theories, his conclusions were
inconclusive. This set the stage for consequent studies in the 1980s.

Researchers in the 1980s (Hutchison, 1987; Klobus-Edwards, 1981;
Stamps & Stamps, 1985; West, 1989; and Woodard, 1988) utilized the
marginality and ethnicity theories and designed their studies to answer the
question of which one of the two theories was correct. Various researchers
provided support and mixed results for both theories. Support for the ethnicity
theory came from Klobus—Edwards (1981) and Stamps and Stamps (1985).

Support for the marginality theory came from Woodard (1988) and West (1989).

13

Mixed results regarding support for marginality and ethnicity came from
Hutchison (1987). I
In the latter part of the 1980s, and particularly in the eariy 1990s,

recreation researchers heralded a new era regarding research on minority

groups and recreation settings. The focus was altered from determining which

one of the two theories is correct to "accepting” either one of the two theories, or

embracing a combination of the two theories and concentrating on other matters.

These "other“ matters included the following:

. utilization of other minority groups that were not Black for points of
comparison (Chavez, 1993; Hutchison, 1987; Floyd, Gramman &
Saenz, 1993);

. creation of projection models, use of demographics, and the
determination of future demand (Christensen & Dwyer, 1994;
Dwyer, 1993, 1994; Murdock, Backman, Hoque 8. Ellis, 1991;
Murdock, Backman, Colberg, Hoque & Hamm, 1990);

. use of observational techniques in the research (Hutchison, 1987);

. critique of the literature, as well as general reviews were more prominent
(Allison, 1988; Dwyer & Gobster, 1992; Gramman, 1996;
Hutchison, 1988; Johnson, Bowker, English & Worthen, 1997);

. definitions of race and ethnicity were debated upon (Hutchison, 1988);

. identification of other factors such as discrimination, family generations,
and neighborhood residence (Phillip, 1994, 1997; Taylor, 1992;
West, 1989);

. and the corroboration of subjective/qualitative measures and

objective/quantitative measures (Allison, 1988; Carr 8 Williams,
1993a; Floyd, et al., 1994; Woodard, 1988).

Throughout the various studies noted above, marginality and ethnicity have
been measured in different ways. The problem with the different approaches to

measurement is that they have not allowed for true comparisons between

14

studies because of two things: (1) different methodological approaches, and (2)
different operationalization of the same variables.

Variables measuring marginality have been approached very differently
from Hutchison’s eariier conception. Woodard (1988), for example, looked at
marginality in terms of 'classism,” which (citing E. Franklin Frazier) he defines as
"the belief that one should only recreate with one’s own social class group" (p.
94). Floyd, et al.’s, (1994) approach to marginality treated social class in
"subjective terms" (p. 162).

The variables measuring ethnicity were consequently divided into race
variables and ethnic variables (recall that this is a deviation from earlier studies
lumping ethnic and racial groups under one category). Taylor (1992) looked at
ethnic expression and participation in ethnic leisure activities as measures of
ethnicity. Carr and Williams (1993b) considered ethnicity in terms of three
dimensions: (1) ancestral origin, (2) generational status, and (3) levels of
acculturation.

Another example of operational inadequacies is residency. Klobus-
Edwards (1981) used residency to determine whether respondents lived
(resided) in Black or racially mixed neighborhoods. West’s (1989) approach
provided a spatial/regional context, while Taylor (1992) defined residency as the
amount of time the respondent has lived in the city.

In addition to variable operationalization issues, Meeker, Woods & Lucas
(1973) mentioned the effect that past injustices and discrimination had on

Blacks’ recreational behavior. West (1989) noted that discrimination was not

15

looked at as a variable affecting use of public spaces by minorities (Blacks in
particular). Discrimination, therefore, was added as a theoretical explanation for
minority underparticipation in leisure settings:

These two paradigms - marginality and subculture theory - have

dominated the thinking and research about minority

underrepresentation in outdoor recreation. However, it is a strange
irony that this body of research has almost entirely ignored another
important potential explanation: the problem of interracial relations

and prejudice (West, 1989, p. 12).

West uses "subcultural theory” in lieu of ethnicity theory because he
believed that Washbume’s initial conception and theoretical assumptions
(regarding ethnicity) were closer to the conception of subcultural life
preferences. West’s measure of subcultural preference was operationalized as
"no interest“ or “prefer to do other things”. In other words, preference or
non-preference indicated ethnicity factors. However, West noted that the
"measures of subcultural influence [were] less robust than other measures”
(1989, p. 17). This indicates that a more complex measure of cultural
preference is needed.

Although West did measure discrimination, he did not consider the
possibility of an interaction between subculture and marginality. Conversely,
Hutchison (1987) mentioned that Black and White differences could be due to a
complex interaction between race and class, but he did not measure the
interaction, nor did he consider the role of discrimination. Lack of attention on
the nature of these relationships is due to a confluence of assumptions

underlying the research variables, an ahistorical approach to dealing with race

and ethnicity, and inadequate measures of the concepts under study.

16

The As um tion R a in R si no

Regardless of how residency has been operationalized in previous
studies (see earlier note), Stamps and Stamps” (1985) assumption that
residential integration facilitates social integration has influenced the literature
on minorities and recreation. The assumption states that Blacks who live in
White or racially mixed neighborhoods will have similar recreation behaviors as
Whites. Although this assumption has existed prior to Stamps and Stamps’
mention of it, it has created two problems. The first problem is that it assumes a
normative order of White recreation. Therefore, if subcultural groups are not
participating in the normative order, they are considered to be under-
represented. Johnson, et al. (1997), mentioned that "Woodard (1988) and Carr
and Williams (1993) [argued] that subgroup recreation should be considered in
terms of the respective ethnic group standards rather than in relation to a
normalized ideal" (p. 4).

The second problem is that the assumption (of equating residential
integration with social integration) is contradictory to other studies specifically
researching the issue or racially mixed neighborhoods. From the Detroit Area
Study of 1978, Whites and Blacks were asked questions concerning
neighborhood racial composition:

it [became] very clear that whites still harbor substantial prejudice
against blacks as potential neighbors, and that their tolerance for
racial mixing is really quite limited This fundamental disparity
between blacks and whites has been confirmed by surveys conducted

[in 1991] in Milwaukee, Omaha, Cincinnati, Kansas City, and Los

Angeles, all of which show that blacks strongly prefer a 50-50 mixture

and that whites have little tolerance for racial mixture beyond 20%
black (Massey & Danton, 1993, pg 93).

17

Furthermore, simply because Blacks are in a racially mixed
neighborhood, it does not signify that they are particularly welcomed.
Historically, so-called "neighborhood associations" became popular in the
19603. Their main purpose was to keep blacks out of the neighborhoods. By
1965, Detroit alone had "at least 192 neighborhood organizations throughout the
city,” founded by Whites (Sugrue, 1996, p. 211). This could hardly be called
welcoming. Perceptions and feelings towards Blacks, because of these
organizations, instill values for the next generation, thereby perpetuating a cycle
of racial discrimination and hostility.

Perhaps a more appropriate measure of how well Blacks and Whites live
together is their sense of belonging on the block. McMillan and Chavis (1986)
identified "membership” as a critical element or criteria for inclusion (both in the
neighborhood setting, specifically and in society, generally). Rather than
inferring from the mixture of the neighborhood (is a respondent in a racially
mixed neighbomood or in a predominantly Black or White neighborhood), one
should consider asking the respondent whether or not they feel they belong. If
there is a strong sense of belonging among minority members of the
neighborhood, than minorities are more likely to use surrounding neighborhood
parks and participate in those public parks.

There are other reasons why sense of belonging in a block is a better
measure than place of residence. In a neighborhood with a strong sense of
belonging, neighbors are more apt to interact and know each other. This leads

to knowing people on the block and less crime because people are more likely to

18

watch out for each other. Less crime would allow people to interact more in
public places, such as nearby parks.
Bag sng Ethnic'fly: Is a Distinction Neadfi Between the Two?

In the recreation literature, perhaps the most critical argument that has
surfaced is the need for distinction between race and ethnicity. Hutchison’s
(1988) critique of the race/ethnicity literature regarding recreation research was
an excellent exposition on the inadequacies of the marginality and ethnicity
perspectives. In particular, he noted that race and ethnicity "have been used
interchangeably without careful definition of either concept, even though the
implied association between race and ethnicity contradicts research in race and
ethnic relations“ (Hutchison, 1988, p. 11). Hutchison, a sociologist by training,
makes a critical observation: some recreation researchers are ahistorical in that
they do not research enough of the wealth of information available on race
relations in the sociological literature to provide a proper context.

However, the question which needs to be asked is: does it matter? The
answer is yes and no. It matters that recreation researchers familiarize
themselves with such things as Jim Crow laws, (racial) restrictive covenants, and
developments in federal and municipal policy which helped to reinforce racial
cleavages in urban settings. 'By 1970, after two decades of urban renewal,
public housing projects in most large cities had become black reservations,
highly segregated from the rest of society and characterized by extreme social
isolation" (Massey & Danton, 1993, p. 57). Given this pattern, one has to

question whether or not Blacks have similar social patterns as compared to

19

Whites. Having a sense of the history of race relations might provide some
insight into why Blacks don’t want to participate in the same public recreation
space as Whites.

Race relations has formally been studied since the late nineteenth
century. While it’s initial focus was on Black and White relations, it has
expanded its scope to include other minority groups. A cursory glance at the
vast race relations literature will show how sociologists have changed their
thinking about the race ”problem."

Blacks were initially looked at as biologically inferior to Whites. Much of
the earlier work was influenced by the Eugenics Movement (this movement
espoused the theory that the human species can be improved by controlling for
heredity) and Social Darwanism. At the turn of the century, sociologists began
to look at blacks as culturally inferior when biological arguments began waning.
From the 1920s, Robert Park and his students from the Chicago School of
Sociology legitimated “the idea that assimilation [italics added] was the inevitable
end point of race relations” (McKee, 1993, p. 110).

Race “problems“ turned to race ”relations.” After World War II, Gunnar
Myrdal’s An Amsn’ssn Dilsmms transformed race relations into a moral issue.
Myrdal chose to title his voluminous work "An American Dilemma” because he
felt it captumd ”his sense of the race problem in the United States: a conflict
between the highest of American ideals and lower parochial interests and

prejudices” (McKee, 1993, p. 226).

20

In particular, Myrdal noted that it was not so much a black man’s problem
as it was a white man’s problem. Myrdal, however, still supported the
assimilationist theory. It wasn’t until the post civil-rights era that this
assimilationist perspective began to change into a pluralistic, multicultural
perspective of race relations; the reasons for which are beyond the scope of this
study.

As mentioned eariiar, it is important that leisure scientists understand the
development of the race ”issue," and put it in its proper context when
researching racial/ethnic matters. However, it is not as important to distinguish
between race and ethnicity for empirical results to be relevant. Race groups are
defined, in the sociological sense, in terms of their phenotype (e.g., hair texture,
skin color, and facial features). Ethnic groups are defined as identifying with a
particular culture (real or imagined). Although Hutchison (1988) points out that,
historically, there is a difference between the two terms, he failed to note that
within the discipline of sociology, there are two schools of thought on this issue.

Theorists who have compared race and ethnicity may be divided

into two camps: those who underscore the likeness of the concepts

and those who view them as sufficiently dissimilar to warrant

distinct perspectives and policies. To proponents of racial and

ethnic similarity, distinction between the two concepts are

frequently seen as based on definitional differentiation. In other

words, distinguishing between the two concepts has validity only

as an exercise in differentiating between what are merely

theoretical constructs. Distinctions between Race and Ethnicity

therefore, it is argued, are untenable and without empirical support

(Singh, 1981, p.2).

This supports the view that racial and ethnic groups could be viewed

together. While Blacks (a racial minority group) are different from other (ethnic)

21

minority groups, given their slavery past, they are similar to other minority
groups in that they are marginalized (economically, politically, and socially) in
some fashion. Again, the question is: does it matter?

This study contends that it is much less important to stress the difference
between race and ethnicity (unless one is researching specifically a race versus
ethnicity issue), than it is to stress the importance of the difference between
minority (marginalized) groups and majority (dominant) groups: Whites and non-
Whites. Given current demographic changes, identified by Dwyer (1993), it will
be increasingly important to look at minority groups (regardless of racial or
ethnic makeup) and how they differ from mainstream (majority) America, with
respect to recreation behavior. The central concern, therefore, should not be
one of definition, but rather one of applicability and practicality.

Following this argument, one has to question whether or not Blacks have
their own culture, which would make them both a racial and an ethnic group by
current definitions. Singh (1981) comments that ethnic groups "may share a
similar language, religion, and/or history" (p. 1). Residential segregation and
social isolation has caused Black speech patterns to differ from Standard
American English.

[Black English] evolved independently from Standard American

English because blacks were historically separated from whites by

caste, class, and region [Black English] has become

progressively more uniform across urban areas. Over the past two

decades, the Black English vemaculars of Boston, Chicago,

Detroit, New York, and Philadelphia have become increasingly

similar in their grammatical stmcture and lexicon, reflecting urban

blacks’ common social and economic isolation within urban
America (Massey 8. Danton, 1993, pp. 162-163).

22

The ramifications of a separate dialect are tremendous. If Black English
is uniform and separate, it creates barriers to communication between Blacks
and Whites, thereby increasing less interaction; even in a racially mixed
neighborhood. Furthermore, it impedes Blacks’ socioeconomic progress
because it will be harder for Blacks to find a job if they do not speak Standard
English well. Wilson (1996) noted that "employers frequently mentioned
concerns about [Black] applicants' language skills and educational training" (p.
116). It will also be more difficult for Blacks to do well in school when they have
to read texts that are in Standard English. It has been found that successful
Blacks “who have grown up in the ghetto literally become bilingual, Ieaming to
switch back and forth between black and white dialects depending on the social
context" (Massey & Danton, 1993, p. 165). In addition to language, Blacks share
a similar history (a.g. slavery) and a similar religion (e.g. spirituals).
lnadsggsts Mfisugs sf Rsce and Ethnicmr'

Given the above discussion, it is easy to see why the categorical scale is
inadequate for measuring a complex variable such as ethnicity or race. Race
and ethnic categories have been too limiting. Although recreation researchers
have classified respondents as Black, White, or Hispanic, they do not know the
extent to which the respondents (pertaining to a particular category) identify with
their respective culture(s). For example, how afrocentric is the Black
respondent? 'Afrocentricity espouses the belief that images and symbols
derived from African and African American life experiences are necessary and

appropriate guides for people of African descent” (West, 1994, p. 28).

23

Therefore, the more afrocentric Blacks are, the more likely they are to be
exposed to African American literature, music, history, and dance. Conversely,
the less afrocentric Blacks are, the more likely they are to reflect mainstream
values and participate in mainstream pursuits. A more complicated measure of
cultural identification (versus a categorical approach) is needed in order to
gauge the extent to which respondents identify with their respective cultures.
Movi_ngForwa_rd

The Jsumal of Leismre Research dedicated an entire issue (1998, volume
30) to minority recreation and leisure. This special issue is the most current
collection providing a critique of previous research. Of particular interest is the
introduction by Myron F. Floyd and the article by Karla Henderson. Floyd (1998)
noted:

Perhaps the most critical issue currently facing the race and ethnic

studies literature is the absence of viable theoretical frameworks.

In general, little systematic thought has been directed at race and

ethnic issues three theoretical explanations have historically

been used the marginality hypothesis, the ethnicity or

subcultural hypothesis, and perceived discrimination

scholarship on race, ethnicity and leisure will advance if past

approaches are reconsidered and alternative theoretical and

methodological possibilities are explored (pp. 4-5).

Henderson (1998) noted:

One of the problems researchers encounter is the tendency to

examine subjects or participants as simply one characteristic when

all of us are more complex than simply being female or white or

able-bodied (p. 159).

The suggestion by both researchers is that there is a need for (1) more

complex forms of measurement, and (2) an integrative approach to reflect the

interrelationships between these different dimensions.

24

Alternative A ches to tud in Mino ' Recreation

Qpnstrug VSEQS a variable approach. A call for new approaches has
been issued by researchers. One such approach is to look at a construct
(conceptual) approach versus a variable approach in order to understand the
relationships between subcultural differences and public recreation participation.
In order to synthesize past research, this investigation looked at the variables
which were used to measure particular phenomena related to ethnic recreation.

In addition to previously used variables, underlying constructs are
normally mentioned by researchers (or assumed) but never specifically
addressed. Table 1 illustrates major concepts and variables which were
addressed in the 1980s and 19903.

From Table 1, one can see the several concepts which were researched.
However, these concepts and variables were typically measured so that one can
see the nature of the relationship among (unidimensional) variables and the
dependent variable: participation. The problem with this approach is that it does
not take into account influences which the independent variables exude on each
other, Le, a multidimensional approach. As a result, researchers often fall short

of understanding or identifying a minority recreation phenomenon.

25

 

Table 1

ConceptsNariables Addressed in the Literature Regarding Ethnicig/Race and
Publis Recreation Participation

 

 

ConceptsNariables Researchers (Year) (Approach to Concept)

 

Marginality Klobus—Edwards (1981)
Stamps & Stamps (1985)
Woodard (1988) (classism)
West (1989)
Taylor (1992)
Floyd, et al. (1994) (subjective)

Race/Ethnicity Klobus—Edwards (1981)
Stamps 8 Stamps (1985)
Hutchison (1987)
West (1 989)
Taylor (1992) (ethnic identification)
Carr & Williams (1993) (3 dimensions of
ethnicity)
Chavez (1993)
Floyd, et al. (1993) (ethnic identification)
Floyd, at al. (1994) (race)

Sex/Gender Edwards (1981)
Woodard (1989)
Taylor (1992)
Residency Edwards (1981) (black or mixed
neighborhoods)

West (1989) (spatial regional context)
Taylor (1992) (time living in city)

Age Woodard (1988)
Chavez (1993)
Floyd, et al. (1993)

Discrimination Woodard (1988)
West (1 989)
Chavez (1993)

 

26

Reconsidering Past Approachessnd ExplonligAltemstive Theoreticail
@nsiderations: Pre-cursors to a Model. This section identifies the undenying
dimensions (whether measured or implied) of previous research. A critique of
assumptions and their consequences will also be offered when necessary. This
section is not meant to be exhaustive. As such, selective studies which have
influenced the scholarship of racial/ethnic recreation research will be considered
for review. The studies are considered chronologically to illustrate development
of scholarship. Lastly, conceptual models are provided to identify constmcts
from past research. The reader should keep in mind that these conceptual
models have been created by the current investigator, not by previous
researchers, for the purpose of illustrating relationships between said constmcts.

Prior to Washbume’s (1978) study, Lindsay and Ogle (1972) and Craig
(1972) studied Black recreation patterns in relation to urban areas. Craig
identified the distance/cost factor as a major impediment to Blacks’ leisure.
Lindsay and Ogle mentioned physical availability as a factor, in addition to cost,
and introduced the opportunity theory (a precursor to marginality theory).

Opportunity theory (as cited from Hendee) was defined by Lindsay and
Ogle as suggesting “that groups long denied opportunities, not only by virtue of
their residence, but because of poverty, ignorance or segregation, might become
participants in available opportunities if these barriers are removed” (1972,

p. 20). Lindsay and Ogle believed that if one were to control for cost, than there

should be no difference in opportunities between low and high income groups,

27

and participation in recreation activities would be equal. Their conceptual model

is reflected in Figure 1.

Figure 1. Lindsay & Ogle’s (1972) Conceptual Model

 

 
     
   

Participation in
Recreation

Opportunity

 

 

 

Figure 1 indicates that one’s socioeconomic status (age, income/cost,
education)‘ affects one’s opportunity (distance, ease of access) for recreation,
which increases the likelihood of participating in a recreation setting.
Socioeconomic status also directly affects recreation participation. There is no
mention of race/ethnicity as having an effect. The study looked at low versus
high income respondents, and users versus nonusers.

Lindsay and Ogle’s approach to opportunity theory was unlike Hendee’s
(1969) conception of opportunity theory (cited earlier). Lindsay and Ogle said
that “participation rates in outdoor recreation vary directly with both cost and
physical availability of outdoor recreation" (1972, p. 20). This definitional
approach varies greatly from Hendee’s definition. For example, segregation

(discrimination) issues were not considered, nor were issues related to being

 

When explainlno figures In text. constructs wlll appear In Italics followed by the variable(s) In parentheses
which reflect the construct. '

28

ignorant of opportunities for recreation. In addition, Lindsay and Ogle’s sample
was mostly white.

In Craig’s (1972) study, his respondents were all southern Blacks. The
reason for Craig’s inclusion in this selective review is that he offers insights that
many researchers have forgotten (or chosen to ignore). Craig (1972) noted the
following:

white people, by their very efforts “to keep Negroes in their

place" have done their utmost to prevent the emancipation of their

leisure (p. 108)

the effects of past experiences on immediate recreational

behavior of Negroes may be greater than the effects of an

increase in income or change in location (p. 111).

Historically, recreational experiences of Negroes living in the Deep

South have been extremely meager and limited. The important

restraining variables were rigid segregation policies, low income,

and lack of leisure time (p. 114).

Aside from the fact that Craig’s arguments support the view that Blacks and
Whites do not have similar leisure patterns, he argued for the ethnicity
perspective, and pointed out that racial identification (not simply race) is an
important variable to consider.

In Figure 2, culture (White attitudes, plantation system) affects attitudes
and behaviors associated with racial identity (racial identification, feelings of
segregation). Racial identity has an affect on rural-urban residency and income.
These constructs have an affect on recreation participation. Figure 2 reflects
Craig's focus on suburban-urban Black differences. Craig made the point that
Blacks define themselves in terms of White’s attitudes towards them.

The reader has to put this study in context, however. The study took

place in the South, in the early 1970s, and Blacks were ”made” urban once the

29

Figure 2. Craig’s (1972) Conceptual Model

 

    
    
   
   

Rural/Urban
Residency

  
 
  
 

Attitudes

Towards
acial lde/ntily/

   
 
  

Recreation
Participation

 

Income

 

 

 

area they lived in ”became" urban. In other words, there was no choice about
living in an urban area. Craig showed the change from what was a rural area to
what became an urban area. That is hard to find today.

Rather than looking at how urban and suburban residency affects ethnic
behavior, one should consider how ethnic behavior differs between urbanites
and suburbanites. In today’s world, a comparable study would look at what
happened to Blacks in urban areas when those areas were "made" inner-city
areas. The pattern of rural to urban to inner-city residential stratification is
reflective of eariier comments regarding residential segregation. Using
secondary data, Washbume (1978) used a sub-sample technique that matched
Blacks and Whites for the same characteristics. Results, however, should not
have been generalized to the larger population. Figure 3 illustrates the

conceptual model from Washbume's study.

30

Figure 3. Washbume’s (1978) Conceptual Model

 

.——'—“
/ \

\
r/ \\
\ w Perceived /
\ \ \ Segregation /

/

\

\

. . __ ,, L 7 .. Participation
Marginality in Recreati-

In Figure 3, marginality (age, income, education, sex, residency, access,

 

 

 

opportunities, poor transportation) affects participation (based on participation
rates). Marginality was compared between Blacks and Whites. The portion of
the model surrounded by the dotted-line circle was mentioned by Washbume,
but never empirically measured in his study. This is Washbume’s greatest
contribution to the scholarship on ethnic recreation studies, and his greatest
shortcoming in his study. As a result, variables regarding preferences, social
organization, and values were mentioned, but not examined.

Klobus-Edwards (1981) provides one of the most elaborate conceptual
models to data. Figure 4 is extremely complex. She considered the major
factor to be race, residency (location in racially mixed neighborhood or in all
Black neighborhood), sex, and marginality (age, education, household income,
number in household). Motivation for recreation participation (physical

conditioning, competition, relaxation, personal contacts, enjoyment, family time

31

[not looked at]) and barriers (lack of interesting programs, inadequate
information, transportation problems, amount of time available for leisure

activities) were intervening factors.

Figure 4. Klobus-Edwards’ (1981) Conceptual Model

 

Ma inal'
Race In W

Motivation .

Leisure
Participation

     
  

 

 

 

 

Leisure participation was a composite dimension made up of recreation
behavior (outdoor recreation, use of public recreation facilities, member in
recreation associations) and recreation preference (willingness to participate in
public sponsored activities). Recreation preference was used as a measure of
leisure aspirations. An underlying assumpfion by Klobus-Edwards was that
resources are located nearer to White areas, therefore Blacks living in racially
mixed neighborhoods will participate fuller in recreation activities.

The concepts of residency and motivation are introduced in a different

manner in this model. Motivation reflects intrinsic reasons for wanting to

32

participate in leisure activities. Klobus-Edwards controlled for marginality
variables in her study, and found that race explained between 9 and 26% of the
variance in recreation involvement.

She argues, but never tests, for a relationship between ethnicity and
marginality perspectives (note that an arrow is missing between race and
marginality). Accessibility to the recreation site (distance from place of
residency) was not used as a variable. Transportation was used as a variable,
but only as an inhibiting factor. The indicas created to reflect leisure
participation behavior and preference were activity-related indicas, not behavior-
related. Knowing what one participates in does not necessarily tell one how
one feels toward an activity or towards leisure in general.

Stamps and Stamps” (1985) study examined the relationships of race and
class to leisure activities of urban residents. In their study, there were two major
assumptions. The first was elaborated upon earlier: social integration occurs
with residential integration. Their second assumption was that social class
status determines ones life style (leisure participation was used as a proxy for
life style). Figure 5 illustrates Stamps and Stamps’ conception of the interaction
between the different constructs. As shown, the arrow from social class to
leisure participation reflects their second assumption.

Social class (education/occupation) status was determined by an index of
social positioning (Hollingshead Two Factor Index) which created a five class
index from education and occupation variables, and which Stamps and Stamps

later collapsed into upper and lower class.

33

Figure 5. Stamps & Stamps” (1985) Conceptual

 

     
 
     
   

Model
Social
Class \
Leisure
Participation
Ethnicity] -
Race

 

 

 

Figure 5 is reflective of the traditional paradigm established by Washbume in
1978. Stamps and Stamps concluded that, as explanatory variables, social
class and race were still inconclusive as predictors of leisure activities.
However, they found (contrary to what they expected) support for race, over
class, as the more important predictor of leisure activities.

Although they tried to approach social class in a new manner, ranking
someone’s choice of leisure activities is not thly reflective of their lifestyle.
lnfonnation came from a quality of life survey. A shortcoming in their study was
that it was not a survey measuring attitudes or behaviors (as they relate to
ethnic/racial or socioeconomic status) specifically, but rather they inferred from
participation rates and preference based on rankings.

As such, Stamps and Stamps recommended that future studies use more
sophisticated multi-variate analysis of specific leisure activities. In addition, they
recommended that leisure researchers abandon attempts to predict leisure
participation from demographic variables. This further reinforces the need for
more advanced models of analysis, as well as, reconceptualization of the

unidimensional race construct.

In contrast to Stamps and Stamps, and previous studies, Hutchison
(1987) argues that a group’s social composition is more important for
understanding differences and similarities between Blacks and Whites than
looking at the specific activity they participated in. Figure 6 illustrates
Hutchison’s conceptual model.

In Figure 6, Hutchison looked at the influence that one’s ethnicity/race
(Black, White, Hispanic) has on values (individual Versus family), social
composition (peer groups, family groups, individual, sex composition, age
composition, size of group), and leisure activities (mobile, stationary, sports
activities). Hutchison’s study was different in some aspects. For example, he
utilized a field observation approach instead of a questionnaire. The unit of
analysis were not actual activities participated in, nor rates of participation, but

rather the nature of the activity.

Figure 6. Hutchison”s (1987) Conceptual Model

Ethnicity! .....
Race

 

 
 

Values

   

Leisure
Activities

    
 

   

Social
Composition

 

 

 

 

Hutchison suggested “that black-white differences, if they exist, are due
not to the simple influence of either class or race, but to a more complex (and
not well understood) interaction between these two factors” (1987, p. 208).

However, Hutchison neither tests this relationship, nor does he consider the role

35

of discrimination. The purpose of his study was to illustrate the advantage of
observational techniques in supplementing quantitative data.

Although Figure 6 illustrates observational techniques, there are some
concerns regarding this technique. Hutchison, using Hispanic groups in his
study, mentioned that distinctions must be made between different Hispanic
groups (Mexicans were mostly looked at in his study). In addition, in
observational data, it is much more difficult to distinguish dark-skinned Hispanics
form Blacks. Therefore, group activity could seem to be intra-ethnic or intra-
racial when, in fact, it is Hispanic. Also, generational issues could be further
explored. If one assumes that the dominant culture exudes influence, then
second and third generation Hispanics should be closely aligned (in their
recreation behavior) to Whites, than their immigrant and first generation
counterparts.

West (1989) noted that discrimination affects marginality and subcultures
(racial/ethnic groups). West’s study took place in Detroit and data was collected
via phone interviews. A critical contribution by West was that he operationalized
the concepts of marginality (limitations on participation due to limited income or
limited access to transportation), subculture (difference in unconstrained
preference by minority groups, i.e., "no interest” and "prefer to do other things"),
and interracial factors (operationalized from open-ended questions and
collapsed into five categories: racism, racial-anti-white, racial-anti-black, race

relations, and uncomfortable/u nwelcomed) (West, 1989, pp. 16-1 7).

36

Figure 7 illustrates how West incorporated discrimination and preference
into the traditional paradigm (see Figure 5). Race (Blacks and Whites) and
marginality (subjective and objective measures of income, transportation) have
an effect on both discrimination (interracial factors (perceived or real), and use of
parks (use/non-use, frequency of use). Race also affects subcultural preference
(no interest, prefer to do other things), which in turn affects use of urban parks.

West, however, called for a more robust measure of subcultural preference.

Figure 7. West’s (1989) Conceptual Model

Subcultural

-- Preference
\ Urban
Use

Marginality

 

     
    

 

 

 

West (1989) found that Whites” use of the tri-county parks and frequency
of use of tri-county parks in Detroit was significantly greater than Blacks’ use of
tri-county parks. In addition, he found that marginality played a small role in
minority underrepresentation, and found no support for Black subculture as a
reason for underrepresentation in Detroit’s tri-county parks. This opened the
door for alternative explanations. West found that transportation was perceived
as a major barrier. It was also surmised that interracial relations may be a

contributing factor in inhibiting use of regional parks by inner—city minorities in

37

Detroit. Furthermore, West raised the issue of perceived geographic distance
(subjective and objective measures) as a possible factor influencing urban park
usage (West, 1989, pp. 18-25).

West’s contribution to minority recreation studies is invaluable because
he underscores three major initiatives. First, although he found no support for
subcultural explanations, he maintains that Blacks are still under represented
because of other factors that have not been measured, or have not been
measured adequately. Such factors could be physical-structural (location),
political (policies regarding or restricting use), economical (lower discretionary
income), or cultural-structural (education system, political system, residential
patterns and housing issues, economic base for infrastructure improvements in
inner-city). The point is that marginality and ethnicity are not explanatory
enough, and they do not operate in a vacuum.

Second, West noted that the policy problem is not one of "access to
recreation areas by minorities, but rather the quality of recreation facilities and
services” (1989, p. 25). Policies regarding urban parks need to focus on
bringing the parks to the users (inner-city Blacks in this case), not bringing the
users to the parks. One need only take a cursory look at conditions in inner-city
neighborhoods to see that there have been disinvestments in inner-city parks (in
part due to crime). Perhaps new initiatives involving coalitions with grass-roots
organizations and police foot and bike patrols may be in order.

West’s third contribution was a challenge to leisure researchers "to face

more directly the issues of race relations"(1989, p. 25). It took a decade after

38

Washbume’s study to incorporate the issue of discrimination into minority
recreation studies; yet another reason why the recreation researchers
(investigating minority issues) should be familiar with the literature on race
relations.

Taylor (1992,1993) advanced the current understanding of the ethnic
recreation literature (Figure 8). Her study was based on personal interviews of
the people and urban parks in New Haven, Connecticut in 1989. She
considered several recommendations of previous recreation researchers. Taylor
considered Woodard”s assertion that “the problem with the ethnicity model is
that is assumes that ethnic groups have a single value and normative order,

[thus], there can be no variation in leisure among [said groups]'” (1988, p. 88).

Figure 8. A Model of the Factors Affecting Ethnic Leisure Participation

 

 

 

 

 

 

 

 

 

 

 

    

 

  
   

 

 

 

 

 

 

 

 

 

 

 

 

 

Social -_.._,-_--fl-.,,.. _. LL
___> Class ‘* , ’
Race—t— , ‘L .....
L: Participate
In Ethnic
gig—mm] j‘i
_ I m “i-Tmem ’
Life Cycle , A ,7. “' ' '[Sex in
Marital Status ' . 'I { Community
Age -_.,, ,,,-_,-,. ‘HTTTT ‘*
I_r____ __,.,..L,,,-_m ,v,,* is, 4
Source: Taylor, D. E. (1992). Identity in ethnic leisure pursuits.
San Francisco, CA: Mellon Research University Press, p. 36.

 

 

 

39

She accomplished this by looking at both racial and ethnic groups
(showing clear support for distinguishing between the two concepts). The racial
groups consisted of Blacks and Whites. Ethnic groups were Jamaican and
African Americans for Blacks, and Italians and "other whites" for Whites. This
allowed for a comparison between and within groups.

Taylor (1992) created a more complex model. She called it the
"differential access model” (see Figure 8). As Figure 8 indicates, ethnic leisure
participation is a very complex phenomenon. The impact of this model is greater
than it seems. In prior studies, differences, which were not explained by
marginality variables, were said to be explained by ethnicity variables, and vice-
versa. Taylor's model illustrates what later researchers have been asserting: the
factors influencing ethnic leisure participation need to go beyond marginality and
ethnicity theories, and consider other factors.

Taylor (1992) addressed the shortcomings of her book. She mentioned
that her next project was to accomplish two things: ”(1) refine and test the model
developed in [her] project, and (2) attempt a more sophisticated analysis of the
questions raised in [her] book" (Taylor, 1992, Preface). Indeed, the model is
rather cumbersome. However, it is a start. In her model, Taylor did not address
the role of racism and discrimination. In addition, her variables could have been
subjected to a factor analytic approach to verify the existence of her constructs.

With respect to her eartier consideration of subcultures, Taylor noted the
following:

The results of this study do not support a subcultural thesis
because there were important and striking differences between

40

African Americans and Jamaicans, the two Black ethnic groups

being studied. It the Black leisure subculture theses were true,

such striking within-group (Black) differences would not be found

(1992,p.259)

In the traditional view of the Black leisure subcultural thesis, the thesis assumes
all Blacks have a normative homogeneous subculture. However, this thesis
considered African Americans and Whites, not ethnic groups who happened to
be considered Black.

The thesis assumes that African American Blacks are a subculture of
Whites. Because of their phenotype, White Americans consider Jamaicans
Black. It does not mean, however, that they share the same values,
experiences, and history with African American Blacks.

The fact that there were differences between Jamaican Blacks and
African American Blacks proves subculture (a subculture within a subculture).
Taylor makes the same mistake other researchers do when lumping Cubans,
Puerto Ricans, Mexicans, and Panamanians under one category: Hispanic. One
has to keep in mind that racial terms such as Black and White, and ethnic terms
such as Caucasian, Asian, or Hispanic are socially constmcted terms for
categorical purposes.

Taylor was looking at two distinct cultural groups. Perhaps a better
method of disproving the subcultural hypothesis is to look at those who identify
themselves as African American Blacks. Perhaps the subcultural thesis needs
to be refined or renamed. Taylor’s use of subculture more appropriately

considers sub-racial , not cultural differences. If differences between Blacks and

Whites don’t hold true, then the sub-racial thesis would also not be supported.

41

In response to the restriction of scope of the marginality-ethnicity
paradigm, Floyd, et al. (1993) incorporated the marginality-ethnicity theory within
a broader context: assimilation theory. "The assimilation perspective is broad
enough to encompass the marginality-ethnicity paradigm and can also
accommodate a perceived discrimination dimension” (Floyd, et al., 1993, p. 84).
This was an attempt to integrate concepts from Washbume (1978) and West

(1989), and is reflected in Figure 9.

Figure 9. Conceptual Model of Explanations of Ethnic Variation in
Recreational Behavior

A c an . Cultural Subcultural
CUE omen” @ \

Structural ___ Socioeconomic

Marginality _.,
Behavioral
Intergroup

Receptional W- . h-

@

Source: Floyd, M.F., Gramann, J.H., 8 Saenz, R. (1993). Ethnic factors and the
use of public outdoor recreation areas: The case of Mexican Americans.

W15 (2), p. 86.

 

   
 

RECREATION
BEHAVIOR

 

 

 

Floyd, et al., (1993) also noted that past approaches involving ethnicity
and recreational behavior focused on Black and White individuals. Their study
used Mexican-Americans as their research population. The data collection
technique used was telephone interviews.

There were three major concepts which came from this study that

contributed to the growing literature regarding ethnicity and recreation behavior.

42

The three dimensions were three of seven subprocesses identified by Gordon in
1964: acculturation, stnrctural assimilation, and behavioral receptional
assimilation. These three dimensions respectively correspond to the three types
of social distance: cultural distance, socioeconomic distance, and intergroup
distance (Floyd, et al., 1993, p. 85).

In the current study’s use of acculturation, although minorities are
”acquiring“ the dominant group’s cultural characteristics, they continue to
maintain their own cultural characteristics as well. The dominant culture exerts
pressure on the minority culture to move from acculturation to full assimilation,
but the process of assimilation is tempered by the minority’s level of exposure to
the dominant culture. Acculturation is therefore closely related to Floyd, et al.’s
concept of ”cultural distance” (M)-

Cultural distance,(acculturation) was measured by a respondent’s
language preference for home, television, radio, and literature. Socioeconomic
distance was measured by level of education. Intergroup distance (perceived
discrimination) was measured by asking questions about feelings of acceptance
at the recreation site, other ethnic groups at the recreation site, and hassle from
police at the recreation site.

Floyd, at al., acknowledged procedural shortcomings. Figure 9 illustrates
Floyd, et al.”s conception of the factors affecting ethnic group participation in
recreation activities. The perceived discrimination factor was found to be
insignificant. Floyd, et al., recommended that future research develop measures

of discrimination that correspond closely to visitation at specific areas. Although

43

their study could have had stronger measures/statistical design to incorporate
relative causality by use of the coefficients, their study is an important
contribution because it reoonceptualized the marginality-ethnicity paradigm

within a larger perspective: assimilation theory.

Chapter 3
METHODOLOGY
Movin Forw : Th D v I m nt of a Model
Given the aforementioned conceptual review, the current researcher
seeks to create a model which identifies major concepts, and considers how
these concepts are (inter)related. Figure 10 illustrates the relationships between
the following factors: Acculturation (ACC), Socioeconomic status (SES),
Subcultural identity (SID), Benefits of Recreation (BEN), Perceived

Discrimination (DIS), and Recreation Participation (REC).

Figure 10. The Ethnicity and Public Recreation Participation

 

 

Model
“a
Perceived
Benefits of _
Recreation ,
Recmatron

H Participation
Perceived
Discrimination

 

 

Subcultural
Identity

 

 

 

The Ethnicity and Public Recreation Participation Model (EPRPM)
illustrates several of the components reviewed in the literature. The EPRPM is a
causal model. It reflects the literature’s support of these concepts and how each

concept relates to others. It is important to understand the order of causation.

45

Davis (1985) supplies some rules concerning the logic of causal order (for two
variables X and Y, distinguish among three possibilities):

(1) X—>Y, X might influence Y but Y does not influence X.

(2) X F’ Y, X and Y might influence each other.

(3) XHY, X and Y might correlate highly, but we do not know
directionality (p. 10).

There are three important things to remember. First, the directions of the
arrows in a causal model indicate "potential” flows of causation, not necessarily
actual flows. Second, the most important principle of causal order is that ”after
cannot cause before there is no way to change the past one way arrows
flow with time" (Davis, 1985, p. 11).

The last thing to remember about causality is the following:

most methodologists agree that causal order is a substantive or

empirical problem to be solved by our knowledge about how the

real world works, not by statistical gyrations. At the philosophical

level this has positive and negative implications. On the negative

side, empirical research is always hostage to empirical

assumptions that might be wrong; on the positive side, computers

cannot substitute for sociologists in analyzing data, because

computers do not know anything about the real worid and

sociologists do know a little bit (Davis, 1985, p. 11).

Given this insight into the development of causal relationships, one can now
consider Figure 10 in its entirety.

Acculturation is believed to measure one”s level of assimilation into the
dominant culture. The causal arrows extend to both socioeconomic status and
subcultural identification. The underlying assumption is that acculturation
begins when one is born and the level of assimilation changes over time.

Acculturation differs slightly from Floyd, et al.’s (1993) conception. Rather

than occurring simultaneously with socioeconomic variables (see Figure 9) and

46

subcultural identification, it is conceptualized, in the current study, as a
precursor to SES and SID.

The stronger one’s level of acculturation (cultural distance) the stronger
one will identify with a majority culture. Conversely, if one”s level of
acculturation is weak, then the socioeconomic status will be lower. For example,
if one speaks fluent English, the prospect of a better paying job will be
increased. This causal relationship reflects acculturation theory.

Following the arrows, one can see a direct relationship between
socioeconomic status and recreation participation. This relationship is
hypothesized by the literature. It reflects the marginality theory. The arrow from
subcultural identity to recreation behavior reflects the ethnicity theory.

Both socioeconomic status and subcultural identity have two intervening
constructs which are perceived discrimination and perceived recreation benefits.
BEN is used as an intervening variable because it was thought that the
perceived benefits of a recreation activity or recreation site will affect actual
participation in the activity or at the recreation site. The ”perception of
recreation benefits” construct incorporates KIobus-Edwards” (1981) suggestion
of integrating ”motivation for recreation participation” as a precursor to actual
participation. The assumption is that motivations or perceived benefits derived
from participation in a recreation activity (or going to a recreation site) is based
on one”s socioeconomic status and cultural perspective.

For example, there are two local parks, and one has a football game and

the other has a baseball game. If you ask a Puerto Rican which he/she would

47

prefer to go to, the odds are that helshe would pick the baseball game over the
football game because football is not a national sport in Puerto Rico, and
baseball is a national sport. Following the causal model, the motivation or
benefit (watching baseball) of going to the park (passive recreation) is
influenced by one’s culture (Puerto Rican).

Conversely, if one had the economic means to attend a professional
baseball game (instead of the local baseball game), or considers the attendance
at a professional baseball game to give one certain social status, than
socioeconomic status would affect the perception of the benefits of visiting a
local baseball park. In addition, the model postulates that socioeconomic status
will affect one’s perception of discrimination.

The assumption here is that as one increases in socioeconomic standing,
helshe may perceive less discrimination because of opportunities afforded by
one”s socioeconomic status. As one’s socioeconomic status increases,
perceived discrimination decreases. A decrease in perceived discrimination
causes an increase in recreation participation.

Perceived discrimination reflects the use of West’s (1989) discrimination
perspective. It is affected by one’s subcultural identity. The stronger one
considers him/herself a member of a subcultural group, helshe will have a
stronger perception of discrimination. This then affects recreation behavior
because the perception of discrimination may act as a deterrent to participation

in recreation activities or sites.

48

Table 2 summarizes the direct and indirect influences of the different
concepts on recreation behavior. The only affect not shown in Table 2 is the
relationship between SES and SID. In the literature, there is no support as to
which construct influences the other. There is support for showing that there is a
relationship, and that this relationship is spurious. The model illustrates that
SES and SID have a common antecedent: ACC, which drives both of them,

thereby making SES and SID spuriously related by definition.

 

Table 2

Direct and Indirect Effects of Ethnicmr' Constructs onLecrestion Participation

 

 

 

 

 

 

 

Constructs Direct Effects Indirect Effects
ACC ACC—iSES ACC—rSES—iREC
ACC—vSID ACC—vSES—+BEN—i REC

ACC—+SES—iDIS—+REC
ACC—+SlD—iDlS—+REC
ACC—iSID—vBEN-rREC
ACC—+SID—9REC

SES SES—iREC SES—iBEN—bREC
SES-iDIS—iREC

SID SlD—iREC SID-+BEN—iREC
SID—iDIS—iREC

BEN BEN—iREC None

DIS DIS—*REC None

 

 

 

 

 

 

Conceptually, one can see that there are both direct and indirect effects
that influence recreation behavior. The EPRPM exemplifies a multiple causation

model. The indirect effects and direct effects combine to provide a more

49

comprehensive picture of the relationships and interrelationships of the different
concepts mentioned in the literature regarding ethnicity and recreation behavior.

The EPRPM reconceptualizes the previous research by offering an
alternative framework. This framework is much more involved than the
traditional marginality-ethnicity framework and incorporates other factors
identified as critical in the assessment of ethnic groups and recreation behavior.
Qgefiionnairs Dssign

Each of the constructs identified in the EPRP Model requires a scale for
measurement of the particular construct. The scales are made up of three to
eight items per construct. The constructs are as follows and each will be
considered in turn: (1) Acculturation, (2) Subcultural Identity, (3) Perceived
Discrimination, (4) Perceived Benefits of Recreation, (5) Socioeconomic Status,
(6) Recreation Participation (park use).

Assglturation. Acculturation was defined by Floyd, et al. (1993) as
”cultural distance.” They measured acculturation by asking respondents to
report their preferred language for use at home, radio listening and television
viewing, and reading newspapers and magazines. Their response categories
were: (1) English only; (2) mostly English, some Spanish; (3) both English and
Spanish equally; (4) mostly Spanish, some English; (5) mostly Spanish. The fifth
choice should have read ”Spanish only" for consistency.

Barona 8 Miller designed a short acculturation scale to ”sample cultural
behaviors from both the context of the family and the context of extrafamilial

social and media influences" (1994, p. 156). Their scale was designed for

50

Hispanic youth. They noted that previous acculturation scales used
sociodemographic data, and recommended that sociodemographic items be
excluded from acculturation scales.

The use of the acculturation scale by Floyd, et al. (1993) and Barona 8
Miller (1993) is consistent with traditional uses of acculturation scales. In a
study on first and second generation Puerto Ricans, Cortes, Rogler 8 Malgady
argue that there is an overemphasis in acculturation scales on language usage,"
and that the level of involvement in Puerto Rican or American2 culture as well as
other elements of culture should also be examined (1994, p. 711). The point is
well taken. The process of acculturation goes beyond language and
incorporates values, norms, and beliefs. Cortes, at al. (1994) consider the
extent to which one is bicultural as an important indicator as well.

They developed a 20 item scale (10 items to reflect Puerto Rican culture
and 10 items to reflect American culture) from a focus group of Puerto Ricans in
New York City. The themes reflected in the items include "language preference
and usage, values, ethnic pride, food preferences, child-bearing practices, and
interpersonal relations” (Cortes, et al., 1994, p. 712).

Saldafia (1994) remarked that acculturation scales often incorporate an
individual”s ethnicity and his/her ethnic identity. Ethnicity, in this case, was
defined as belonging to a particular ethnic group. Ethnic identity "may more

accurately reflect internal factors associated with level of acculturation” (p. 117).

 

2 In ordertodlstlngulshtrom PuerloRIcans, 'Amerimns' will be used because It alludesto
ctrlmraldlfferencesbasedonnatlonallsflcllnes. PuertoRIcansraciallybelongtobothWhlte,
BlackandNatlveAmericanraces,butethnicallybelongtonelther.

51

The Saldafla study reflects the current study”s operational use of subcultural
identity (see next section). Saldafia believed that acculturation and ethnic
identity should be two constructs. Of particular relevance to the current study is
Saldafia’s ”behavioral preference" sub-scale as a measure of acculturation.
This sub-scale (factor) accounted for 52% of the variance and used items that
were similar to the Floyd, et al., and Cortes, at al., studies.

Items from these previous studies were used to measure acculturation in
the current study by considering items that loaded highest, items which were
similar across studies, and alpha levels of .75 or higher. In addition, a pilot
study of 20 Puerto Ricans helped to discern the most relevant items. Table 3
shows the items and the articles they were taken from.3

In Table 3, the responses for items 1-7 are based on a five-point scale
measuring (1) only Spanish, (2) more Spanish than English, (3) both equally, (4)
more English than Spanish, and (5) only English. The responses for items 8-11
are also based on a five-point scale: (1) very important, (2) important, (3)
somewhat important, (4) a little important, and (5) unimportant.

Subcultural identig. Saldafia’s (1994) ”ethnic loyalty" sub-scale reflects
ona”s ethnic identity. She considered ethnic identification as complimentary to,
but different from, acculturation. Figure 10, reviewed previously, reflects this
conceptual relationship. Four items made up the scale (alpha = .81 ):
identification with culture, bilingualism, cultural pride, and participation in cultural

actiVIties.

 

3 Theltemsllstedlnthetablesretlectthelnltlalwordinganduseoftheltemstorthepretests.
ThereadershouldrefertothequestionnalrelnAppendleforthefinalwordlngoftheltems.

52

 

Table 3

The Aflltupstipn Scale

 

Item Source

 

1. Preferred language at home Floyd, at al. (1993)

 

2. Language preference for radio Floyd, et al. (1993), Saldaiia

 

 

and television. (1994)

3. Language preference for Floyd, et al. (1993), Saldaila
reading magazines, and (1994)

newspapers.

4. Language preference for Saldafta (1994)

music.

 

5. What language do your close Barona 8 Miller (1994)
friends speak?

 

6. You prefer going to parties Barona 8 Miller (1994)
which speak which language?

 

7. Which language do the people Barona 8 Miller (1994)
you visit or that visit you speak?

 

8. How important is it to celebrate Cortes, Rogler and Malgady
holidays in the American way? (1994)

9. How important is it to celebrate Cortes, Rogler and Malgady
holidays in the Puerto Rican way? (1994)

10. How much are Puerto Rican Cortes, Rogler and Malgady

 

 

 

 

 

 

values a part of your life? (1994)
11. How much are American Cortes, Rogler and Malgady
values a part of your life? (1994)

 

 

She correctly noted that "[little] research has addressed the relationship
of ethnic identity to level of acculturation, and these two constructs are often

treated synonymously'” (Saldafla, 1994, p. 117). Because of the synonymous

53

nature involved in ethnic studies, most scales measuring identity have been
developed for racial identity (see Burlew 8 Smith, 1991, for an excellent review).

Scales specifically measuring ethnocentrism have been lacking. Reasons
for this range from operationalization of ethnicity and race to transitional
definitions between race and ethnicity. ”For member of racial minorities, social
categories of race or national origin override the importance of ethnicity. Their
“choice” is limited because their identity is constrained to labels imposed by other
members of the society or by custom” (Xie 8 Goyette, 1997, pp. 549-550).

Rodriguez accentuates this particular point with the Puerto Ricans in New
York City.

The experience of Puerto Ricans in New York City points up more

clearly than any researched materials the chasm that exists

between whites and blacks in the United States and the racism that

afflicts both groups. For within the US. perspective, Puerto

Ricans, racially speaking belong to both groups; however,

ethnically, they belong to neither (1996, p. 25).

Few scales have been developed specifically for, or tested among, Puerto
Ricans. Because of this inattention, the scale measuring Puerto Rican cultural
identity was adopted from the African Self-Conscious (ASC) scale (Baldwin 8
Bell, 1985). The scale consists of 42 items measuring beliefs, attitudes, and
opinions of Black people. It is an overall measure of afrocentricity with reliability
of .90. There are both individual and group measures of subcultural identity.

A later analysis of the ASC scale suggested only 32 of the 42 items be
retained (Stokes, Murray, Peacock, 8 Kaiser, 1994). Whether 42 or 32 scale

items, it is an overwhelming number of items for the current study. The

researcher limited the items to seven (six of which came from the ASC scale).

54

Items were selected according to how the wording reflected the subcultural
identity construct. Item 4 was reworded to reflect a stronger Individual focus.
Table 4 illustrates the original items and how they were re-worded to reflect the

current study population.

 

 

 

Table 4
Sub ltu I I nt' cal
Original Wording of ASC Scale Ra-Wording
1. Blacks born in the United States are 1. I strongly identify myself as a Puerto
Black or African First, rather than Rican first, rather than as an American.

American or just plain people.
2. Blacks in America should try harder to 2. Puerto Ricans in America should try

 

 

be American rather than practicing harder to be American rather than
activities that link them up with their practicing activities that connect them to
African cultural heritage their cultural heritage.

3. Regardless of their Interests, 3. Regardless of their interests,
educational background, and social educational background, and social

achievements, I would prefer to associate achievements, I would prefer to associate
with black people than with non-blacks. with Puerto Ricans.

 

4. Black children should be taught that 4. When I was growing up, my parents
they are African people at an early age. taught me that my Puerto Rican culture is
very distinct, different from the American

 

 

 

culture.
5. I feel little sense of commitment to 5. I feel little sense of commitment to
Black people who are not close friends or other Puerto Ricans who are not close
relatives. friends or relatives.
6. White people, generally speaking, do 6. Americans, generally speaking, do not
not respect Black life. respect the Puerto Rican culture.
7. Not in ASC Scale. 7. If I had a choice, I would rather hold
the Puerto Rican flag, than the American
flag.

 

 

 

 

 

As one can see from Table 4, in addition to the six items from the ASC
scale, a seventh item was added to reflect a characteristic particular to the

Puerto Rican cultural identity: nationalism. Items 4 and 7 reflect this dimension.

55

Unlike Blacks in the US, national and cultural symbols (e.g., Puerto Rican flag,
el coqui, aljibaro) are defining characteristics of the Puerto Rican, not racial
identity.

Perhaps the primary point of contrast is that in Puerto Rico, racial

identity is subordinate to cultural identity, while in the US, racial

identification, to a large extent, determines cultural identification.

Thus when asked that divisive question, "What are you?” Puerto

Ricans of all colors and ancestry answer, "Puerto Rican,” while

most New Yorkers answer black, Jewish, or perhaps ”'of Italian

descent'” (Rodriguez, 1996, p. 26).

Perceived gissrimination. Several researchers have looked at the issue
of perceived discrimination and comfort level with respect to recreation
(Woodard, 1988; West 1989; Chavez, 1993; and Falk, 1995). Feagin (1991)
looked at antiblack discrimination in public places and focused on (1) the sites of
discrimination, (2) the character of discrimination actions, and (3) the range of
coping responses by Blacks to discrimination (p. 101). Feagin listed
discriminatory action against Blacks as five progressively serious actions:

(1) avoidance actions, such as white couple crossing the street

when a black male approaches;

(2) reject actions, such as poor service in public accommodations;

(3) verbal attacks, such as shouting racial epithets in the street;

(4) physical threats and harassment by white police officers; and

(5) physical threats and attacks by other whites, such as attacks by

white supremacists, in the street (1991, p. 102).

Given these factors, items were developed by borrowing from two studies.
Two items (Items 1 and 2 in Table 5) came from Floyd, et al.”s (1993) study. The
remaining three came from Sénchez 8 Brook’s (1996) study (T able 5). Item 2
was mentioned by several researchers as a factor in public areas (Feagin, 1991;

Chavez, 1993; Floyd, et al., 1993). Sénchez 8 Brock looked at the relationship

56

between perceived discrimination among Hispanics in the work setting. Their
hypotheses suggested a positive relationship between subcultural identity and
perception of discrimination. The items from the Sénchez 8 Brock scale were

reworded to reflect public parks instead of a work setting.

 

 

 

Table 5
Perceiv i ' in ti n.
Item Source
1. In some Southbridge parks, Floyd, Gramman, 8 Saenz (1993)
Puerto Ricans are not very
welcomed.

 

2. In Southbridge parks, police Floyd, Gramman, 8 Saenz (1993)
often hassle Puerto Ricans.

3. In Southbridge parks, the Sénchez 8 Brock (1996)
presence of other groups that are
not Puerto Rican makes me feel
uncomfortable.

4. In Southbridge parks, I feel that Sénchez 8 Brock (1996)
Americans exclude me from
activities because I am Puerto
Rican.

5. In Southbridge parks, Sénchez 8 Brock (1996)
Americans look down on me
because I participate in customs of
my culture.

 

 

 

 

 

 

 

 

Psrgivfl mnefits pf [sorgation There are a number of studies and

scales which have been developed to assess the cognitive nature of leisure and

recreation (lso—Ahola, 1980; Neulinger, 1981; Witt 8 Ellis, 1989). The items in

57

the Perceived Benefits scale (items 1A-1G, see Appendix A) reflect key concepts
in the recreation literature (e.g., relaxation, escape, socialization, family time,
and physical fitness). The items are based on lso-Ahola”s Benefits of Leisure
Scale. The wording was altered from ”benefit of leisure" to ”benefit of going to
Southbridge parks.” The idea was to make the term ”leisure” more specific to
the leisure activity under study: park use.

W. The most commonly used scale for
socioeconomic status is Hollingshead”s Two-Factor Index of Social Positioning
(1965), or his more current version: the Four-Factor Index of Social Status
(1975). Edwards-Hewitt and Gray (1995) provide an excellent review of other
socioeconomic scales, as well as a comparison between objective and
subjective reporting formats on these scales by ethnic group members. For this
study, the researcher contacted Gray to obtain the latest version of
Hollingshead”s index. However, the index was obtained after the current study
was undenrvay. As a result, the researcher altered the original index and
adapted it to reflect the socioeconomic questions asked in the current study.
Hollingshead’s Four-Factor Social Status scale used the items of marital status,
education, occupation, and gender.

Because the demographics section of the current study’s questionnaire
(see Appendix A) did not reflect the Hollingshead categories (e.g., the
questionnaire did not ask for both the husband and the wife’s information to
combine for a particular status score), it was decided to use Hollingshead”s

original Two-Factor Index which included the education and occupation as

58

items. Edwards-Hewitt and Gray also noted (from Gordon,1958) that some
studies use occupation in their socioeconomic scales, while others use income.
The current study does not use income because income was reported on a
household basis, whereas occupation and education are based on an individual
basis.

In addition, the researcher felt that gender and marital status (used in the
Four-Factor Index) are inappropriate variables to include in a scale of
socioeconomic status because of the nominal nature of the items. In the actual
Four-Factor Hollingshead scale, the gender and marital status are incorporated
into the education and occupation measures. Therefore, one still has a two
factor scale. Additionally, gender and marital status could have an impact on
other parts of the model (e.g., they could be intervening or mediating variables),
and to include them in only one construct does not seem meaningful for this
study.

The socioeconomic scale used for this study uses the items education,
occupation, and income. Education was scored from 1 to 7. In the
questionnaire, there are nine categories, but the last three were combined into
one category (graduate school) in order to match Hollingshead’s seven-point
scale. Occupation was scored from 0 to 9 based on Hollingshead’s 1975
designation for occupation. Income was scored from 1 to 9. The categories for
income reflect the US. Census breakdown (see Appendix A).

W. Park use is the final endogenous construct in

the EPRP Model. There are five public parks in the study area. Recreation

59

participation was measured by asking the annual frequency of use (by season)
for each of the parks (see Appendix A). The five items measuring park use are
the total visits to each park by a respondent. The seven scores for park use by
season were:

(1) never (scored 0);

(2) once (scored 1);

(3) almost never/ 2-4 times (scored 3);

(4) sometimes! 5-13 times (scored 9);

(5) frequent! 14-20 times (scored 17);

(6) very frequent! 21-29 times (scored 25);

(7) a lot/ 30+ times (scored 35).

Each score represents the middle value. For example, if a respondent
checked the fourth box on the questionnaire (sometimes), helshe falls between 5
and 13, and would be given a score of ”'9.” The five items measuring parks were
combined to give an overall assessment of participation in each of the city's
public parks by respondent. For the purpose of this study, the interest is not in
the actual activities engaged in the park, but whether or not the ethnic group
members visited the public recreation sites (parks), and if so, how frequent. As
a result, each respondent was given a "scale score” from 0 to 6 for use of each
park. The constmction of the scale categories is reflected in Table 6.

Using Table 6, if one were to go to West Street Park (see Appendix A)
14-20 times in the Summer, once in the Fall, never in the Winter and 5—13 times
in the Spring, then the respondent would be assigned the following designation:

17+1+0+9 = 27. The total use for West Street Park is 27 and it falls within the

“total range" of 13-36, therefore the respondent would be coded a "3"; a

60

"somewhat frequent“ user of West Street Park. This is done for every park. All

five parks are then combined into an index to measure park use.

 

Table 6

Scoring of Paik Use for Esch Park

Ssason Annual Total Scale Annual Use

Scars Maximum Rsnge Me Dssignation
for Each Park

 

 

0 0 0 Never

1 4 1-4 1 Very Few

3 12 5-12 2 A Few Times

9 36 13-36 3 Somewhat
Frequent

17 68 37—68 4 Frequent

25 100 69-100 5 Very
Frequent

35 140 101-140 6 A Lot

 

Sub'ects and m lin Frame

In order to cany out this study, access to an ethnic group population was
needed to empirically test the different theoretical constructs. The ethnic group
members selected were Puerto Rican. This ethnic group has not been
researched in previous ethnic recreation studies. Because of the relationship
between the United States (US) and Puerto Rico, Puerto Ricans provide

researchers with an opportunity to look at acculturation processes, perceived

61

discrimination, socioeconomic factors, issues regarding cultural identity, and a
unique perspective on their perceptions of parks.

Puerto Ricans are part of the growing Latino population. It was important
to consider a Latino group for three reasons. First, from a demographic
standpoint, Latinos will be the largest ethnic group at the beginning of the next
millennium. Second, previous research regarding ethnicity and recreation has
focused primarily on African Americans. Third, the researcher needed to have
access to the ethnic group members. Because the researcher is Puerto Rican,
and a native of the study area, the researcher had access to the population, and
an understanding of the population and its cultural nuances to facilitate
participation in the study.

GQgpaphis profils. This study was conducted in Southbridge,
Massachusetts. Southbridge is located in Worcester County - south-central
Massachusetts (MA), and borders northern Connecticut. The city is hilly,
particularly in the northern and southern part of the city. The central part of the
city is in a valley. Southbridge is approximately 60 miles west of Boston, MA.
All parks are located approximately within one mile of the downtown area (see
Appendix B).

W According to the 1990 Census, Southbridge”s
population was 17,816. The median household income for Southbridge residents
is $27, 834. Puerto Ricans constitute the largest ethnic group in Southbridge.

Puerto Ricans represent 12% (2,109) of the city’s population, and 91% of

the city's Hispanic population. Over half of the Puerto Rican population (56%)

62,

was born in Puerto Rico. Spanish is spoken in nearly 10% of the entire
population, and by about 75% of all Hispanics (United States Census, 1990).

The first Puerto Rican family arrived in Southbridge in 1957 (Brown,
1982). Puerto Rico, at the time, was transitioning from an agricultural economy
to an industrial economy. As a result, Puerto Rico’s agrarian labor force turned
to the US for economic relief. Specifically, Southbridge’s Prest-Wheel Company
hired many Puerto Ricans in the late 19503 and early 19603(Brown, 1982; Datz,
1998)

The late 19503 marked the beginning of serious Puerto Rican

migration into Central Massachusetts. For the first few years, the

most frequent landing spot was an industrial town accustomed to

employing immigrants: Southbridge. (Datz, 1998, p. A23).

By the late 19603, Puerto Ricans were attracted to Southbridge because
of its need for an unskilled labor force to work in industry. Those Puerto Ricans
who arrived in the 19603 paved the way for the next wave in the 19703. The
majority of those arriving spoke only Spanish. Strong family ties to Puerto Rico
were maintained because families often left siblings and parents behind.
Southbridge Puerto Ricans travel to Puerto Rico quite often and send money
and clothing to their extended families in Puerto Rico.

The majority of the Puerto Ricans arriving in Southbridge, if not related,
already knew each other from their barrios, neighboring towns, or family friends.
As a result, an enclave of Puerto Ricans was established in Southbridge which
reinforced ethnic cohesion. Southbridge Puerto Ricans are particularly suited to

this study because of the strong ties to their homeland. The respondents would

be emigrant, first, or second generation Puerto Ricans in Southbridge.

63

This study comes at a time that marks Puerto Rico’s 100‘” anniversary
under US rule. A plebiscite was recently conducted in December of 1998. The
researcher considered what possible impact this might have in the nationalistic
tendencies of some of the Puerto Ricans in Southbridge.

The impact was considered to be minimal. The local and county
newspapers (Southbridge Evening News, Worcester Telegram 8 Gazette)
reported opinions of the plebiscite and the results of the plebiscite. In both
instances, the papers reported the Puerto Rican communities (on both sides of
the Carribean) as being divided about the issue of statehood for Puerto Rico
(Datz, 1998; Echegaray, 1998; Rabinowitz, 1998; “Puerto Ricans,” 1998). As a
result of the focus on Latinos in Worcester County, Massachusetts, the county
paper reported that the majority of Latinos in the county are from Puerto Rico or
of Puerto Rican descent, and that of the six communities in the county with the
largest number of Latinos, Southbridge has the largest Latino population for any
one city in Massachusetts:16.7% (2,999) of the Southbridge population as of
1995 (Echegaray, 1998).

Population and sample size. The subjects were selected from the Puerto
Rican population of Southbridge, MA. Subjects were 14 years of age or older.
Because high schoolers are active users of Southbridge parks, it was important
to include them in this study.

High schoolers offer a broader age variance which may illustrate
generational influences on participation at public recreation sites. In addition,

because a majority of Puerto Ricans in Southbridge were born in Puerto Rico, it

64

is expected that their perspective on American values and their acculturation
and assimilation will be different than their children’s and grandchildren’s
perspectives.

The sample size for this study was determined by three values: (1) alpha
value (for Type I error), (2) power (for Type II error), and (3) effect size (how
large a difference one wants to be able to detect in one”s study) (Shavelson,
1996). The level of significance for this study will be set at the .05 level. The
probability of rejecting a false null hypothesis is set at .90 (1-8). The researcher
would like to have statistical tests detect a standardized difference of .20 (effect
size).

Using Table M in Shavelson’s (1996) text, the approximate sample size
for a one-tailed test design is N=215. For a two-tailed test design, N=262. The
majority of the hypotheses in this study consider a one-tail design, however
there is one hypothesis that requires a two-tailed test.

Therefore, a minimum of 262 subjects is considered adequate for this
study. This is consistent with Kline’s assertion that the larger the sample the
better, and that samples of “less than 100 could produce misleading results” in
factorial designs (1994, p.180). In addition to the statistical power needed, 262
subjects represent 19% of the Southbridge Hispanic population that is older than
14 years (1,397), a more than adequate sample size for inference to the
population. 1

WWW For the selection of Puerto

Rican high schoolers, the researcher met with the superintendent of schools and

65

the high school principal to obtain permission for the administration of the
questionnaire during home room period. A copy of the questionnaire, the human
subjects approval form from Michigan State University, and a letter of
introduction were provided for them prior to administration approval. It was
mentioned that participation was strictly voluntary and that all information is
anonymous and confidential. Permission was verbally granted from both the
superintendent and the principal.

The high school liaison was the head of the Social Studies curriculum in
the school system. Explanations and instructions were given to him on how the
researcher wanted to conduct the study. Home room teachers were instructed
on how to administer the questionnaire and to return the completed
questionnaires to the Principal’s Office. A count of the number of Puerto Ricans
in the High School was obtained and questionnaires were provided in both
English and Spanish.

According to a breakdown by home rooms, there were a total of 135
Latino high school students (W. Gosk 8 J. P. Bailey, personal communication,
January, 1999). The high school liaison was instructed to deliver the box of
surveys to the researcher’s sister (who works in the school system) (see
Appendix B). She mailed the surveys back to the researcher.

Prosfiures for adult population sample selection. In order to sample the
Puerto Rican adult population, key persons in the Puerto Rican community had
to be contacted. These community leaders have access to lists of names, or

have contact with Puerto Ricans at Puerto Rican-owned establishments. This

66

process involved tapping into the Puerto Rican community’s social capital by
utilizing formal and informal networks to obtain verbal consent and addresses of
prospective respondents. Such contacts were made in the following community
areas:

- St. Mary’s Parish (Catholic Church with Spanish Mass and Puerto Rican
congregation);

- Catholic Charities (non-profit organization which has a heavy Puerto
Rican clientele);

. Mane Creations (Puerto Rican-owned beauty parlor frequented by Puerto
Rican men and women);

- Hispanic-American Store (Puerto Rican—owned and frequented);

- Women, Infant and Child Program (VVIC) personal contact;

- Key community persons at different work settings that have personal
contacts with co-workers;

- Annual census at the Town Clerk’s office.

In addition to the above list, the researcher solicited family and friends to
help “spread the word.” A letter explaining the purpose of the study was given to
each of the community leaders listed in the above areas, and they were given
instructions to ask their clients/parishnerslco-workers to participate by giving
their address on an address “sign-up" sheet (see Appendix B).

The community leader from the Head Start Program had access to
several Puerto Ricans. Although she compiled nearly 35 names, she felt it best
if the researcher mailed the surveys to her and that she administer the
questionnaire to her clients. The researcher acquiesced, but still had to get the
list of the names to (1) make sure that the respondents were not sent two
surveys, and (2) know whether to send an English or Spanish questionnaire.

The contact at the Head Start Program notified the researcher’s parents once

67

the questionnaires were completed so that the Head Start Program would not
have to pay for their delivery.

In addition to word of mouth and contacting community leaders, the
researcher was granted the opportunity to address the Puerto Rican public at
one of the heavily attended masses of the year: Christmas Eve Mass. The priest
allowed the researcher to address the congregation. The speech was in
Spanish (see Appendix B). The priest also situated a desk and chair for the
researcher so that parishioners could sign up for the study after mass ended.

Selsgion of subiscts: Problems and solutions to creating a list. While the
method for the creation of a list is somewhat unorthodox, it is a functiOn of the
population under study based on the researcher’s knowledge of the population.
Therefore, alternative methods for a list were needed. Researchers call for
creative solutions to this problem. For example, Salant and Dillman (1994)
suggest creating a list from multiple sources or using a purposive sample
design.

Many Puerto Ricans in Southbridge do not have listed phone numbers,
therefore, the phonebook was not trustworthy. Additionally, the Puerto Rican
population is very mobile. Often times they will move one or two times a year,
move in with extended family, or relocate to Puerto Rico. These situations
create problems with using the telephone book as a list source. These

Another traditional list source is the city’s annual census‘. This list has

the same problems as phone listings. The town clerk mentioned that the census

 

Under Massachusetts (MA) election law. MA cities are required to conduct an annual census. The census provides
names, dates of birth, precinct and occupations. The annual census is used mostly for grants and public funding.

68

is not as accurate as they would like due to a lack of cooperation on the part of
Puerto Ricans, undercounting, and the lack of Puerto Rican census takers
(Helen I. Lenti, personal communication, December 22, 1998). Therefore,
traditional sampling techniques need to be augmented to increase the possibility
of an individual’s participation in the study.

In an article on researching diverse populations, Henderson (1998) noted
that “[methods] are important, but the strategies used to get information are
essential. Researchers may need to stray from research protocol to obtain data
and create an environment of social support.” While her comment was oriented
towards qualitative research, it is appropriate for this study. Social support for
this study was first created for the study to be successful in the adult Puerto
Rican population.

Knowledge of the population is critical to getting enough responses to
perform useful analyses. The study population requires informal and formal
lines of communication. For example, the researcher “informally” solicited
names via personal contacts throughout the community in order to make Puerto
Ricans aware of the study. It is culturally more acceptable to first “ask” if the
subject’s name and address can be used for a mailing, and then perform the
actual mailing.

Watson (1992) noted that conventional sampling methods have been
ineffective in reaching minority populations. He identified three sampling
techniques and the problems associated with ethnic minorities. They are as

follows:

69

1. Random sampling - inadequate as many in the ethnic
community have not been on the electoral registers.

2. Quote sampling - insufficient data have existed from the census
on which to sample targets and selected sample points may not
reflect where ethnic groups actually are.

3. Random digit dialing - how do you establish ethnic origin with
any degree of certainty by telephone? Maison, 1992, p. 339).

Although his comments deal with ethnic minorities in the United Kingdom,
researchers have encountered similar problems in the United States. Cox
(1990), for example, argues for non-traditional designs and unconventional
methods for researching minorities in the US.

“Sutton and Schurrnan note that conventional methodology calls
for all respondents and for investigators to refrain from disclosing
the details of the research objective to the respondent. They found
that both guidelines were unworkable in their own research
[regarding minorities], however, and thus followed no rule. They
acknowledge that they made a conscious decision to violate these
rules despite the potential effects of bias on the results. They
concluded that considerations such as an ability to obtain data and
to create an environment of social support in which interviewees
will provide responses must sometimes take precedence over
traditional notion of scientific “objectivity.” I believe that their
findings are applicable to other emotionally sensitive topics such
as racioethnicity, and that they illustrate the need for new
paradigms of research methodology (Cox, 1990, p. 11).

The above quote illustrates several issues related to researching ethnic
minorities. First, it is difficult to obtain truly representative samples of minority
populations. Second, because of the difficulty in obtaining a list, other
nontraditional methods are required to augment or create a list. Third, an
environment of social support (over objectivity) is needed for participation.

Through contact with community leaders, the researcher “spread the
word” and Southbridge Puerto Ricans were more responsive. The social

support was created along informal lines. Asking Puerto Ricans to fill out the

70

questionnaire when one first visits would be considered improper. An initial
house call or visit should be informal. One can talk about business, but not
actually conduct business. This is a similar concept to what Winter and Chavez
(1998) referred to as taking time to “visit” for successful data gathering.

If one were to conduct the personal interviews at the place of residence it
would require at least two visits. The first visit would be to establish social
support. This will typically involve sitting down for a cup of coffee, catch up on
social, political, or family events in the community or Puerto Rico, and then
discussion on the survey. To talk “just business” or visit quickly would be
considered “rude” and would most likely not assure one of participation. A
second visit would be required to conduct the actual interview where it would be
considered an “official” visit. As one can gather, the cost and time for this
method of ensuring an adequate sample size and response rate would be quite
large.

Pret tin h uestionnaire

Two pretests were administered. The first pretest involved Puerto Ricans
from Michigan State University’s (MSU) Puerto Rican Student Association.
Because this particular group of Puerto Ricans were not from Southbridge, they
provided excellent feedback on the wording of the questions (from English to
Spanish), as well as remarks on the sensitivity of the questions being asked. In
addition, the group of 10 MSU Puerto Ricans answered the questions regarding

cultural identity, values, language preference, and perceived discrimination.

71

After evaluating the initial pretest, corrections were made, and the second
pretest was administered to Puerto Ricans in Southbridge.

From December 20, 1998 to January 4, 1999, the researcher met with 15
Southbridge Puerto Ricans and administered the questionnaire. Of the 15
Puerto Ricans in the pretest, eight questionnaires were administered in English
and seven in Spanish. The following is a synopsis of the respondents:

Mean time living in Southbridge: 21 years (Min: 2, Max: 35)
67% born in Puerto Rico, 33% in United States

33% Male, 67% Female

Mean age: 35, (Minimum: 18, Maximum: 53)

40% Single, 53% Married, 7% Divorced
87% Employed, 7% Self-employed, 7% Other

Each respondent was given the option of taking the survey in English or
Spanish. The respondents were timed. After the respondents finished, the

researcher asked the following questions:

- Was there anything in the questionnaire you didn’t understand or found
particularly difficult?

- Were any of the questions offensive?

- Do you think other Puerto Ricans would want to answer this survey? Why
or why not?

- Do you think the questionnaire is too long?

Pretfit analysis of ssales. In assessing the reliability analysis the author

took into consideration the N of 15. This is not a large enough N to warrant any
conclusions, however it does provide some insight into the questions
themselves. With an N of 15, the author expected anything above a .6 to be
adequate. The alpha reliability is affected by the size of N, and by the number of
items. In general, it is better to have more items so that if there were a need to

delete one or two, one could still have a respectable scale. What follows is a

72

summary of the scales, their alpha reliabilities and the comments provided by
the respondents‘from both pretests (See Appendix A for corresponding
alpha-numerals in parentheses).

Percein panefits of parks (IA-1G). The alpha reliability for the scale on
perceived park benefits was .8783. The weakest measure was 1F. Without this
item, the alpha reliability increased to .8880. Given this weak increase, and the
low N, the researcher decided to retain the scale without any modifications.
Respondents found these questions rather straight-fonrvard.

Publis rgreation participation (Section 2A-2E). This is the dependent
variable measuring park use. For this section, the author tried two different
scales. The first scale for park use (not on the final version) utilized a scale that
measured only the annual use of each park. The second scale divided the year
into seasons. The decision of which scale to use was decided by (1)
preference of the respondents, and (2) by the amount of information gathered by
each of the different scales.

In the first case, three of the respondents were visited twice (on two
different days) by the researcher in order to see if the same information would be
obtained and to see which scale was preferred by the respondents. In all three
cases, the respondents preferred the “seasons” scale over the “annual” scale.

In addition, the “season” scale offered more insight into actual use. In all three
cases, the season scale gave more use. Respondents mentioned that it was

easier to recall what they did by season than by year. Respondents also

73

mentioned that the use of “almost never” to “a lot“ above the categories was
helpful. Park use was coded as:

0 = Never

1 = Once

3 = 2-4 times

9 = 5-13 times
17 = 14-20 times
25 = 21-29 times
30 = 30+ times

The middle value of the number of times was given for each category.
The seasons were then summed and a value was arrived for annual use for
each park. The summation of all the parks is the “overall” park use value
assigned to the respondent. The park which had the greatest use was Westville
Dam (mean = 10.33, min: 0, Max: 52), followed by Henry Street Park (mean =
8.4, min: 0, max: 51). This was an interesting finding. Geographically, Westville
Dam is furthest away from where the majority of Puerto Ricans live, and Henry
Street is the closest to the Puerto Rican population.

Acculturation sgle (Section 3A-3G and 4A-4D). The original wording is
reflected in Table 3. The acculturation scale combined all the items from 3A to
40. However, it was found that the combined scale of language and values
gave a .70 reliability. If one were to divide the two scales into language and
values. The first acculturation scale (language) had an alpha reliability of .9280.
Oddly enough, the weakest item was “preferred language at home.” With the
removal of this item, the alpha reliability became .9452. It was decided that due

to the small N, all items would be kept.

74

This second acculturation scale (values) tried to get at Puerto Rican
versus American value systems. This was the weakest of all the scales and
gave a -1.92 alpha reliability. This is an impossible number because the alpha
reliability should be between 0 and 1. The researcher reverse-coded 4A and 4C
because this reflected American values. If they are high on American values,
then this should reflect a lower Puerto Rican valuation. With the inversion, the
alpha reliability became .52.

In the initial pretest, the items for the values scale were worded slightly
different. The wording was changed for the second pretest because of
recommendations from the MSU Puerto Ricans. However, a factor analysis
showed two very distinct factors - an American and a Puerto Rican - on both
pretests. It was decided to change all the items to reflect only Puerto Rican
values because it seemed confusing to the respondents.

Some respondents wondered why the researcher was asking on both
American and Puerto Rican values, on the same issues. Although Cortes, et al.
(1994) meant for the original scale to reflect biculturality, it seemed redundant to
the respondents and it did not make sense to them. The change was made to
reflect the feedback from both the analysis and the respondents. It was
hypothesized that the change will create a stronger value scale. The combined
scale of 11 items (language and values) for acculturation will be tested again in
the actual study. If the two concepts do not converge into one scale, the
researcher will use only the language items to measure acculturation because it

is the stronger scale.

75

Parceivad discriminatipn (Section 5A-5E). The perceived discrimination
scale had a reliability of .66. With the deletion of 5B the alpha became .70.

With the deletion of 58 and 5A the alpha became .73. There were several
issues regarding this particular scale. For example, statement 53 (see Table 5)
was controversial with the respondents. Some questioned the relevance of the
statement. However, when the researcher explained the reason for its inclusion,
they agreed that it should be left in.

The respondents felt that the “not sure” middle category should be
changed to “neither agree nor disagree” as a better option. With this change,
the researcher felt that it was acceptable to proceed with the scale and include
both 5A and 5B in the scale. In addition, “feel” was include in the wording of all
the items for the sake of consistency (compare the Table 5 wording with final
wording in Appendix A).

Subsgltural identfly (Section 6A-6F). The subcultural identity scale had a
reliability of .58. With the deletion of 6B, the alpha reliability became .70 (see
Table 4). Items 6B and 6E were re-coded in the reliability analysis. In the final
cepy of the questionnaire, 6B remained the same, but the wording of GE seemed
confusing to respondents because of the use of “not“ twice in the statement. As
a result, SE was re-written so that it would not have to be re-coded. Although 63
may be too long, re-wording it would have been problematic because a shorter

version would not give the same meaning in English or Spanish. If BB proved

problematic during the actual study, it could always be taken out of the scale.

76

Dempgraphic section of pretest. There were two inconsistencies between
the English and Spanish surveys on the last page of the survey. In the English
survey, there was no “7A2" question which asked the respondent if helshe had
ever lived in Puerto Rico, and if so how many years. This is getting at the
acculturation issue and it is a counterpoint to how long they have lived in the
United States.

The second inconsistency involved the use of “total family income“ in the
English version, and “individual income” in the Spanish version. It was an
oversight on the part of the researcher, and was pointed out by a respondent
who was editing the surveys for English/Spanish consistency in the translation.
In addition, the income year was corrected; it was changed from 1997 to 1998.

Below is a summary of the alpha reliabilities for each scale (with all

items):
Name pf Scale Alpha reliabil'sy
Perceived Park Benefits .88
Language Preference .92
Values .52
Perceived Discrimination .66
Cultural Identity .58

The values scale when combined with the language preference scale
yielded a .70 alpha reliability. This was to be a combined scale to measure
acculturation as a reflection of language and values. The combined scale had a
high reliability, but there may not be reason for combining them as
recommended by previous research (Cortes, et al., 1994). The original scale

version was weak on the “values” scale. The two scales listed above, which

77

were below .60, had items re-worded or changed entirely in order to make the
statements more understandable.

There was one final inconsistency regarding the demographics section.
Items 7B3 and 7B4 asked for data concerning only one set of grandparents.
The researcher did not correct this because it would have been problematic for
the typesetting. As a result, this aspect of the study could not be incorporated in
acculturation measures accounting for generational status. In addition, the
socioeconomic scale was not tested because the researcher did not have the
Hollingshead Index at the time of the pretest.

Quastipnnaire distribution. The first step, discussed in previous sections,
was to collect names via purposive and snow-ball sampling techniques to create
a list. The reader should note that the researcher started to “spread the word”
about the study about three weeks before the letters were sent to community
leaders. The letters and sign-up sheets were sent to community leaders about a
month prior to addressing the congregation on Christmas Eve. The second step
was to cross-reference the names with the various lists to make sure that the
names do not appear twice.

The final list for the Puerto Rican adult population yielded 690 mailing
addresses (539 signed up and 151 addresses were from the phone book and the
town census), and 35 additional surveys to be delivered to the Head Start
Program. There were 135 potential respondents who were Puerto Rican high

schoolers. The total number of surveys to be distributed was 860 (62% of the

78

1397 resident Latinos aged 14 and over). The total time it took to collect the
names and addresses was approximately two and a half months.

The researcher opened “formal” lines of communication. Instead of
making a “formal” second visit, adult subjects were contacted by mail. The
surveys were mailed to those whose names were solicited. In addition, the
cover letter should be made to look official by use of letterhead. It was expect
that if Puerto Ricans voluntarily gave their names, and the study was being
conducted by someone from their own community, they would be more inclined
to participate. This approach was recommended by several leaders in the
community.

Coll ion Da

In order to enhance a good response rate, the collection process included
techniques suggested by Dillman (1978). Everything prescribed by Dillman’s
total design method (T DM) was not followed due to monetary and time
constraints. The basic concepts, however, were applied. The mailing procedure
for the collection of the data involved the following steps:

(1) mailing the introductory letter and autobiography;

(2) mailing the cover letter and questionnaire;

(3) mailing the follow-up letter and replacement questionnaire.

The initial mailing involved the use of a letter and autobiography. The
letter acted as an announcement and solicitation for completion of the
questionnaire (See Appendix B). In addition, a short autobiography of the

researcher was included so that Puerto Ricans had an update on the researcher

since the time he left the community. The autobiography acted as a proxy for

79

the researchers’ personal visit, and allows for the respondent to have an idea of
the researcher’s character. This is an extension of the social support concept.
The initial mailing was sent by first class mail and used department envelopes
and letterhead.

One week after the initial mailing, the cover letter and questionnaire were
mailed out (See Appendix B). The cover letter provided information on the
purpose of the study, what the information will be used for, and how their names
were chosen. In order to save on costs, the first wave of surveys were mailed by
third class bulk mail. Accounting for a mailing time of 7-10 days, the return
window given was approximately two weeks.

There were some problems with the third class bulk mailing. After
speaking with the post office in Southbridge, the researcher found that third
class mail gets distributed very poorly and is not always sorted the same day it
arrives because it is not considered priority mail. The first surveys were arriving
haphazardly. As a result, the researcher waited an additional two weeks for
responses to arrive before mailing the follow-up survey. In the cover letter, the
researcher tried to convey an understanding of the problem with the mail, while
at the same time expressing a necessity for having full participation.
W

The purpose of this study is to establish a substantive model and test it
empirically. This model is a causal model. The hypotheses set forth in Chapter
1 reflect two things: (1) the proposed nature of the relationships between the

theoretical constructs in the literature, and (2) the hypothesis that the data are

80

empirically consistent with the theoretical model. The strength of conducting a
path analysis is that it can provide insight into the nature of the relationships
between constructs. As a result, it illustrates where proposed causal
relationships may not exist or at best very weak. Because the causal arrows in
the model are “potential“ flows, if the data are not consistent with the model, the
researcher can postulate an alternative model that is both empirically consistent
with the data, and substantively consistent with the original theoretical
underpinnings of the initial model as discussed in the literature review chapter.

The data was analyzed with both a confirmatory factor analysis and a
path analysis. The former assesses the validity and reliability of the indicators of
the different constructs. The latter assesses the fit of the model. Step 1, than, is
to run a confirmatory factor analysis on the total number of items and obtain the
correlations among the constructs. Step 2 takes the correlations of the
constructs as the basis for the path modeling. These steps will be discussed in
more detail in the next chapter.
Valig'mr and Reliabil'mr

To the extent that the internal consistency theorem fits the data, this is a
model of validity. The relationship between validity and reliability is
conceptualized in a particular way. It does not make sense to calculate a
reliability coefficient until one knows that all indicators are ultimate indicators of
the same thing.

A measure is reliable to the extent that it measures whatever it measures

consistently. Reliability ranges, theoretically, from 0.0 to +1.0. Zero would mean

81

no reliability and one would mean perfect reliability. "What” it measures is a
measure of validity, not a question of reliability. Reliability is based on an
algorithm that tells one how consistent the measure is, but it says nothing about
what one is measuring.

There are several measures of reliability (e.g., test-retest, split-half), but
the most often employed is Cronbach’s Alpha. The derivation of Cronbach’s
Alpha is based on three assumptions: (1) that the unidimensional factor model is
correct, (2) that the variances are equal, and (3) that all items are equally
reliable (i.e., their factor loadings are the same).

If the assumptions are wrong, then the alpha is an inaccurate assessment
of the reliability coefficient. Robustness refers to the ability of an estimator to
have its assumptions violated and still provide a useful estimates. Alpha is
relatively robust with respect to the third assumption.

However, If the unidimensional factor model is incorrect, the reliability
coefficient does not make sense, it is a meaningless number. The reliability is
something that is not a constnict. By employing a CFA, one can test the
unidimensional factor for all possible parallel factors with the internal
consistency and parallelism theorems. The first assumption - the idea that one
has content validity - can be checked.

The second assumption is that there are equal variances. That
assumption is relaxed by standardizing the data. Standardizing the data forces
all the variances to be equal. By forcing all the variances to be equal, the

second assumption no longer applies.

82

The method of estimating reliability that is employed in this study, when
we standardize all the variables, is Standardized Item Alpha. The Standardized
Item Alpha is given in Program CFA”s output. It is the Speannan-Brown
coefficient applied to the scaling problem and it is derived in the following

manner:

 

NF

SI” = l+(N—1)F (1)

 

 

The "F" is the mean of the correlations (which is an estimate of the
reliability of one factor/construct). The 'N” is the number of indicators of the
construct. If for example, there are four indicators, then that would mean that
one is quadrupling the test. What this means, substantively and statistically, is
that the more indicators a construct has, the higher the reliability (given
unidimensionality) of the construct.

Given the validity and reliability of one’s measures, one could safely sum
indicators to create a composite variable (factor/construct) which could be used
in a path analysis. Program CFA supplies the researcher with a factor
correlation matrix corrected for attenuation. This is the matrix with the
correlations among the constructs, which will be used in the path analysis.
Assassment of Errors/Testing for Fit

A final word on residuals is necessary for the reader to fully comprehend
the next section. Errors are assessed in order to find whether the indicator is

within sampling error of zero. As such, the null hypothesis is that error equals

83

zero. In other words, one wants to fail to reject (accept) the null hypothesis and
find it NOT significant, i.e., no significant error.

If one is assessing errors (i.e., testing the fit of the model) using the
significance test, one is looking for a probability value greater than or equal to
.05 (when testing at the 95% confidence interval) in the output. Program CFA
uses the chi square test of fit to assess the fit of the model. The chi square test
reverses the role of the null hypothesis. The research hypothesis is the null
hypothesis.

Therefore, a low chi square is good. A large chi square indicates
departure from the model. If chi square is not significant, than the model fits.

An alternative to the significance test is to compare the observed versus
the expected correlations for the items (during both the internal consistency and
parallelism tests) and determine whether the errors are within sampling error of
zero by using the confidence interval approach. A combination of both methods
for assessing the errors are employed in this study, with greater emphasis

placed on the confidence interval approach.

Chapter 4
DATA ANALYSIS
This chapter is partitioned into three distinct sections for the purpose of
clarity. The first section is an overview of the descriptive statistics regarding the
survey respondents. Section two addresses the findings of the CFA on the
study data. Of particular interest in this section are the items which were used in
the final indicas. The last section of this chapter examines the path analysis,

and assesses the fit of the model.

Section One: Dessriptiva Statistiss

W8:

A total of 690 questionnaires were mailed to Puerto Ricans in Southbridge
(see Table 7). Of the 690 surveys initially mailed, 77 (11%) were returned due
to incorrect addresses. The majority (45) of the addresses for the returned
letters were addresses from the phonebook. Thirty percent of the phonebook
address were incorrect, while only five percent of the sign-up list yielded
incorrect addresses. This reinforces the notion that purposive sampling was
indeed a better way to identify the desired respondents than using the telephone
directory, especially given the mobile nature of the population under study. A
total of approximately 613 questionnaires was probably delivered.

A total of 304 Puerto Ricans responded by mail. This produced a
response rate of approximately 50%. The low level of response was probably a
result of the third class postage for the first mailing of the questionnaires.

Factors which may have influenced the overall response included the following:

85

. lack of priority given to third class mail;

. lack of current addresses in phone book;

. no forwarding address.

Because of problems related to mailing and delivery, the response rate of 50% is
a conservative estimate. The amount of actual delivered questionnaires is

unknown.

 

Table 7

Mailed Surveys and Response

 

Sam le Initial lncorr Prob I
Source Mailed A r 33 Dalivered
Phone Book 151 45 106
Sign-up List 52$ 32 5_07_
Totals 690 77
Completed 8 Returned by mail: 304 (50% response rate)
High School: 57
Community Leader". 29
Unusable Surveys: :6

Total usable surveys (N) -

 

In addition to the 304 returned by mail, 29 were received from the
community leader at the Head Start Program, and 57 came from Southbridge
High School (T able 7). The number of high school students who responded
could have been higher as well. Accessibility to all 135 Latino high schoolers
was not possible for several reasons:

questionnaires were distributed one day only (during home room period);

- questionnaires were distributed the first day after students returned from
Spring Break, and attendance may have been lower that day;

86

eligible students could have been late or sick, and;
. not all Latino high schoolers are Puerto Rican.

On the day the questionnaire was administered, there were 57 Puerto Rican
students in the high school during home room period.

Six of the surveys were unusable because of incomplete or ambiguous
answers, or because the respondents were not Puerto Rican. The total amount
of usable surveys (N) totaled 384. Of the 384 surveys, 209 (54%) were in
Spanish. Nearly 61% of the adult population preferred the questionnaire in
Spanish, and 73% of all respondents were born in Puerto Rico.

Sampla and Pppulation Dampgraphics

Because there is no information on non-response, the sample was
compared to population figures from the 1990 US. Census to assess
representativeness5. Puerto Ricans constitute 90% of the Latino population in
Southbridge. The researcher used Latinos (Hispanics) as the reference
population when comparing the sample to the population. The 1990 Census
does not have a breakdown by Puerto Ricans. According to the 1990 Census,
persons 14 years of age and older, of Latino origin, constitute 1,397 total
persons.

The mean age of respondents in the study is 34, with the youngest
respondent being 14 years of age, and the oldest being 80 years of age. The
1990 Census Latino high school age group (14-17 year olds) for Southbridge is

made up of 129 persons. This corresponds to the 135 Latinos at Southbridge

 

There are differences beMen the sample and the 1990 population statistics. However, one may speculate whether
acensusthatisnearlyadecadeoldistrulyrepresentative.

87

Hig
Stu

Inc

High School. The study does not, however, include high school drop outs. This
study includes 42% (57) of the high school age Latinos. In addition, this study
includes 27% (384 of 1,397) of the entire Latino population of Southbridge.
Table 8 illustrates the frequencies between age and gender in the
sample, and expected frequencies based on census percentages. In both
cases, the observed frequencies do not equal the expected frequencies. The

chi square for age is 61.92 08m (0,: _05. a: ,0, = 18.31 ). Chi square for gender is

31-29 (xzcrltlcal (an .05, dl= 1) = 334)-

 

Table 8

A nd Gend rBr kdown in nsus an am I

wt 0 N= 1 Samm “ Census °
(°/°) (W
1 . 14-1 7 14 9
2. 1 8-21 8 12
3. 22-24 5 9
4. 25-29 14 16
5. 30-34 1 5 17
6. 35-39 10 13
7. 40-44 10 9
8. 45-49 9 5
9. 50-54 5 3
10. 55-59 5 4
1 1. 60+ 5 3
Gend r Cat 0 N=3 2 Sample“ ansus "
(°/°) (‘70)
1. Male 40 54
2. Female 60 46

a - Survey respondents.
b - From 1990 Census

 

88

The sample is over-represented in the 14-17 age group and in ages 40
through 54, and 60 and over. The sample is under-represented between the
ages of 22 through 39. Nearly 60% of the respondents are over the age of 30.
This suggests that a relatively older group of residents were sampled.

Table 9 shows the educational attainment for persons 25 and over of
Latino origin in the sample. The observed frequencies do not equal the
expected frequencies. The observed chi square is 58.13 (sz (a. .05. d: a, =
12.59). Table 9 shows that the sample is generally over-representative of
education higher than a high school diploma. Half the respondents attained a
high school diploma or some college. These data suggest a wide range of

educational attainment.

 

Table 9

Education for Persons 25 and Over of Latino ri in in ensus and Sam Ie

Egucation Categog (N=271) Sam Ie' Census "
(%) (%)
1. Less than 9‘” grade 22 32
2. Some high school 13 24
3. High school diploma 33 21
4. Some college, no degree 17 13
5. Associate’s degree 9 4
6. Bachelor’s degree 3 4
7. Graduate school 3 2

a - Survey respondents.
b - From 1990 Census

 

89

Univariata Analysis.

Appendix A offers a breakdown of the distribution of responses for each
variable in the study. The check boxes in the English survey were replaced by
percentages. Means, standard deviations, and the range was provided for non-
categorical variables.

Ancestry. In this sample, 73% of Southbridge Puerto Ricans were born in
Puerto Rico. In addition, nearly 100% of all Puerto Ricans had parents which
were both Puerto Rican. There are almost no respondents of mixed ancestry.
This finding suggests a homogenous ethnic group. In addition, the majority of
respondents either emigrate or are of first generation status.

The data also suggest that cultural ties are not only reinforced by familial
ties, but by ties to the island of Puerto Rico. Neany 80% of the respondents
have lived in Puerto Rico. When comparing the average years (18.5) living in
Puerto Rico and in the United States (19.5) the means are very similar. The
standard deviations and the minimum and maximum years are also similar (see
Appendix A). This supports the view that there are strong ties to the island and
that Puerto Ricans in Southbridge can be generalized to other Puerto Ricans. In
addition, one could speculate that there is continuous migration back and forth
between Puerto Rico and Southbridge.

Agulturation ans subcultural identfl' . Because the acculturation
processes and cultural identity may play a role in recreation, the means of time
spent living in Puerto Rico by age cohort was examined. Table 10 is instructive

for several reasons. There is a significant difference between the means of the

90

various age groups (F=16.94, df = 11, ci= .000). The means show a positive
relationship between one’s generation or age (indicative of acculturation) and

time spent living in Puerto Rico.

 

Table 10

ANOVA of Years_i_n Puerto Rico by Age Cohort

 

 

 

 

_Aga N 2am Mean gs, Minimum Maximum
Qohort omen Deviation
14-17 32 57 6.22 4.53 1 15
18-21 17 53 10.71 6.59 1 20
22-24 12 66 7.42 5.16 1 19
25-29 35 66 14.43 7.00 2 26
30-34 46 80 17.57 9.08 1 32
35-39 29 76 18.69 10.14 1 37
40-44 35 88 22.91 9.72 1 38
45-49 34 100 24.53 8.70 10 40
50-54 19 95 25.89 9.80 14 41
55-59 14 100 27.36 10.30 16 47
__60+ 21 100 29.31 13.02 15 59
Total 294 77 18.54 11.14 1 59
Sum of Sguarg S1! Mean [3 _S_ig._
. Sguare
Between Groups 14481.701 11 1316.518 16.944 .000
Within Groups 2191 1.224 282 77.699
Total 36392.925 293

 

The percent of all respondents in each age cohort that have lived in
Puerto Rico is decreasing by each newer generation. For example, 53% to 57%
of pre-adult (ages 14-21) respondents have lived in Puerto Rico. This compares
to 66% for young adults in their twenties; 76% to 88% who are middle age (ages

30-44); and 95% to 100% of respondents over age 45 who have lived in Puerto

91

Rico. The descriptive evidence suggests that this study sample is reflective of
the population.

The implication of this finding, with respect to the EPRP Model, is that it
forecasts a strong relationship between acculturation and subcultural identity. A
limitation of this finding is that there is no data which would indicate what period
of their lives was spent in Puerto Rico. Knowledge of the period spent in Puerto
Rico might shed light on the relationship between acculturation, subcultural
identity, and discrimination.

Park use. The average park use was 74 visits per year to Southbridge
public parks (N=383). The park used the most was Henry Street ()1 = 24.01 ),
followed by Westville Dam (52 = 18.38), McMahon Field (52 = 13.91 ), Morris Street
(x = 8.77), and West Street (>7 = 8.48). Appendix B provides a map of
Southbridge and the spatial distribution of its parks.

Although age was not incorporated in the current model, there can be
significant difference in park use between students and adults. In the study,
Puerto Rican high school students have mean annual park visits of 124 visits per
person (N=57), compared with Puerto Rican adults who averaged 65 visits to
public parks (N=326). There is a relationship between park visits and whether or
not the visitor is a student or an adult ()8: 7.01, dfi1, or: .008, 2-tail). There is
also a significant difference between the means of student and adult park use (t
= -6.03, df = 381, or: .000, 2-tail). It can be argued that age or cohort could be a

factor on park use.

92

The researcher investigated whether students differed from adults (on
average) with respect to their views on the benefits of parks. Table 11 indicates
that there is no difference between students and adults in terms of how each
group views the benefits of parks as; (1) a place to for exercising, (2) a place to
socialize, (3) a place for kids to go, or (4) a place for providing open space.
Adults and students did differ on their perceptions of parks as a place for (1)
escape and (2) enjoyment, with adults showing stronger agreement. There was
barely a significant difference between students and adults on their perception of
parks as a place where family and friends could meet. The two top ranked
perceived benefits of going to a park by both groups were that (1) parks provide
a place for kids to go, and (2) parks provide a place where family and friends
meet. This is consistent with Huthison’s (1987) and Chavez’ (1993)

observations that Latinos are very family oriented.

 

Table 11

Mean Difference in PQEIVQQ Banefits pf Parks Between Students and Adults

Perceiv Tptal Studant Agult 1 g a
Banefrts Maan Maan Mean (twp;tail)

Enjoy nature 4.06 3.58 4.15 4.20 380 0.00
Escape 4.12 3.79 4.18 2.90 380 0.00
Socialize/contacts 3.95 4.07 3.93 -0.99 372 0.32
Exercise 4.08 3.88 4.12 1 .82 378 0.07
Family/friends 4.33 4.1 1 4.36 2.08 378 0.04
Open Space 4.09 4.09 4.10 0.06 379 0.95

For kids to go 4.46 4.46 4.46 0.04 380 0.96

 

93

Based on the researchers understanding of the Puerto Rican culture, the
Puerto Ricans in Southbridge are very family oriented. A strong family oriented
park atmosphere that is conducive to family and children-oriented activities
would be draw for Puerto Ricans. Henry Street park is the most visited park by

Southbridge Puerto Ricans. This is probably due to several reasons:

. it is the closest park to where the majority of Puerto Ricans reside;

- structural changes (playground improvements) were made to
accommodate more children;

. cultural (Puerto Rican festivals) and family-oriented activities were
targeted at Puerto Ricans;

0 Henry Street Park has three baseball diamonds (the most for any park in
Southbridge);

. the tennis courts at Henry Street Park were changed to basketball courts.

These sports, baseball and basketball, are the number one and number two
sports, respectively, among Puerto Ricans.

Westville Dam, however, is the furthest from the majority of Puerto Ricans
in Southbridge. This park may be popular and have high visitation because it is
very unique. It is the only recreation area that has a river running through it.
This offers Puerto Ricans the opportunity to fish. Westville Dam is the only
place that has grills and picnic tables which are appealing to family oriented
Latinos (Hutchison, 1988) . Many church outings are also held there.

The reader should not make light of these findings. Puerto Ricans hold
their families (immediate and extended) at the core of their culture. Ties along
familial lines are very strong.

The surprising finding was the use of McMahon Field. Further inspection
reveals a large discrepancy between mean use by students (39.09) and mean

use by adults (9.52) which is reflective of its location. McMahon Field is located

94

across the street from Southbridge High School, and it is where most high
schoolers ”hang out“. It is also located in a more affluent part of Southbridge
where less Puerto Ricans live. In addition, the park is unique because it offers
the only track field and football field in the city.

As a final note, mean use of parks is misleading because it does not take
into account seasonal variation for each park. The distribution of the frequency
of use by season for each park (see Appendix A) shows that the mean is not a
good indicator of use. Lack of use in the winter, spring and fall months.
However if one looks at only the Summer, public parks are used more widely
and between 50% and 90% of respondents use the parks in some capacity.

D ta in and lean-u

The initial run of the correlation matrix between all the factors gave a
lower N than expected. The majority of housewives, students and retired
persons were left out of the analysis because they were not coded for
occupation. Following the coding for the Hollingshead Index, housewives were
coded "0" on their occupational score, students were coded "1” for their
occupation score, and retired persons were coded “2”.

Excluding cases listwise, the final N used for the CFA and path analysis
was 311. This included the housewives, students and retired persons. In
addition, not all items were used in the initial CFA run.

Income was not used in the socioeconomic construct because of reasons
listed earlier. The socioeconomic construct had two indicators: education and

occupation. Additionally, questionnaire item 6B (see Appendix A) was not used

95

as an indicator for the cultural identity construct. During the pilot test, this item
seemed problematic and it was kept after it was re-worded. However, the item
continued to present problems in the consistency of responses to it. The item
was not used because it was long and unclear.

Lastly, the items for Acculturation (3A-3G, see Appendix A) were re-coded
so that ”1" reflected ”only Spanish” and "5” reflected "only English". The
researcher realized that the more Spanish-oriented the respondent was, the less
helshe was acculturated. Therefore, the stronger one’s preference towards
Spanish, the lower that respondent should be in their acculturation process.

Section Twp: Cpnfimiatppr Factor Analysis

The model to be tested consists of six constructs: (1) Acculturation (ACC),
(2) Subcultural Identity (SID), (3) Socioeconomic Status (SES), (4) Perceived
Benefits of Parks (BEN), (5) Perceived Discrimination (DIS), and (6) Recreation
Participation (REC). The testing of this model will be done one construct at a
time. The initial CFA began with 35 items. This created a 35x35 matrix yielding
595 correlations. The N for the CFA is 311. Three criteria are used to assess
the fit of the model: (1) face validity, (2) internal consistency, (3) parallelism.

Face validity was assessed in the previous chapter and will not be
discussed further in this analysis. It is assumed, based on the results of the two
pilot studies, and the literature review, that all items meet the criterion for face
validity. The qualitative and substantive evidence suggest that all the items
measure their respective constructs. This data analysis concerns itself with the

latter criteria and addresses each accordingly with respect to each construct.

96

lntemal Consistensy

lntemal consistency examines how the items in a cluster (construct) relate
to each other (see Appendix C). The pattern of interrelationships should be
consistent with the hypothesis that all items measure the same construct. This
will be considered for each construct that has more than four indicators.
Constructs having less than four indicators cannot be tested for lntemal
consistency without making some assumptions. Unless otherwise noted, the
lntemal consistency tests allow for a gradient in the correlation matrix (allows for
variation in item quality, i.e., the factor loadings are not equal).

Psrceivfi benefits of parks. The initial BEN construct started with seven
indicators (see Items 1-7 in Table 13). The value of chi square was 62.42 with a
p-value of .00. The chi square is not near zero, and the p-value is not above
.05. This implies that the construct is not internally consistent.

The size of the errors for the original seven items were then assessed by
looking at the item with the largest errors (sum across the row for each item in
the correlation matrix). Next, the researcher observed what happened to the
sums when indicators were removed. If questionably large errors remain, the
last check is to see if the observed correlation is within the expected
correlation’s confidence interval for those correlations whose errors are large.

In Table 12, the obtained correlations are located in the bottom left
triangle of the first matrix, and the predicted correlations on the upper right

triangle in bold. Predicted correlations were derived by employing the product

97

rule and multiplying the factor loadings (designated “FL” in Table 12) for the
corresponding items. For example, the predicted correlation between Item 1 and
Item 2 is the product of the factor loadings for each item (.66 * .66 = .44). The
absolute difference between the obtained and predicted is then put in the error

matrix.

 

Table 12

Assessing Errors for BEN Const ct

 

    
    

 

Obtained/Predicted Correlations Error Matrix
1 2 3 4 5 6 7 )1

1 .44 .42 .50 .53 .46 .40 1 .34
2 .59 .42 .50 .53 .46 .40 2 * . . . . .34
3 .40 .43 .48 .51 .45 .38 3 .02 .01 ** .00 .01 .00 .01 .05
4 .51 .47 .43 .6 .52 .45 4.01 .03 .00 ** .03 .00 .03 .10
5 .4353 .52 .57 .5543 5.05 .00 .01 .03 ** .07 .01 .17
6 .45 .40 .45 .52 .56 .42 6 -0 .06 00-00-07 “” -06 20
7 .30 _31 .39 .43 .55 .49 7 f .01 .03 .01 .06 ** .30

 

 

FL .66 .66 .64 .75 .80 .70 .60

 

Looking at the error matrix in Table 12, one can see that Items 1, 2, and 7
have large errors associated with them. For a more thorough assessment, the
researcher viewed the confidence interval around the predicted correlation
between Item 2 and Item 7 (the lowest of the high correlations): p (.31 $.40 $.49)
= .95. The obtained correlation of .31 is barely in the confidence interval,

indicating that errors of .09 will not make this model fit the data.

98

Therefore, Items 1 (Enjoy nature at parks), 2 (Parks allow me to escape),
and 7 (Parks are a place for kids to go) were discarded. The modified BEN
construct consisting of Items 3, 4, 5, and 6 (see Table 13) yielded a chi square
of .306. The p-value was .998. The chi square is close to zero, the p-value is
greater than .05, and the errors are within sampling error of zero. The BEN
construct passed the internal consistency test with four indicators and it is
unidimensional (all items measure the same construct).

Table 12 illustrates the process which is repeated for each construct. The
significance test aside, each construct was submitted to the same process of
obtaining correlations, comparing them to predicted correlations, and assessing
the errors. As such, the obtained/predicted correlations and the error matrix for
each construct will not be shown. The reader should use Table 12 as a model.

Agulturatipn. The initial ACC construct consisted of 11 indicators, Items
8-18 (see Table 13). The original conception of the ACC construct consisted of
both language and value variables. The 11-indicator ACC construct had a chi
square value of 1217.53 and a p-value of .00. The enormous chi square and p-
value indicated significantly huge errors.

The factor loadings for Items 15-18 were very different in size from the
factor loadings of Items 8-14. This indicates that perhaps there are two
constructs. The first construct measuring acculturation in terms of language
preference, and the second construct measuring acculturation in terms of
values. Assessment of the error matrix showed that Items 15-18 contained large

errors, with individual item residuals having errors over .30. Items 15-18 were

99

omitted from the construct cluster and errors were assessed based on seven
items (Items 8-14). These seven items reflected acculturation based on
language preference and use. The factor loadings were re-estimated and the
chi square of the 7-indicator ACC construct was 252.16 with a p-value of .00.
Again, the chi square was considerably large and the construct had significant
errors. Assessment of the correlation matrix led to the omission of Items 9, 10,
and 13 (see Table 13). Item 13 had the most error associated with it. Item 9
had the second most error and did not match the strength of the factor loadings.
Item 10’s largest error did not fall within the confidence interval of the predicted
correlation.

This left the ACC construct with four indicators. The abridged scale
yielded a chi square of 10.04, with a p-value of .07. The errors are within
sampling error of zero with this set of indicators, and the significance test states
that the errors are not significant. The ACC construct is internally consistent and
therefore unidimensional.

Perceived dissrimination. The DIS construct initially consisted of five
indicators, Items 19-23 (see Table 13). The five items yielded a chi square of
41.07 with a p-value of .00. The original construct had significant errors.
Assessment of the residual matrix showed that Items 19 and 20 had the same

summed total of errors.

100

Table 13

Descriptive Statistig For ltams Used in the CFA (N = 311)

Standard
Mean

Construct

Deviation

Perceived
Benefits

c
.9
..

E

:3
+_-
g

Perceived
Discrimination

Subcultural
Identity

Public
Recreation

 

economic

' are

" - Shaded items used in final confirmatory factor analysis.
‘ - Shaded items passed tests for lntemal consistency and parallelism.

101

When each item was individually removed from the scale to assess how
each affected the factor loadings, the factor loadings were more uniform with
Item 19, than with Item 20. In addition, Item 20 did not match the initial factor
loading gradient, and it’s highest error did not fall within the predicted
correlation’s confidence interval. Item 20 was removed from the scale.

The four indicators measuring DIS were Items 19, 21, 22, and 23. This
yielded a chi square of 6.84 with a p-value of .23. The errors are not
significantly different from zero and the construct is internally consistent yielding
a unidimensional model.

Subcultural ident’ny. SID was measured by five indicators, Items 24-28
(see Table 13). The chi square for the five indicator model was 31.03 with a p-
value of .00. The model was not internally consistent and had significant errors.

Assessment of the errors disclosed Item 28 as having the most error
associated with it. Item 28 also failed the confidence interval test. With the
deletion of Item 28, the SID construct’s chi square became .56 (which is very
close to zero), with a probability of .99. This indicates almost perfect internal
consistency and one can conclude that the four indicators are indeed measures
of the same construct. The construct is unidimensional.

Public racreation. The REC construct consisted of five indicators, Items
29-33 (see Appendix C). The five items yielded a chi square of 42.39, with a p-
value of .00. The initial REC construct was not unidimensional.

Assessment of the residual matrix found that absolute errors larger than

.09 would place the obtained correlation outside of the predicted correlation’s

102

confidence interval. Item 29 had two larger errors than .09 (.11 and .13). With
the deletion of Item 29 from the set of indicators, the chi square became 6.19
with a probability of .29. The four items did not have errors significantly different
from zero. The REC construct is unidimensional.

Socioeconomic status. SES could not be tested for internal consistency
because it only has two indicators. A construct with less than four indicators is
under-identified. This means that one cannot estimate the parameters unless
one makes an assumption, mainly that the parameters (factor loadings) are both
equal. One cannot test the fit of the model (lntemal consistency) with an under-
identified status.

Test of Parallelism

After the internal consistency analysis, there were 23 total indicators.
There are 253 correlations among 23 indicators. One should expect 12 or 13
errors to be large by chance alone. The biggest error found in the matrix was
.23. The correlation associated with that is -.09. The sample size is 311. The
standard error is .06. That means that the 95% confidence interval around that
correlation is :t .12. The reader should remember that this is in the parallelism
blocks of the matrix. Parallelism is calculated using three correlations, all of
which are subject to sampling error. Given this, an error of .14 may not be so
unusual

Errors that were .15, or above, were considered large given the sample
size of 311 and the sampling error associated with the three correlations used in

estimation. There were seven correlations which were .15 or above, and four of

103

the highest (.16, .17, .19, and .23) were associated with Item 33 (Harry J.
McMahon Field). Through inspection of the correlation matrix alone, Item 33
would not pass the parallelism test because of the high amount of errors
associated with it and items in other constructs.

The researcher inspected the original correlation matrix and noted that
Item 33 correlates higher with other items in other constructs. This is a violation
of parallelism. Therefore, Item 33 was deleted from the analysis. The final
parallelism check was conducted on 22 items. The next step is to perform a
CFA and test for parallelism with the remaining 22 items.

The test for parallelism is quite involved. A test of parallelism for the SES
construct will be presented to illustrate the process. After the SES example, a
parallelism synopsis for all constructs will be provided.

The parallelism theorem states that the correlation between items of two
different (parallel) constructs should equal the triple product of the factor
loadings for each item and the correlation between the constructs. To the extent
that this does not hold true, there is error. The test of significance for the
parallelism theory tests the idea that the errors are within sampling error of zero,
i.e., the errors are not significantly different from zero. However, more important
than the significance test is the analysis of the errors.

§E§ Qpnstgig as an exampla pf parallelism. Table 14 presents the factor
loadings and factor correlations (from Program CFA output) needed to obtain the
predicted correlation between items of different constructs in order to perform

the parallelism test. To check whether the items for SES are parallel to the

104

items for BEN one assesses the correlations between the items for each

construct.

 

Table 14

Item py Factor and Factor by Factor Correlation Matrix ' ”

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BEN A C - -DIS ID REC SES

3 4 5 I3 8 liifiTZ '14 19 I21 22 I23 24 I25 26 I27 30 31 32 34 35

Le» lssjsrs I.» I." -.31 -.02 .15 l .38 [-27
Incc -.31 .14 I." I." I.» -.48 -.69 .09 .27
DIS '02 --48 .70 Isa so I.” ' 53 -.12 -.16
SID .15 -.69 .53 .54 I.» lsrz [.ss .05 -.26
REC .38 .09 -.12 .05 .52 I.“ .si -.10
p53 -.27 .27 -.16 -.26 -.10 e... I...

 

 

 

 

 

 

- - the factor loadings of each item on each factor (construct) is in bold/smaller font
-. the correlations between each factor (construct) is in normal font (off-diagonal)

 

The following is the obtained correlations between the items for SES (Items 34

and 35) and BEN (Items 3, 4, 5, and 6):

3.4.
-.18

-.09
-.16
-.04

I0) I01 I-h I0)

35
-.16
-.13
-.22
-.20

The predicted correlations from the parallelism theory are obtained by

multiplying the factor loading for the first item by the factor loading for the

second item, and multiply this product by the correlation between the constructs.

105

 

For example, to obtain the predicted correlation between Item 3 and Item 34,
one multiplies Item 3’s factor loading (.65) by Item 34's factor loading (.75) by
the correlation between BEN and SES (-.27). The triple product [(.65)(.75)(—.27)]
is -.13. One does this for each of the correlations. This yields the following

predicted correlations and corresponding errors:

Pgdicted Correlations: Absolute Errors:
a 35 a 352

3 -.13 -.13 _3 .05 .03

5 -.15 -.15 _4 .06 .02

a -.16 -.16 5 .00 .06

_6 -.14 -. 14 _6 .10 .06

The errors are not large and one can conclude, by the assessment of the
errors, that the items for BEN (Items 3, 4, 5, and 6) are parallel to the items for
SES (Items 34 and 35). This operation is performed for every set of items on
every construct. The next step is to see if the SES items are parallel with
respect to the ACC items, and so on. Appendix D presents the final 22 item
correlation matrix used to assess parallelism. It illustrates the errors in the lower
left portion of the matrix and the predicted correlations in the upper right corner
of the matrix.

Once SES is assessed for parallelism among all the other factors, a “total”
parallelism measure is computed by Program CFA. This is the extent to which
SES (or any other construct) is parallel to all other constructs. It is also a chi
square test and is interpreted in the same manner as the chi square for the
lntemal consistency theorem. The null hypothesis is that error is equal to zero.

If one fails to reject chi square, then the items in the construct are not affecting

106

the items in the other constructs. Table 15 provides the chi square analysis of

the parallelism given by Program CFA.

 

Table 15

Parallelism Ana sis

Construct th Sguare P-Value Gradient/Flat
BEN (Items 3.4.5.6) 11.20 .74 Gradient
ACC (Items 8,11,12,14) 22.62 .09 Flat

DIS (Items 19, 21.22.23) 23.05 .08 Gradient

SID (Items 24.25.26.27) 23.88 .07 Flat

REC (Items 30.31.32) 16.94 .08 Flat

SES (Items 34.35) 9.02 .11 Gradient

 

The parallelism theorem held for all constructs (T able 15). The p-values are
larger than .05. Therefore. the errors are not significantly different from zero.
The model fits.

In addition to the significance tests, the errors for all the correlations in
the off-diagonal cells of the 22-item correlation matrix (parallelism blocks) were
within sampling error of zero. Thus the test for parallelism held for all items and
all constructs. Once the constructs pass the internal consistency and parallelism
tests. they are considered valid and can be summed to create a composite
variable. The reliabilities of each of the constructs is given in Table 16. The
next section discusses the steps taken in the path analysis to test the fit of the

causal model.

107

 

Table 16

R Ii bil’ An sis

 

Construct Standardized Item Alpha
BEN (Items 3,4,5,6) .81
ACC (Items 8,11,12,14) .89
DIS (Items 19, 21.22.23) .83
SID (Items 24.25.26.27) .70
REC (Items 30.31.32) .63
SES (Items 34,35) .72

 

Section Three: Path Analysis on EPRP Model

In the path analysis, the constructs are now considered (composite)
variables in the EPRP Model. Program Path is the software program used to
perform the statistics for the path analysis. It was initially developed by Hunter
(May 3, 1984) and revised by Hunter and Hamilton (August 17, 1997). The main
input for Program Path is the correlation matrix (taken from Program CFA) and
the specification of the path model. From this, Program Path’s main output is the
path analysis itself which gives the original correlation matrix. the reproduced
correlation matrix, the error matrix and the chi square and p-value.

Path analysis is based on a set of structural equations. The precise
mathematics behind path analysis will not be discussed here. For a more
comprehensive explanation of the mathematics behind causal models, see
Duncan (1975), Heise (1975), Kenny (1979), Kline (1998), and Loehlin (1987).

It is important, however, to have a basic understanding of the concepts involved

108

with path analysis in order to understand this section and the interpretation of
the Program Path output.
Path AnaNsis Qongms

Path models use standardized data. Therefore, the path coefficients are
equivalent to regression coefficients. The path model is a null hypothesis. What
one is testing is that the data are consistent with the model, i.e., the errors
associated with the model are equal to zero.

Over-jdentificatipn status. In order to evaluate the model, one has to
estimate the parameters (path coefficients) and assess the fit of the model. To
estimate the parameters. the model must be over-identified. The identification
status of the model refers to the amount of information that one has in the data
set relative to the amount of information one has to estimate the parameters. A
model is over-identified when the number of correlations exceed the number of
parameters one has to estimate. Simply put, subtract the number of paths in the
model from the number of total possible correlations and that is the identification
status.

Total number of possible correlations between variables is derived by

using the following formula:

 

[CZ—k (2)
2

 

 

Where “k” is equal to the number of variables in the model. In the case of the
proposed EPRP Model, there are six variables. Using the formula, there are 15

correlations to estimate.

109

Next, one looks at the number of arrows (paths) in the model. There are
10 paths. The number of paths is the number of correlations that are
constrained (used up in estimating the path coefficients) to equal their obtained
value by definition. Where ever there is a direct effect postulated, that

correlation is used for estimating the parameters.

The remaining five correlations (meme... News, rAccmEC, [9.0553, and

rams) are not used up in estimating the parameters (see Figure 11). Their

predicted value is NOT constrained to equal their obtained value. These five
parameters are what one has to estimate. These are the correlations that are to
be tested to assess the fit of the model, i.e., it leaves five degrees of freedom
which can be used to test the fit of the model. If the model fits, the observed
minus the predicted value (errors) should be equal to zero.

Visualizing the correlation matrix. As one adds variables to a causal
string, the correlation between the first variable in the causal string and the last
variable will get progressively smaller as a result of the product rule. The
relevance of this is that if one orders the variables in the correlation matrix
according to their causal order, the correlations should decrease as they leave
the diagonal of the matrix. This is a diagnostic tool for assessing errors and

evaluating the logic of the model and its resultant fit.

110

Figure 11. The EPRP Model (with non-constrained paths)

”“3

 

   
   
 

Recreation
Participation

 

 

 

 

Local vsrsus global test. The five non-constrained correlations are the
basis for local tests of the model. Program Path calls the local tests the
individual link analysis. This is the test that examines whether the errors for
each non-constrained link is within sampling error of zero. The error obtained
for each local test, divided by the standard error of the difference is distributed
as a z-value. One then uses the 2 test to examine the hypothesis that the errors
are within sampling error of zero.

If the z-value is less than -1.96. or greater than 1.96, then it is a large
error at the .05 level. It is statistically significant at .05 (it should not be
statistically significant, i.e., not statistically different from zero). If 2 is not as low
as -1.96 or higher than 1.96, then the error is not statistically different from .05.

This is the 2 test and it is performed for each of the five links, giving us five

111

different standard errors. Program Path computes each of these standard
errors.

The global test is the test of the model as a whole. The global test is
distributed as a chi square. Chi square by definition is the distribution that is the
sum of squared normal deviates. In this case, there are five normal deviates
(five z-values). Program Path squares each of the five z-values and adds them
up, yielding a chi square with five degrees of freedom. The chi square should
not be statistically significant, if the model is to fit.

The importance of the local and global tests is that. if the model does not
fit the data, the tests allow one to see where the model failed. The model as a
whole could work because the chi square is a summed value, but perhaps one
or more of the links are faulty. It is possible to have no errors on four of the links
and have a huge error on one link. The global test would not pick this up“.

Correlations and path coefficients. The final concept that one should
understand for path analysis is the difference between a correlation and a path
coefficient. The correlations between constructs include direct. indirect and
spurious relationships. The path coefficient, however, is the measure of direct

effects only. The path coefficients control for indirect and spurious effects.

 

This is the advantage of Program Path over LISREL. LISREL does not give the local tests. it only gives the global
test It also does not give a significance test using corrected coefficients. Using uncorrected correlations. the model
falls every time (Frank J. Boster, personal communication, February 9, 1999). LISREL uses maximum likelihood
estlrnation, while Program Path uses ordinary least squares. Although maximum likelihood is a more efficient
estimate. it spreads the error across the correlation matrix. As a result, one can never figure out where one went
wrong. Ordinary least squares localizes the error. It is important to look at both the local and global tests to see
where the model could be improved. The most important step Is to diagnose the errors. The significance test is a
slrnple decision rule, but the errors are what really matter.

112

The Path AnaNsis

The first step in the path analysis was to estimate the parameters for the
non-constrained links in the proposed EPRP Model. The corrected correlation
matrix used for the Program Path analysis is shown in Table 17.

Although Program CFA corrects for the correlations among the
constructs, it does not correct for attenuation in the correlations among the
indicators. The researcher used SPSS 9.0 for \MndowsO to compute the
uncorrected correlation matrix between the constructs. Using the Iistwise option,
the matrix with the correlations among the constructs yielded an N of 318 for the
22 items that make up the constructs. When the correlations went from 35 items
to 22 items, the sample increased. The path analysis, therefore, is based on an

N of 318.

 

Table 17

Obtained Corralatign Matrix for Path Analysis (N=318)
.899. fig §_E_§ 91$ BEN BED

M 1 .00 -.69 -.29 -.48 -.34 .09
§l_Q -.69 1.00 -.27 .55 .16 .08
§_E_S .29 -.27 1.00 -.16 -.29 -.13
_D_I§ -.48 .55 -. 16 1.00 -.02 -.10
_B_E_N -.34 .16 -.29 -.02 1.00 .38
REC .09 .08 -.13 -.10 .38 1.00

 

The uncorrected construct matrix from SPSS was used with the

reliabilities for each construct provided by Program CFA. This is the equivalent

113

of correcting for attenuation. The correlation matrix from SPSS, corrected for
attenuation by the reliabilities from Program CFA is more reflective of the exact
structure of the data file.

The correlations have been corrected for attenuation. To test the fit of the
model. one has to assess the over-identified paths in the model. The over-
identified paths are illustrated by the dotted grey arrows, and the identified paths
are illustrated in the solid black arrows in Figure 12.

The five non-constrained links were used to test that the errors in each of
the non-constrained paths are equal to zero. Confidence intervals around the
predicted correlations were used to assess the influence of sampling error.

They are as follows:

- 1-(-.20)2 / s/(317)=.054*(1.96)=.11 --> p(-.31 $-.20 $-.09)= .95 (SES/SID)
- 1- (-.14)2 I s/(317)=.055*(1.96)=.11--> p(-.25 $-.14 $-.03)=.95(ACCIBEN)
. 1- (-.38)2 I 7(317) =.048*(1.96)=.09--> p( -.47 $-.38$-.29)=.95 (ACC/DIS)
- 1- (.08)2 I s/(317) =.056*(1.96)=.11 --> p( -.03 $.08 $.19)=.95 (DIS/BEN)
- 1-(-.06)2 l\/(317) =.056*(1.96)=.11-—-> p( -.17 $-.06 $.05)=.95 (ACC/REC)

The obtained correlations (see Table 17) for ACC/BEN. ACC/DIS, and
ACC/REC are not within the confidence intervals. This indicates problems with
the causal model as it has been conceived. Program Path’s local tests and
global test for the analysis are as follows:

Z value between SES and SID is -.64 (prob = .52);
Z value between ACC and BEN is -2.22 (prob = .03);
Z value between ACC and DIS is -1.17 (prob = .24);
Z value between DIS and BEN is -1.08 (prob = .28);

2 value between ACC and REC is 1.49 (prob = .14);
Chi square for overall fit is 10.98. df=5, p-value is .07.

114

Focusing on the above 2 scores and chi square, the model does fit the
data, but there are problems with the model. The local test between ACC and
BEN failed. This is evident because the z-value is less than -1.96, and the p-
value is less than .05. In addition, the z-value of -1 .17 for the ACC/REC
correlation is close to the 1.96 and -1.96 marks. Although considered significant
by the z-test. these links failed the confidence interval test. The overall model
fit the data, but there are some weak links. The probability for the overall model
is greater than .05, and the obtained chi square is smaller than the critical chi
square (df=5) of 11.07.

Path analysis allows one to assess the errors and path coefficients, and
consequently. respecify the path model. Observing the original error matrix,
there appeared to be considerable problems with how ACC was related to DIS,
BEN. and REC. There also was a large error with how DIS and BEN are related.
These huge discrepancies indicate that perhaps these relationships should have
been direct relationships to account for the large error. The researcher modified
the paths to reflect direct relationships between each of the highlighted

relationships in Table 18.

115

 

Table 18

Matrices for Original EPRP Model from Program Path

 

 

   

Prflictfi correlations/Error Matrix‘ Path Coefficients
ACC SID SES Dis BEN REc ACC SID SES DIS BEN REC
ACC . -_, . Acc
SID -.69 -.07 .oo .02 . SID -.69
SES .29 -.2o -.04 .01 . SES .29 .38 ,
Dis -.38 .55 -.12 . DIS .00 55 :61
BEN -.14 .14 -.28 .08 .02 BEN .00 .09 -26 00

REC -.06 .07 -.13-.06 .36 REC .00 .09 £53- 14 35
'-Error Matrix in Bold

 

Because path coefficients indicate the direct paths, these coefficients
should be faiiiy ample (greater than .10). The size of the direct paths for
SES/DIS (-.01), and SES/REC (-.03), are relatively low coefficients when
compared to the other direct paths. These direct paths were omitted in the next
path model. The researcher decided to keep the two direct paths for SID/BEN
(.09) and SID/REC (.09) to see how the path coefficients would be affected with
the respecified model.

A clearer picture of causality emerges as a result of the re-specification
procedure. Thus. four additional direct correlations are introduced into the
model, and two direct relationships are expunged from the model. One can see
from the original correlation matrix that the direct correlations between ACC/DIS
(-.48). ACC/BEN (-.34) are ample. However, the correlations between ACC/REC

and DIS/BEN (-.02) are minor.

116

Keep in mind that correlations consist of direct. indirect, and spurious
paths. If paths creating the correlation are both negative and positive, the
combination could yield a small number. That is why the errors and path
coefficients are assessed. The assessment is used as a basis to respecify how
the model should have been entered into Program Path. Anderson and Gerbing
cautioned the following:

we recognize that most often some re-specification of the

measurement model will be required. It must be stressed,

however, that re-specification decisions should not be based on

statistical considerations alone, but rather in conjunction with

theory and content considerations. Consideration of theory and

content both greatly reduces the number of alternate models to

investigate and (cf. Young, 1977) reduces the possibility of taking

advantage of sampling error to attain goodness of fit (1988, p.

416)

The implication of this is that the reconceptualization (revision) has to make
substantive (theoretical) sense as well as statistical sense.

In summary, the proposed EPRP Model was consistent with the data.
However, as a result of the individual link analysis, the evaluation of the error
matrix, and path coefficients, it was found that the EPRP Model’s constructs are
related to each other, but some of the hypothesized links were weak or
hypothesized incorrectly. The path analysis allowed for a clearer picture to
emerge. As such, a revised. theoretically consistent model involving the same
constructs, and similar causality, was developed to test if the current insights

hold We The next section discusses the insights which led to the revised

model.

117

The Revised Model

Substantively speaking, there is not much difference between the
proposed EPRP Model (Figure 10) and the revised EPRP Model (Figure 12).
They both hypothesize a similar array of relationships. The major difference is
that in the initial EPRP Model, some relationships which were considered direct,
were in fact indirect or spurious. and some relationships which were conceived
as indirect were in fact direct. Figure 12 illustrates the revised version of the
EPRP Model as a result of the initial path analysis. The major differences
between the initially proposed model and the revised model are from three
variables: ACC, SES, and SID.

In the revised model, ACC has a direct effect on both BEN. DIS, and
REC. This relationship was indirect and spurious in the original model. In
addition, the causal arrows between SID/BEN and between SES/BEN were
omitted. The revised model hypothesizes an indirect relationship between SES
and REC, rather than direct, as originally hypothesized. Lastly. the revised
model introduces a new relationship between DIS and BEN. They are

hypothesized to be directly and spuriously related.

118

Figure 12. The EPRP Model (revised)

 

 

___ _ Parcaifllad __ _ _
\ Discrimination

 

 

 

The revised model hypothesizes acculturation to be the direct antecedent.
or driving force, behind the other constructs which are hypothesized in the
literature as having an effect on recreation participation by ethnic group
members. The extent to which one is acculturated has an impact on one’s
socioeconomic standing, subcultural identity, perception of discrimination, and
one’s perception of the benefits of parks, and actual recreation participation in
the parks. The implications of this will be discussed later. This revised model
must first be subjected to a path analysis to see if it is consistent with the data.
Path Analysis for Revised Model

The revised model has three non-constrained links to test. The following
are the three correlations used to test the model: rsmes, rsems, and rsssmsc.

The matrices given from Program Path is shown in Table 19.

119

 

Table 19

Matrices for EPRP Model from Prpgram Path (revised)

 

Predicted correlatignlerror Matrix‘ Path Coefficients
ACC SID SES DIS BEN REC ACC SiD SES DIS BEN REC

ACC .oo .oo .oo .oo .00 ACC
SiD -.69 -.07 .oo .02 .01 SID -.69
SES .29 -.2o -.02 .01 -.os SES .29 .oo
DIS -.48 .55 -.14 .oo .oo DIS -.19 .43 .00
BEN -34 .14 -.3o -.03 .01 BEN --44 as --22 52.3
REC .09 .07 -.07 -.1o .37 REC -48 -36 .00 £05.38

' - Error Matiix in Bold

 

Considering the non-constrained paths, one tests the hypothesis that the
errors in each of these paths are equal to zero (see error matrix in Table 19).
Assessing the influence of sampling error, the confidence interval around the
predicted correlations for the non-constrained paths are as follows:
- 1-(-.20)2 l J(317)=.054*(1.96)=.11--> p(-.31 $-.20 $-.089) =.95(SES/SID)
- 1- (-.14)’ I \/(317)=.055'(1.96)=.11—-> p(-.25 $-.14 $-.03)=.95(SESIDIS)
- 1-(-.07)2 / ~/(317)=.056*(1.96)=.11-—-> p(-.18 $-.07 $.04) =.95 (SES/REC)
All of the obtained correlations are within the confidence intervals of the
predicted correlations. As a further test, the model was subjected to Program
Path’s significance tests for the local tests and global test, they are as follows:
Z value between SES and SID is -.64 (prob = .52);
Z value between SES and DIS is -.17 (prob = .86);

2 value between SES and REC is -.55 (prob = .58);
Chi square for overall fit is .74, df=3, p-value is .86.

Note that all the z-values are greater than -1.96 and less than 1.96. In

addition, the probabilities for the individual link analysis are all estimated above

120

 

.05. The overall chi square is very much smaller than the chi square for the
initial model (.74 versus 10.09) with a p-value of .86 (as compared to .07 for the
initial model). Both the local and global tests, as well as the confidence interval
tests confirm that the model is consistent with the data, and the model is not
rejected. The hypothesis that the data are consistent with the model is
supported. If the model fits, then all correlations are significant.

Although significant, the enor matrix and the path coefficient matrix were
assessed to discern if the model could be made stronger (see Table 19).
Rudimentary inspection of the error matrix reveals that the errors are small.
However. the path coefficient matrix indicates two relatively low coefficients for
SID/BEN (-.08) and DIS/REC (-.05). These path coefficients should be
substantial (relative to the other path coefficients) because they are direct links
and the errors are not different from zero. This indicates that perhaps the
relationships are indirect, rather than direct. One final path analysis was
performed on the data.

The Final Revised Model

The difference between Figure 12 and Figure 13 is that there is no direct
arrow from SID to BEN, and no direct arrow from DIS to REC. Substantively
what this means is that the previous direct relationships are now indirect.
Perceived discrimination has a direct effect on the perceptions of benefits of
public recreation rather than on actual participation. Actual participation is

influenced by discrimination via discrimination’s effect on perceived benefits. In

121

addition, subcultural identity does not affect perceived benefits directly. It is re-

conceived as influencing perceived benefits with discrimination as a filter.

Figure 13. The EPRP Model (final version)

 

    

Discrimination

 

 

 

Path Analysis for Fi_n_al Revised EPRP Model

The final revised model has five non-constrained links to test. The

following are the five correlations used to test the model: rsmss. rswms, ram“.

rsemc’ and rmc. The matrices given from Program Path are shown in Table

20. The confidence intervals around the predicted correlations are as follows:
- 1-(-.20)2 I J(317)=.054 *(1.96)=.11—-> p(-.31 $-.20 $-.09) =.95(SESISID)
. 1- (—.14)2 I ~/(317)=.055 *(1.96)=.11--> p(-.25 $-.14 $-.03)=.95(SES/DIS)
- 1-(-.18)2 / ~/(317)=.054 *(1.96)=.11-—-> p(.07 $-.18 $.29) =.95 (SID/BEN)
- 1-(-.07)2 I J(317)=.056 *(1.96)=.11—-> p(-.18 $-.07 $.04) =.95 (SES/REC)
- 1-(-.06)2 I t/(317)=.056 *(1.96)=.11—-> p(-.17 $-.06 $.05) =.95(DISIREC)
All of the obtained correlations were within the confidence interval. The fit
of the model was also assessed using the local and global tests. The z-values

and chi square are as follows:

122

Z value between SES and SID is -.64 (prob = .52);
Z value between SES and DIS is -.17 E (prob = .86);
2 value between SID and BEN is -.19 (prob = .85);
Z value between SES and REC is -.55 (prob = .58);
Z value between DIS and REC is -.30 (prob = .76);
Chi square for overall fit is .87, df=5, p-value is .97.

 

Table 20

Matrices for EPRP Model from Prggram Path (final version)

 

 

 

2mm correlationsfiror Matfl‘ Bath Coefficients
ACC SID SES DIS BEN REC ACC SID SES DIS BEN REC

ACC 1.00 .00 .oo .oo .oo .00 ACC .oo
SID -.69 1.00 -.07 .oo -.02 -.01 SID -.69 .oo
SES .29 -.201.00 -.02 .oo -.01 SES .29 .00 .00
DIS -.48 .55 -.141.00 .00 -.03 DIS -.19 .43 .00 .00
BEN -.34 .18 -.29 -.03 1.00 -.01 BEN -.40 -00 521 =25 -00
REC .09 .09 ..07 -.06 .39 1_oo REC .50 .34 .00 .00 .49 .00

" - Error Matrix in Bold

 

All of the z-values are near zero, and have p-values well over .05. In
addition, the overall chi square was .87. This chi square was slightly higher than
the first revised model’s chi square (.74). This is due to five degrees of freedom
(versus the three degrees of freedom in the first revised model). The p-value is
higher than the first revised model. This is indicative of less error and therefore
a stronger model. All of the path coefficients are substantial, and the errors are
within sampling error of zero. The model is consistent with the data. The final
model is the best possible model given substantive and empirical measures.
Figure 14 illustrates the final model with the corresponding path coefficients. A
comparison of all three models and their respective errors, z-values, and p-

values is included in Appendix D.

123

Interpretation of the Model

Figure 14 represents the path coefficients in the model. The reader
should keep in mind that in a standardized linear one-predictor case, the
parameter estimate or path coefficient (beta weight) is the correlation. The
indirect and spurious paths, however. also have to be taken into account when
assessing the full impact that one construct can have on another. As an
example, the direct relationship between ACC and DIS is -.19. However, DIS
has an indirect relationship between ACC via SID of (-.69*.43) = -.29. The total

effect of ACC on DIS is (-.19 + -.29) -.48.

Figure 14. The EPRP Model (final version)

 

 
  

of *- ~-19 PM
3 6 Discnmrnation

S bcult I
“1|;an um ] .34

 
   

 

 

 

This reflects both the direct and indirect effects that ACC has on DIS. If
one observes the obtained correlation matrix, one notices that -.48 is the
correlation between ACC and DIS (see Table 19). The implication of this is that

40% (-.19/-.48) of the relationship between ACC and DIS is accounted for by the

124

direct relationship, and 60% is accounted for by the indirect relationship with

SID.

This process (statistical procedure) of path analysis can be considered as

a series of regressions in which each variable in the model is considered in turn

as a dependent variable to be predicted from its causal antecedent variables.

The path coefficients are equal to standardized regression coefficients (beta

weights). Therefore, the relative influence one variable (construct) has over

another variable in the model can be compared. Having mentioned that, the

hypotheses can be examined.

Examination of Hypotheses

Hypotheses
1. Subcultural identity [SID] and socioeconomic status [SES] are

significant predictors of public recreation participation [REC].
Subcultural identity will be a stronger predictor of park use than
socioeconomic status.

There is a significant relationship between perceived discrimination
[DIS] and public recreation participation, and perceived benefits of
recreation [BEN] and public recreation participation.

There is a significant relationship between the subcultural identity
factor [SID] and socioeconomic status [SES].

There is a significant negative relationship between acculturation
[ACC] and subcultural identity. There is a significant positive
relationship between acculturation and socioeconomic status.

There is a positive relationship between subcultural identity and
perceived discrimination.

There is a negative relationship between perceived discrimination
[DIS] and public recreation participation [REC].

The path model hypothesis:

125

Ho: the data are consistent with the model (Error = 0).
H,: the data are inconsistent with the model (Error :0).

Subcultural identity and socioeconomic status on glblic recreation usage.

SID and SES are significant predictors of park use (REC). Subcultural identity
was hypothesized as a stronger predictor of park use than socioeconomic
status. Looking at Figure 14, one can see that the correlation for SES/REC via
BEN is (-.21 *.49) -.10, and the correlation for SID/REC is the correlation via
DIS and BEN and the direct relationship between SID and REC: [(.43 *-.25 *
.49)+(.34)] = -.29. The hypotheses are supported. Subcultural identity and
socioeconomic status are significant predictors, and subcultural identity is the
stronger predictor.

Perceived discrimination and perceived benefits on park use. There is a
significant relationship between perceived discrimination and park use, and
perceived benefits of parks and park use. The relationship is not as originally
hypothesized in terms of its causality, but nonetheless the relationships between
DIS and park use, and BEN and park use are significant. As Figure 14
indicates, BEN is a significant cause of park use with a .49 effect. Perceived
discrimination has a lesser (-.12), but still significant effect on recreation
participation. There is a negative relationship between the perceived
discrimination and park use, and a positive relationship between perceived
benefits and park use. These correlations support the hypotheses that the
relationship between perceived discrimination and park use is negative, and the

correlation between perceived benefits and park use is positive.

126

Sufiultugl ident'y and socioeconomic status relationship. There is a
significant relationship between the subcultural identity factor and

socioeconomic status. The path model would support this hypothesis. The
correlation between SID and SES is (-.69 * .29) -.20. This finding offers support
for the marginality and ethnicity theories. According to this model, they are
spuriously related. From a substantive view, the relevance of this is that, in
previous studies, they were often viewed as separate or dissimilar to warrant
different studies or approaches. There is a strong relationship between the two,
especially given the antecedent of acculturation as hypothesized by the
researcher.

Acculturation/subcultural identity and acculturation/socioeconomic status.

 

There is a significant negative relationship between ACC and SID, and there is a
significant positive relationship between ACC and SES. The model clearly
presents the highest direct (and negative) correlation between ACC and SID as
-.69. ACC also has a strong positive relationship (.29) with SES. These
hypotheses are supported and acculturation has a significant impact on
socioeconomic status and subcultural identity.

Subcultural identity and perceived discrimination. There is a positive

 

relationship between subcultural identity and perceived discrimination. The path
model reveals a .43 correlation between SD and DIS. The hypothesis that

subcultural identity positively influences perceived discrimination is supported.

127

mm Fin i

This summary is organized in a fashion consistent with path analysis,
whereby one discusses the individual links, and then provides an overall
assessment. The researcher first discusses the individual constructs and the
relationships among those constructs. This provides a comparison to results of
other studies. After the constructs are discussed, the researcher provides a
summary of the overall model, and discusses how to de—construct the model to
compare it to other studies.

A ti n rceiv nefits and erceived discrimination.
Acculturation, as conceived in this study, reflects a preference by Puerto Ricans
to maintain the use of Spanish as the primary mode of communication. The
assumption is that this indicates a preference to maintain the heterogeneity of a
Puerto Rican subculture within the US. supra culture. As such, it indicates the
extent to which a Puerto Rican, as a member of an ethnic group, is assimilated
into the mainstream. The indicators were similar to those used in Floyd, et al.’s
(1993) study.

The current study supported Floyd, et al.’s idea that acculturation (cultural
distance) plays a critical role in the relationship between one’s subcultural
identity (intergroup distance) and socioeconomic status (socioeconomic
distance). Although the concepts were operationalized differently, they
nonetheless held true. Acculturation was found to be an important causal

antecedent to several variables in the EPRP Model.

128

A major finding regarding acculturation was the strong direct effect on
perception of park benefits (-.40) and perceived discrimination (-.19). Perceived
discrimination was found to negatively affect public recreation, as expected.
However, the direct relationship was not as robust as the other direct
relationships. The path analysis revealed the relationship to be indirect.

Perceived benefits of the recreation environment have not been explicitly
looked at in previous studies. Often times, it is assumed that there are
perceptions of parks, but those perceptions are often not measured because it is
a "given.” The model illustrates the importance of measuring perceived benefits
because it is a mediating variable between three different causal processes.
The negative relationship between acculturation and perceived discrimination
supports a marriage between West's (1989) discrimination theory and Floyd, et
al.’s (1993) acculturation theory.

The measurement of values was initially incorporated in this study as
suggested by Washbume (1978) and Hutchison (1987). It differed from
Hutchison’s model (see Figure 6) in that Hutchison conceived of ethnicity/race
as an antecedent to values, whereas, this study argues the opposite. Cortes, et
al. (1994) specifically argued for language and value items in the formation of a
unidimensional acculturation scale. This study did not support that argument.
The value and language items, some of which were taken from Cortes, et al.’s
study, did not pass the internal consistency test for unidimensionality. As such,
items measuring values were left out of the analysis and not incorporated into

the acculturation construct.

129

Socipflnomic status and subcultural identity. These two constructs
reflect the marginality and ethnicity theories, respectively. The inclusion of both

constructs is consistent with the recent Iiterature’s embracement of both
theories. Washbume (1978) mentioned that reasons for underparticipation were
not simply due to SES, and forecasted declining utility for SES. While this study
supports the notion that participation is not due solely to SES, the study does not
support declining utility for SES.

This study supported the idea of an interaction effect between SES and
SID, initially considered by Hutchison (1987). According to the model, the
relationship between SES and SID is spurious. As hypothesized, and in support
of Stamps and Stamps (1985) and Washbume, ethnicity was found to be a more
important predictor than marginality. In addition, it was found that both SES and
SID’s effect on using public parks was minimal compared to the effect of other
factors in the model.

SID was operationalized a little differently than in Taylor's (1992) study.
She looked at ethnic expression of leisure activities and inferred SID from this.
This study took a more subjective and direct approach and solicited information
reflecting ethnic expression (see questions measuring SID, Appendix C).

The SID measure was also more robust than West’s (1989) measure of
subculture, and was similar to Floyd, et al.’s (1993) conception. West found no
support for the relationship between subcultural identity and recreation in his

study. This study did find support for the relationship, albeit a spurious one.

130

Socioeconomic status was measured using an adopted version of the
Hollingshead Two-Factor Index of education and occupation which was used in
Stamps and Stamps’ (1985) study. This differed from Floyd, et al.’s study which
only used education as an indicator of SES. Unlike Stamps and Stamps, this
study did not collapse the sample into upper and lower classes, but instead
maintained the two indicators as reflective of one construct.

Public reereetion participation. Unlike previous studies that observed
what people do at parks, or asked about which recreation activities were
participated in, this study focused on frequency of visits to parks over a one year
time period. Park visitation was used as a proxy for public recreation
participation. The interest in this study was in park use as measured by
visitation. The advantage of measuring park use in this manner is that one can
procure users and non-users of the parks. Not all parks were used because
they did not meet the internal consistency criterion for unidimensionality of the
public recreation participation construct.

Chapter Summeg

The EPRP Model was submitted to stringent validity tests of internal
consistency and parallelism. The constructs held and were used as summative
variables in a path analysis where the model was then analyzed to estimate the
parameters and test for the fit of the model.

The EPRP Model’s initial conception in terms of direct or indirect causality
was respecified after assessing path coefficients, errors, and confidence

intervals. The result was a revised model (Figure 14) which was statistically

131

consistent with the data, and substantively consistent with the theoretical
underpinnings of this study.

The result of this section was the testing of what became a sound
theoretical and statistical model: The Ethnicity and Public Recreation
Participation Model. The next, and final chapter summarizes the major points of
this study, discusses the deconstruction of the EPRP model, and considers the

implications for future studies.

132

Chapter 5
SUMMARY AND CONCLUSIONS

The purpose of this study was to identify factors which were important in
examining the relationship between ethnicity and recreation behavior. This was
accomplished in one of two ways. The first approach was to create a theoretical
model that underscored the relationships between the different factors as
hypothesized in the literature regarding ethnicity/race and recreation.

The researcher identified six factors: (1) ethnicity (subcultural identity), (2)
marginality (socioeconomic status), (3) acculturation, (4) perceived benefits of
public recreation, (5) perceived discrimination, and (6) public recreation
participation. Based on prior empirical studies, the six factors were assembled
to provide the building blocks for the Ethnicity and Public Recreation
Participation Model. This led to the second approach in which the factors were
examined.

The theoretical model was subjected to a confirmatory factor analysis to
validate the existence of the factors/constnrcts as measured by the items that
made up each construct. Once validity was assessed, the model was then
empirically subjected to a path analysis in order to estimate the parameters of
the EPRP Model, and test the fit of the model by assessing whether or not the
model was consistent with the data.

The researcher revised the model based on the path coefficient and error
analysis and found that the revised model was not different, conceptually, from

the initial model. As a result, this study provides a theoretically and empirically

133

sound model to consider in reconceptualizing the relationships between ethnicity
and marginality factors and their effect on public recreation participation.
mm

Dflpnstructing the EPRP Model: Reflections of previous literature. The
purpose of this section is to provide the reader with an in—depth look at the
nuances of the EPRP Model. Because the model is reflective of causality, and it
is statistically and substantively sound, all or parts of the model should provide
points of comparison to other studies. The researcher de-constructed the model
to illustrate, pedagogically, how the model reflects prior literature and
reconceptualizes the relationships. One method of approaching this de-
construction is to discern what would happen with the removal of one or several
parts of the model.

The two key constructs are acculturation and perceived benefits. The
former because it is a significant driving force in the model. The latter because it
is both central to the model, and it is a ”new" constmct advocated for and
introduced in this study. The removal of a construct does not change the
correlation between the constructs, rather the changes occur in the path
coefficients, i.e., in the nature of the direct relationships.

Rempvel pf ammatiop. The EPRP Model incorporates theories
espoused specifically by Floyd, et al. (1993), Washbume (1978) and West
(1985). These are acculturation, marginality and ethnicity, and discrimination,
respectively. If one removes Floyd, et al.’s contribution, the acculturation

construct as a direct driving force is extracted. The resulting path model is

134

exhibited in Figure 15. The curved arrow indicates that SID and SES are
spuriously related, but that one does not hypothesize a common antecedent. If

one compares Figure 15 with Figure 14, one sees that the path coefficients have

 

 

 

 

changed.

Figure 15. The EPRP Model without Acculturation
{Mama ' ° m
tatus -

Recreation

Ix

0.

-.27 '

Perceived
Discrimination
Subcultural

[Identity ,, ,, .02

 

 

 

In causal modeling, if any variable is thrown out of a model, it reduces to
a simpler model. The implication of this is that if the larger model fits, then the
resulting simple model or causal strings will also fit. The model in Figure 15 has
a chi square of 2.05, with df=4, and a p-value of .73.

Acculturation as a driving force behind DIS and SID increases the effect
that each of those constructs have on BEN and REC, respectively. The
inclusion of ACC in the model has the effect of lessening the effect SID and SES

will have on other variables (and themselves). A general rule of causal analysis

135

is that correlations are going to be progressively smaller as variables are farther
away from each other in the causal string.

The implication of this is that without consideration of ACC as the driving
force behind DIS and SID, the effect of DIS on BEN decreases to a coefficient of
-.07 (versus -.25 with ACC in the model). Unlike, the final EPRP Model, the
model without acculturation (Figure 15) would show a stronger direct link
between DIS and REC (-.10) than between DIS and BEN. Figure 15
reconceptualizes and combines the relationships in Figure 4 (Washbume) and
Figure 7 (West).

Removal of perceived benefite. Figure 16 reflects the combined concepts
of Figure 4 (Washbume), Figure 7 (West) and Figure 9 (Floyd, et al.). The
model has a chi square of .44, with df=2, and a p-value of .81. The following is a
comparison between the path coefficients in the full EPRP Model versus the

model without BEN in it:

Qaueal Path Model with BEN Model without BEN
ACC/SID -.69 -.69
ACC/SES .29 .29
ACC/DIS -.19 -.19
ACC/REC .50 .28
SID/REC .34 .31
DIS/REC --- -.16
SES/REC ---- .09

The comparison of the path coefficients is instructive for several reasons.

First, one can see that paths that were not affected by BEN were unchanged, as

136

the model should predict. Second, because the direct relationship between ACC
and BEN was eliminated (-.40), it attenuated the relationship between ACC and
REC. Third, because BEN is not in the model, the relationship between
DIS/REC is direct and it changes the relationship between SID/REC slightly.
Related to this, the relationship between DIS/REC and SES/REC were not
defined as direct in the EPRP Model, but they are when the BEN mediator is

removed in order to be consistent with causality. Lastly, this is the first model

Flgure 16. The EPRP Model without Perceived Recreation
Benefits

 

 

 

 

 

that presents a positive correlation between SES/REC. All prior models have

shown the relationship to be negative.

Removal of acculturation and perceived benefits. If ACC and BEN were
removed from the model, the result would be a model reflecting the three

underlying theories of marginality, ethnicity, and discrimination (see Figure 17).

137

It comes closest to Stamps and Stamps (1985) model (see Figure 5), and unites
it with West’s (1989) model (see Figure 7). This is the ”basic” model as

conceived today by most of the researchers studying ethnicity and recreation.

Figure 17. The EPRP Model without ACC and BEN

 

-.27 Participation

Perceived _
Discrimination '20

 

       

    

 

 

 

 

When comparing path coefficients of Figure 17 with those of Figure 14,
the importance of the role of discrimination is noted it ACC and BEN are not
accounted for. The strong negative path coefficient (-.20) is the strongest it has
been in any of the models. This affirms West’s claim that discrimination is a
valid construct when looking at ethnicity and recreation. Without ACC driving
the model, or perceived benefits mediating relationships, the relative impacts of
SES and SID on REC are about the same. This provides insight into why
studies have been ambiguous in the past as to which of the two is the stronger
predictor of recreation participation. The coefficient of .15 between SiD/REC

does not support the notion that minorities are under-represented in public parks

138

because of ethnicity. This relationship is mediated by other factors as was
illustrated in the previous models. The result of this deconstruction is that
perceived benefits and acculturation are both important constructs to look at
when considering ethnicity and recreation. Through causal modeling, the
relative impacts of the inclusion or removal of a construct can be seen and
compared.

Conclusions

Based on the findings, and within the limitations of this study, the
researcher felt that this study answered the call from Floyd (1998) and
Henderson (1998) for a more integrative approach, and more complex
measurement of the relationships between ethnicity and recreation. The more
complex measurement came in the form of identifying critical dimensions from
the literature and empirically testing these dimensions through confirmatory and
path analytic approaches. This study found the relationships in the model to be
consistent with those hypothesized by the literature, although the nature of the
relationships are reoonceptualized.

From a conceptual standpoint, subcultural identity and acculturation were
reoonceptualized: the former in terms of how it is operationalized, the latter in
terms of how it related to other variables. Subcultural identity was
reoonceptualized as a subjective response to the extent to which one identifies
with a subcultural group. This is different from previous studies that simply ask
or denote (in observational studies) whether one is Black, Caucasian, or

Hispanic, and is consistent with Washbume’s (1978) initial conception.

139

Acculturation was reoonceptualized as a precursor to the entire model. In
previous studies it was conceptualized as occurring simultaneously with
socioeconomic status or subcultural identity.

The study answered the first research question regarding which of the two
factors (ethnicity or marginality) was the better predictor of recreation behavior.
In terms of their direct effect, ethnicity has a more profound effect. The second
research question considered the individual and combined importance of
ethnicity and marginality factors. The combined effect, as measured by the SES
and SID constructs, was found to be significant (see findings). The last
research question considered the relationship between acculturation, perceived
discrimination, and perceived benefits. initially, it was thought that the
relationship was indirect via subcultural identity and socioeconomic status,
respectively. The path analysis of the EPRP Model, however, supported the
idea that there were considerable direct effects from acculturation to both
perceived benefits and perceived discrimination.

The EPRP Model not only determined the relative influence of the factors
in the model (see Figure 14), but also offered a "snap shot" of the relationships
between the most prominent theories regarding ethnicity and recreation. As a
result, the EPRP Model contributes to the literature by offering a model for
practitioners and researchers to consider. The model is not exhaustive of, but
rather, instructive of possible factors influencing ethnic participation in public

recreation settings.

140

Shortcomings and Limitations

There are some shortcomings and limitations with respect to the model.
From a procedural standpoint, the study had to alter the original use of the SES
scale. Because individual income was not used, the household income was not
appropriate to use where education and occupation were individual measures.

Although the concepts are based on previous studies, variables were
operationalized differently. For example, subcultural identity in the current study
is operationalized as the extent to which a person identifies with his/her
subculture. In previous studies, it was a categorical variable asking for the
person to “check” his/her ethnic/racial background. Because variables were
operationalized differently, the reader should be careful when comparing
findings to other studies.

The dependent variable, REC, was limited to frequency of park visits. i
does not consider purpose of visit or the activities engaged in. Furthermore, it
was limited to three parks instead of five. Several factors could explain this.
Possibilities include the location of Morris Street Park or the heavy use by
students at McMahon Field. it may also be because of non-response bias. If
only 27% of the eligible Puerto Rican population was surveyed, the other 73%
could have a different view of those parks. Whatever the reasons, the construct,
includes 60% of the public parks in Southbridge. But, of the parks that did
remain in the construct, they are the closest or most heavily used parks by

Puerto Ricans in Southbridge.

141

Related to REC is the perceived recreation benefits constnrct. BEN
measures whether respondents perceive that, in general, they have benefits
versus what specifically those benefits are. As such, the BEN construct is a
measure was intended to be a measure of overall perceptions, versus individual
motivations for visiting a park.

The acculturation constnrct presents another limitation. ACC appeared to
be very central to the causality of the model. This may have been due to the
population under study. Puerto Ricans, in general, have a particular historical
relationship with the United States that is beyond the scope of this study.
Southbridge Puerto Ricans, in particular, have a very strong tie to Puerto Rico.
This may influence the acculturation process.

Lastly, the study population is from a specific ethnic group. Although a
Latino group, these findings should not be generalized to all Latinos. The
findings should also not be generalized to other ethnic group members. Some of
the items which reflect the constructs are specific to Puerto Ricans.
Acculturation, subcultural identity, and discrimination measures should reflect
the specific ethnic group under study.
lmplicatione

Congptuel implications. From a conceptual standpoint, the model
synthesizes past research and looks at the relationship between the different
factors. Marginality and ethnicity theories as explanations are insufficient, and

do not operate in a vacuum. They are affected by each other and other factors

142

in direct, indirect and spurious ways. The challenge is to try and make sense
out of what these relationships really mean.

This model also helps to fill a void in theory-building approaches to
studying ethnicity and recreation constnrcts. In particular, the EPRP Model
introduces the perceived benefits of park use/recreation as a major factor
intervening or mediating the relationships between SES, ACC, and SID on
recreation participation. This incorporates leisure theory into the model, a
theory which had not been directly incorporated in prior studies.

This study looks at affective or behavioral traits, rather than objective
manifestations. In addition, the perspective or point of reference is from the
actual ethnic group, and therefore the norm is the Puerto Rican norm. In
previous studies, subcultural group members are often compared to whites. This
assumes a normative structure based on white’s recreation. If one is to tnrly
understand the nature of ethnicity with respect to recreation, one should study
ethnic groups without using Whites as a standard. The comparison to Whites or
other subcultural groups brings in the pedagogical issue of race versus ethnicity.
This study circumvented the issue by not considering it at all. As stated eariier,
the issue is one of perspective.

Practical impligtions for the fielg. The last conceptual implication relates
to a practical one. This study illustrates that Puerto Ricans do use parks. Some
more heavily than others. To continue to view ethnic group members as being

”under-represented” in a recreation site is to assume a normative order based

143

on Whites. From a demographic standpoint, when the minority populations
become the majority, this will no longer be an adequate point of reference.

The model shows that asking minorities about how they perceive the
resource is important to understanding why they visit a recreation site. This is a
better approach than observing that there are not as many Blacks as Whites
visiting the park. If one understands the perceptions or motivations behind the
visit, one could better cater the product to the user. This is implicit in most
studies, but it should be made more explicit.

The understanding that acculturation is a significant driving force has
several ramifications for practitioners. It indicates that the sense of belonging
one has to American society impacts public recreation. If one does not feel
accepted in the United States, one is not likely to participate in public places,
regardless of subcultural identity. One can have a strong subcultural identity,
but if one feels helshe belongs, one will still participate in the public sphere.
The implication of this is that parks should be promoted as cultural centers, and
recreation programs need to be more inclusive. It is worthwhile to emphasize
the importance of understanding the different leisure patterns and needs of
various ethnic groups.

If acculturation does indeed drive several of the other factors influencing
public participation in recreation areas, practitioners need to take the time to
know more about the background of subcultural groups using the resource. In
Southbridge, for example, a majority of the respondents preferred the survey in

Spanish and/or preferred to speak Spanish. This knowledge should be

144

incorporated into the management of the resource. For example, if there are
rules which regulate use of the parks, and the majority of ones users are non-
English speakers, then managers need to ensure that the signage is bilingual.
Cultural nuances are important to keep in mind when trying to enforce policies.

Because acculturation affects ones perception of discrimination,
managers can do one of two things. Managers and employees can be more
sensitive to this issue. By acknowledging that there is already sufficient
discrimination outside of the park, recreation managers and staff should work
harder to provide a welcoming atmosphere, with no tolerance for discriminatory
practices. Related to this, managers should hire a diverse group of employees
who are members of various ethnic groups and who can ”speak their language.”

Policies today need to reflect the users of tomonow. A more proactive,
assertive stance on reaching out to minority groups is recommended. In the
case of the Latino population, studies show that they are very brand loyal
(Wagner 8. Soberon-Ferrer, 1990). Starting a loyal following, based on cultural
understanding today, will ensure visitation in the more pluralistic environment of
the future.

For example, Henry Street Park was visited more than any other park in
this study. Administrators of the park realized that the majority of the users were
Puerto Rican, and that the majority spoke Spanish. Nearly a decade ago, tennis
courts were changed to reflect the pattern of use of the Puerto Ricans who
preferred to play basketball instead of tennis, and signage is bilingual. These

small but significant changes did not go unnoticed by the Puerto Rican

145

population. Over the years, Henry Street Park has hosted several Puerto Rican
festivals. As a result, perceptions of discrimination were lessened and the
perception of the park changed, thereby leading to more use.

As a recreation professional in a public service agency, exploration of
reasons for why people do not visit a recreation area are sought. Often times,
because the area is public, and the economic impact is minimal, minorities in the
public sector constituency are overlooked. A needs assessment should go
beyond recreation needs, and incorporate the cultural needs of the community.
Recommendations for Further Study

Procedural recommendations. If the study were to be conducted again,
more time should be allotted to the distribution and collection of the survey.
Related to this, third class mailing should be avoided if possible. Most of the
problems with responses and follow-ups resulted from problems with third class
mailing. Specifically, there were problems with the time that it took for the
Southbridge Post Office to distribute third class mail to the population because it
was not a priority. Less of a recommendation, and more of a reminder, the
researcher stresses the importance of making contact with several leaders in the
community. Time has to be given for the contacts to establish credibility for the
researcher and the study.

Questipnnaire recommendations. In order to better reflect the
Hollingshead Index, response categories should have an option for the spouse

so that total household income could be include. In addition, the employment

146

status of all working-age household members, and their occupations, should be
asked.

More items are needed to capture the essence of each of the different
construct areas. Given more space on the survey, the researcher would have
included more items to measure the constructs. In general, the more items one
has the better the reliability of the scales. Specifically, more items measuring
subcultural identity and socioeconomic status would have strengthened those
scales.

In the demographic section of the study inclusion of both sets of
grandparents should have been included (see Appendix A). If one looks at
section 7B1-7B5, it is clear that the second set of grandparents was not asked
about. This was an oversight on the part of the researcher. In order to use the
information, it would have to be assumed that when the respondent answered,
they were referring to both sets of grandparents. This is not an unreasonable
assumption given that most of the respondents are emigrants or of the first
generation in Southbridge.

Statistical reggmmendations. Generally, there has been little use of
confirmatory factor analysis and path analysis techniques in the recreation
literature. As the technique has demonstrated, causal models are very
unforgiving of error. The validity check is very rigid, and the reliability of ones
measures is incorporated into the measurement model. The process also allows

one to control for statistical artifacts such as measurement error.

147

As a result, it causes the researcher to be more precise in his or her
theorizing and in the tests of those theories. Diagnosing errors is a skill that is
sharpened with experience. The statistical techniques used in this study were
very useful diagnostic tools in assessing where the model went wrong, or where
the researchers thinking was incorrect, and where the model could be
strengthened. These statistical techniques are exceptional for theory
development and testing.

@ngptual recommendations. As mentioned earlier, the model is not
exhaustive, but there are certain variables from the study which could be
explored in future studies. Age, for example, might moderate the effect of
acculturation or subcultural identity on one’s perception of discrimination. Age
might also have a bigger impact.

Location was another common factor which was mentioned in the
literature but not assessed in the current model. It is recommended that a
spatial analysis be used as a means to determining if place of residence
(determined by physical location) has an influence on (specific) park use. It
would be useful to compare whether location or subcultural identity is a better
predictor of which park one would use.

Generational influence was not used in this study for reasons noted in the
questionnaire recommendations. However, in a future study, generational status
could be looked at to see what effect it has on acculturation. Conceptually, it is
conceived as a precursor to acculturation so that level of acculturation is

affected by generational status. It could be hypothesized that as generational

148

status increases (O=emigrant, 1=first generation, etc), the level of acculturation
(into the mainstream) increases.

Lastly, the study needs to be replicated. Specifically, the study should be
replicated among other ethnic groups. Use the same constructs, but alter the
substantive nature of the itemsfrndicators to reflect the particular ethnic group
under study. The concepts would still have to be subjected to a confirmatory
factor analysis. The purpose would be to see if the model holds true for other
ethnic groups.

Summagy

This study makes several contributions to the existing knowledge
regarding ethnicity and recreation behavior. First, it provides support for the
ethnicity, marginality, discrimination, and acculturation theories. Second, it
provides support for the inclusion of leisure variables when assessing recreation
behavior - these are variables often overlooked. Third, this study found that the
acculturation process is a significant factor. It is hypothesized to be the
antecedent of all the other factors. Fourth, this study uses statistical techniques
not utilized in prior research on this topic. Fifth, the sample collection technique
could be used as a model for other difficult to reach ethnic groups. Finally, this
study provides a model for the reconceptualization of the relationship between
ethnicity, marginality, and related factors, and their effect on public recreation

participation.

149

APPENDIX A

English and Spanish Questionnaires with Descriptive Statistics

150

Southbridge

   
    

The Department 'ecreation,
and Touri;" rces

Michigan sity
East La “~61; ‘ t u A 2

151

INSTRUCTIONS

The survey is meant to capture the opinions of the Puerto Rican community in Southbridge,
Massachusetts. The purpose of this study is to better understand how Puerto Ricans interact in public
recreation settings. The person whose name appears on the envelope should be the 'one responding to the
questionnaire. It is important that you answer the following questions as honestly as possible so that
the study can accurately represent the Puerto Rican community. Please take your time answering the
questions. The questionnaire should take from 10-12 minutes to complete. When you are finished, fold
the questionnaire once and place it in the yellow envelope provided for you. You do not need to put a
stamp on the envelope, simply drop it in the mail.

If you received two questionnaires (one in English, the other in Spanish),
please return only one completed questionnaire.

Thank you!

Any questions or concerns please contact:

Edwin Gomez, Doctoral Candidate
Department of Park, Recreation and Tourism Resources
131 Natural Resources Building
Michigan State University
East Lansing, MI 48824-1222
Phone: (517) 353-9145

E-mail: gomczedw@pilotmsu.edu

152

Page 1.

—
SECTION ONE: PARK QUESTIONS

 

1. People have suggested many reasons why they visit parks. Below is a list of the benefits of going
to a park. Please indicate whether you strongly agree, agree, neither agree nor disagree,
disagree, or strongly disagree.

 

Neither
rec
Strongly or Strongly
Agree Agree Disagree Disagree Disagree
"/ °/ ‘V °/ "/

A. A benefit oifioing to Southbridge parks ( o) ( o) ( o) ( o) ( o)

is that they ow me to enjoy nature. 38.2 39.5 14.4 6.0 1.8
B. A benefit of oing to Southbri c parks

is that they low me to escape or a while. 39.0 43.5 10.7 4.2 2.6
C. A benefit of goin to Southbridge parks is they

allow me to socifiizdcreate personal contacts. 29.1 48.9 12.8 6.4 2.7
D. A benefit of going to Southbridge parks is that

they allow me to get some exercise. 37.4 42.9 11.8 6.6 1.3
E. A benefit of going to Southbrid e arks is that

they allow me to spend time wi amily/friends. 50.4 37.6 6.1 4.5 1.1
F. A benefit of having parks in Southbridge is that

they offer a place with lots of open space. 38.1 43.6 10.0 6.6 1.8
G. A benefit of having parks in Southbridge is that

they ofi'er a place for children and youth to go. 61.5 29.8 3.4 3.7 1.6
2. How 0 en during each season do you use or visit each of the following public parks in

So ridge?

A. MORRIS STREET PARK
(Almost Never) (Sometimes) (Frequent) (Very Frequent) (A Lot)
Never Once 2-4 times 5-13 times 14-20 times 21-29 times 30+ times

(%) (%) (%) (%) (%) (%) (%)

Summer (June-August) 43.8 1 1.7 18.2 17.7 4.2 1.8 2.6

Fall (September-November) 70.3 7.6 14.8 4.4 2.3 0.5 0.0

Winter (December-February) 85.9 4.4 7.0 1.8 0.8 0.0 0.0

Spring (March-May) 63.0 6.8 14.3 8.9 2.9 1.9 0.5
B. WEST STREET PARK

(Almost Never) (Sometimes) (Frequent) (Very Frequent) (A Lot)
Never Once 2-4 times 5-13 times 14-20 times 21-29 times 30+ times

(%) (%) (%) (%) (%) (%) (%)

Summer (June-August) 42.3 12.3 21.9 16.4 3.9 0.8 2.3

Fall (September-November) 71.4 7.6 10.7 7.3 2.1 0.5 0.5

Winter (December-February) 89.6 2.6 5.7 0.0 0.5 0.0 0.0

Spring (March-May) 64.1 6.8 15.4 9.1 2.1 2.1 0.5

153

Page 2.

C. HENRY STREET PARK
(Almost Never) (Sometimes) (Frequent) (Very Frequent) (A Lot)
Never Once 2-4 times 5-13 times 14-20 times 21-29 times 30+ times
(%) (%) (%) (%) (%) (%) (%)
Summer (June-August) 19.3 8.6 19.3 16.7 12.0 7.0 17.2
Fall (September-November) 50.0 8.1 19.5 12.0 4.7 2.3 3.4
Winter (December-February) 79.1 5.2 9.1 3.9 1.6 0.5 0.5
Spring (March-May) 41.9 8.3 18.5 11.2 10.2 3.6 6.3
D. WESTVILLE DAM RECREATION AREA
(Airnost Never) (Sometimes) (Frequent) (Very Frequent) (A Lot)
Never Once 2-4 times 5—13 times 14-20 times 21-29 times 30+ times
(%) (%) (%) (%) (%) (%) (%)
Summer (June-August) 8.1 9.9 24.7 29.4 14.8 4.9 8.1
Fall (September-November) 51.3 12.2 22.9 7.3 3.4 1.8 1.0
Winter (December-February) 82.3 8.1 6.5 1.8 1.0 0.3 0.0
Spring (March-May) 40.1 9.1 23.7 15.6 7.3 2.3 1 8
E. HARRY J. McMAHON MEMORIAL FIELD (Dresser St.)

(Almost Never) (Sometimes) (Frequent) (Very Frequent) (A Lot)
Never Once 24 times 5-13 times 14-20 times 21-29 times 30+ times

(%) (%) (%) (%) (%) (%) (%)
Summer (June-August) 46.1 10.4 14.3 12.8 8.6 2.1 5.7
Fall (September-November) 62.8 8.9 10.9 7.0 4.9 2.1 3.4
Winter (December-February) 89.1 4.4 2.9 0.8 1.6 0.8 0.5
Spring (March-May) 62.2 6.5 11.7 8.3 6.0 2.1 3.1

 

SECTION TWO: PUERTO RICAN

 

CULTURE QUESTIONS
3. Please indicate which language(s) you prefer to use.
More More
S anish Use En sh
Only an Both an Onl
Spanish English Equally Spanish Eng sh
(%) (%) (%) (%) (%)
A. Preferred language at home 3.9 14.1 29.2 34.4 18.5
B. Preferred language for radio and television. 12.1 21.8 34.9 19.4 11.8
C. Preferred language for magazines/newspapers. 22.2 16.2 25.3 16.4 19.8
D. Preferred language for music. 5.3 10.8 35.1 25.3 23.5
E. Preferred language your close friends speak. 9.2 13.1 30.2 28.1 19.4
F. Preferred language at parties and get-to others. 6 1 12.7 34.2 28.6 18.3
G. Preferred language when you visit poop e or 4 8 12.1 34.4 31.2 17.5
when people visit you.

4. Please indicate how important you feel these are.

Veg Somewhat A Little

lmpo nt important Important Important Unlm rtant
(%) (%) (%) (%) °o)

A. How important is it to celebrate

holidays in the Puerto Rican way? 1.6 3.7 8.4 22.5 64.0
B. How important is it to raise your

children with Puerto Rican values? 1.6 2.4 6.5 20.2 69.4
C. How important is itto maintain

the use of the Spanish language? 1.3 2.3 6.0 15.1 75.3
D. How important is it to maintainPuerto

Rican values as a part of your hfe? 1.3 1.3 5.5 15.9 76.0

Page 3.

Some people have reported that other people do not treat them fairly in Southbridge simply because
they are Puerto Rican. Others have not experienced this. How do you feel about the following
statements? Please indicate the extent to which you strongly agree, agree, neither agree nor
disagree, disagree, or strongly disagree with the following statements.

Neither

ee
Strongly Agr" Strongly
Agree Agree Disagree Disagree Disagree
(%) (%) (%) (%) (%)
. 1n Southbridge, i feel that Puerto Ricans
are not very welcomed. 12.6 22.5 34.8 22.8 7.3
. ln Southbridge, I feel that police often
hassle Puerto Ricans. 22.3 24.9 29.6 18.3 5.0
In Southbridge, the presence of other groups that
are not Puerto Rican makes me feel uncomfortable. 4.5 8.9 20.2 41.2 25.2
. In Southbridge, I feel Americans excluded me
from activities because 1 am Puerto Rican. 6.8 13.9 28.7 34.2 16.3

In Southbridge parks, i feel Americans look down
on me because I participate in customs ofmy culture. 11.1 14.7 27.1 35.0 12.1

Please indicate the extent to which you strongly agree, agree, neither agree nor disagree, disagree,
or strongly disagree with the following statements regarding your views on your cultural identity.

Neither
Strongly Nrr Strongly
Agree Agree Disagree Disagree Disagree
(%) (%) (%) (%) (%)
. 1 strongly identify myself as a Puerto Rican first,
rather than as an American. 66.4 18.5 10.9 3.9 0.3
Puerto Ricans in America should try harder to be
American rather than practicing activities that connect
them to their cultural heritage. 24.3 38.4 17.8 13.1 6.5
1 would prefer to be with Puerto Ricans,
rather than with Americans. 13.0 9.6 45.3 5.0 1.3
. Puerto Rican culture is very distinct and
very different from the American culture. 38.6 40.3 14.7 5.0 1.3
1 identify with other Puerto Ricans, even if
they are not close friends or relatives. 31.8 49.3 12.9 3.9 2.1
1 would prefer to display the Puerto Rican
flag, rather than the American flag. 22.3 18.6 39.6 14.4 5.0

155

Page 4.

SECTION THREE: DEMOGRAPHICS
_

7. These final questions are about you. Please answer to the best of your knowledge.

A1. How long have you lived in United States? Years (M = 19.53, gal = 10.91, min/max. = 0 yr. / 50 yrs)
A2. Have you ever lived in Puerto Rico? No 21.2% Yes 79.8% Years (M = 18.54, ed = 11.14,
min/max. = lyr. / 59 yrs.)

CI Puerto Rico D United States D Other

Bl. Where were you born? 72.9 % 27.1 % 0.0 %
82. Where was your father born? 99.7 0.3 0.0
B3. Where was your mother born? 99.2 0.3 0.5
B4. Where was your grandfather born? 98.4 0.8 0.8
BS. Where was your grandmother born? 98.7 0.5 0.8

C. What do you consider your cultural background to be?

8.2 % 0 Puerto Rican

5.3 C] More Puerto Rican than American
19.5 CI Puerto Rican and American equally
11.6 DMoreAmericanthanPuertoRican
55.4 0 American

D. What is your sex? 39.5 % Male 60.5 % Female
B. What is your age? (M = 34.27, g] = 13.96, min/max. = 14 yrs. / 80 yrs.)
F. What is your marital status?
41.0 % Ci Single 45.2 % D Married 5.3 % Cl Separated
5.3 Ci Divorced 1.9 D Widowed 1.3 CI Other
G. What is your work status?
64.6 % Cl Employed 2.9 % D Self-employed 10.9 % D home maker
5.1 CI Retired 6.9 D Unemployed 9.6 D Other

H. What is your occupation?

 

I. What was your total family income 1998 (before taxes)? [Include parents and children’s income]

8.1 % 0 Under $5,000 18.4 % Cl $15,000 to $24,999 9.7 % Cl $50,000 to $75,000
6.8 CI $5,000 to $9,999 17.2 CI $25,000 to $34,999 2.9 CI $75,000 to $99,999
14.9 0 $10,000 to $14, 999 21.4 C1 $35,000 to $49,999 0.6 CI $100,000 and over

J. What is the highest grade that you completed in school?
16.8 % Cl Eighth Grade or Below 16.3 % D Some College 1.9% CI Some Graduate School

26.9 0 Some High School 6.4 D Associate’s Degree D Master’s Degree
27.7 C] High School Graduate/GED 4.0 CI Bachelor’s Degree D Doctoral Degree

156

Page 5.

Thank you very much for taking the time to answer this questionnaire and
for sharing your opinions with me. You have been a great help!

Please return this survey to me in the enclosed self-addressed envelope.

157

 

    
     
 

d6 .
Pu efios

The Departm k, Recreation,
and Tozpe‘ .1 ources

‘ I I‘ 'ty
, g, MI 48824-1222

158

INSTRUCCIONES

El questionnario propone capturar las Opiniones de la comunidad puertorriquefia dc Southbridge,
Massachusetts. Este estudio intents obtener su punto de vista acerca de como los puertorriquefios
interaccionan en sitios dc recreacidn pr'rblica. La persona cuyo nombre aparece en cl sobre debe ser
e1/la que responds a1 cuestionario. Es importante que usted contests las siguientes preguntas lo mhs
honestamente posiblc para que el estudio pueda representar correctamente la comunidad
puertorriquefia. Por favor de coger su tiempo contestando las preguntas. El cuestionario debe cogerte
desde 10 a 12 minutos para completar. Cuando usted termine, dobla el cuestionario una vez y ponlo
dentro del sobre amarillo provisto para usted. No necesita ponerle un sello, simplemente envielo por
correo. .

Si usted recibio dos cuestionarios (uno en ingles y el otro en espafiol),
por favor devuelva uno solamente.

[Graciasl

Cualquiera pregunta 0 problems comunicate con:

Edwin Gomez, Doctoral Candidate
Department of Park, Recreation and Tourism Resources
131 Natural Resources Building
Michigan State University
East Lansing, MI 48824-1222
Phone: (517) 353-9145

E-mail: gomezedw@pilot.msu.edu

159

Pagina 1.

PRIMERA SECCION: PREGUNTAS DE LOS PARQUES

1. Se ha sugerido varios beneficios de tener parques en Southbridge. A delante se presenta una lista de oraciones.
Por favor indique si usted estd muy de acuerdo, de acuerdo, ni de acuerdo ni en desacuerdo, en desacuerdo, a
may en desacuerdo.

Ni De
Acuerdo
Muy De De NI En Des- En Muy En
_ . Acuerdo Acuerdo acuerdo Desacuerdo Desacuerdo
A. Un beneficro dc tener parques en Southbridge es

que permiten disfrutar de la naturaleea. D D D C1 D

B. Ul'l beneficio dc tener parques en Southbridge es
que ofrecen actividades recreativas. D C1 C1 C1 Cl

C. Un bencficio dc tener parques en Southbridge es
que chosen an sitio donde se puede socializar
o establecer contactos personales. Cl D C1 C1 C1

D. Un beneficio dc tener parques en Southbridge es
que ofrecen un sitio donde hacer ejercisio. E] El D D D

E. Un beneficio dc tener parques en Southbridge es
que ofrecen un sitio donde se puede compartir

con familiares y amigos. C] D D . D C]
F. Un beneficio dc tener parques en Southbridge es

que ofrecen un sitio con espacios abiertos. D C1 Cl C] D
G. Un beneficio dc tener parques en Southbridge es

que ofrecen un sitio donde pueden ir niflos/jovenes. D 0 C1 C1 C1
2. En 1998, gCon cudntafi'ecuencia usted utilize las siguientes parques en Southbridge? lndr'que cudntas

veces usted va a los parques por cada estacidn del ado en cada parque.

A. MORRIS STREET PARK
(Cmi Nuncs) (A veces) (rm) mm mm) (Muy We)
Nunca Una vcz 2-4 veces 5-13 veces 14-20 veces 21-29 veces 30+ veces
Verano (Junio-Agosto) D D C1 C1 C1 C1 C1
Otofio ( tiernbre-Noviembrc) D C1 C1 Cl C] D D
Inviemo 'embre-Febrero) C1 C1 C1 C1 C1 C1 C1
Prirnavera (Mano-Mayo) Cl Cl C1 D D C1 C1
B. WEST STREET PARK
(Cal Nines) (A veces) (Frecumte) (But-us Frequente) (Muy Preemie)
Nunca Una vcz 2-4 veces 5-13 veces 14-20 veces 21-29 veces 30+ veces
Verano (Jimio—Agosto) C1 C1 Cl C] D D Cl
Otofio (Septiembre-Noviernbre) C1 C1 C1 D E] El
Inviemo (Deciembre-Febrero) C1 C1 Cl C1 C1 C1 C1
Primavera (Mam-Mayo) C1 C] D D D C1 C1

160

Pagina 2.

C. HENRY STREET PARK
(Cui Nurse) (A veces) (Freer-cote) (But-rte Frequurtc) )(Muy W)
. Nunca Una vez 2-4 veces 5-13 veces 14-20 veces 21-29 veces 30+ veces
Verano (Junro-Agosto) C1 0 Cl C1 Cl C1 C1
Otofio ( tiernbre-Noviernbre) D D D C1 C1 C1 C1
lnvierno 'ernbre-Febrero) D C1 D D C1 C1 C1
Primavera (Mano-Mayo) D D C1 C1 C1 Cl C]
D. WESTVILLE DAM RECREATION AREA
(c-i Nncs) (A veca) (Freer-eats) (one We) (Muy W)
Nunca Una vez 2-4 veces 5-13 veces 14-20 veces 21-29 veces 30+ veces
Verano (Junio-Agosto) D D C1 C1 Cl Cl Cl
Otofio (Septiembre-Noviernbrc) D C1 C1 C1 Cl D C]
lnvierno (Deciembre-Febrero) D D C1 Cl C1 C1 0
Prirnavera (Mano-Mayo) C] D D C1 C1 C1 C1
E. HARRY J. McMAHON MEMORIAL FIELD (Dresser St.
W) M) (BI-flue M) (May mic)
Nunca Una vez 24w veces 5-13 veces 14-20 veces 21-29 veces 30+ veces
Verano (Junio-Agosto) D D D D 0 C1 C1
Otoiio ( ticrnbre-Noviembrc) Cl D D C1 Cl D D
lnvierno ernbre-Febrero) C] D D C1 C1 C1 C1
Prirnavera (Marzo- Mayo) C] Cl C1 C1 C1 C1 C1

 

SEGUNDA SECCION: PREGUNATAS SOBRE LA CULTURA PUERTORRIQUENA

 

3. Por favor indique curil(es) idioma(s) usted prefiere usar.

M‘s Mhs

alol Ambos In es
Solo e For us Solo
Espaiiol Ingles lgual Espaiiol Ingles

A. Idioms més utilizado en la case. 0 C1 D D D
B. Idioms preferido por radio y/o television. C1 C1 C1 C1 Cl
C. Idioms preferido en revistas y/o periodicos. D C1 C1 C1 C1
D. Idioms preferido en la musica. D D C1 C1 D
E. LQué idioms hablan sus ami os mes cercanos? C1 C1 C] D C]
F. LQue idioms se habla en las estas y reimiones? D D D Cl C]
G. gQué idioms hablan las personas que usted

visits o que lo visitan a usted? El Cl E] D D
4. Por favor indique la importancia de los siguientes.

M h Bastan an
lmpgs't:ncia Importancla ImAdfiirltlhhcia Importancia lm pagoda

A. fights importancia tiene ara usted cl cele-

los dias feriados a] lo puertorriquefio? C1 C1 C1 Cl C]
3' tfiiifi’Wé gfi‘étifirie‘ifiifi“ D u a D u
C' ‘fihi‘mm cit-.1 iieTguEJTespaaor? Cl 0 C1 C1 D
D imamztssrsmssm o a a o a

161

Pagina 3.

Algunas personas han informado que otras personas los(as) han tratado injustamente en las sitios
priblico de Southbridge simplemente porque son puertorriquerios. 0tras personas no han sufrido esta
experiencia. Quisiera saber su experiencia al respecto. Por favor indique en la medida en que usted
estd: Muy de acuerdo, de acuerdo, ni de acuerdo ni en desacuerdo, en desacuerdo, o muy en
desacuerdo con la siguiente.

Ni De
Acuerdo
Muy De De Ni En Des- En Muy En
_ . Acuerdo Acuerdo acuerdo Desacuerdo Desacuerdo
En Southbridge, los puertornquefios

no son bienvenidos. C1 C1 Cl C] D

En Southbridge Ia policia molesta a los
puertorriquellos. E] Cl C1 D 0

En Southbridge, la presencia dc otros grupos que
no son puertorriqueiios me haw sentir incomodo. C1 C1 C1 C1 C1

En Southbridge los americanos me cxcluyen de
sus actividades porque soy puertorriqueflo. C1 C1 C1 C1 C1

En Southbridge los americanos me rniran por
encima del hombro (con menosprecio) debido
a que yo participo en costumbres puertorriqueflas. Cl C1 C1 Cl C]

Algunas personas sienten que las puertorriquer'ios deben adoptar valores americanos, otras sienten que
puertorriquerios deben mantener sus costumbres y tradiciones. Por favor indique lo que usted opina de las
siguientes oraciones.

Ni De

Acuerdo
Muy De De Ni En Des- En Muy En
Acuerdo Acuerdo acuerdo Desacuerdo Desacuerdo

Me identifico primero como puertorriquefio
en vez dc americano. D 0 Cl 0 C]

Los puertorriqueflos en los Estados Unidos deben

deesforzarsemdsporadaptarscalacultura

americana en vez dc practicar actividades que los

identifiquen con su cultural. D D D C] D

Yo prefiero asociarme con pucrtorriquefios,
en vez dc americanos. C1 D D C1 C1

La cultura puertorriquefia se distingue y es
difercnte a la culture americana. D C1 C1 C1 C1

Yo me identifico con otros puertorriquefios
aunque no seen amigos o familiares. C1 Cl 0 Cl Cl

Yo prefer-la mostrar la bandera de Puerto Rico
en vez de la bandera de los Estados Unidos. Cl . Cl C1 Cl [:1

162

Pagina 4.

TERCERA SECCION: DEMOGRAFICAS

7. Estas rillimas preguntas son sobre usted. Por favor conteste segrin su mejor conocimiento.

A1. gPor cuantos afios ha vivido usted en los Estados Unidos? Afios

A2. gHas vivido en Puerto Rico? D No Cl Si -0 Afios

Bl. gDonde nacio usted? Cl Puerto Rico 0 los Estados Unidos CI Otro
BZ. gDonde nacio su padre? Ci Puerto Rico CI los Estados Unidos Ci Otro
B3. gDonde nacio su madre? Ci Puerto Rico Cl 108 Estados Unidos D Otro
B4. LDondc nacio su abuelo? C1 Puerto Rico CI 108 Estados Unidos D Otro
BS. gDondc nacio su abucla? Cl Puerto Rico CI los Estados Unidos C1 Otro

C. (A cuai culture usted pertenece?

D puertorriquefio/a

C1 més puertorriquefio/a que americano/a
Ci arnbos por igual

Cl mes arnericano/a que puertorriquefio/a
Cl americano/a

D. (,Cual es su sexo? Ci Hombre Ci Mujer

B. Low es en edad? Aflos

F. LCual es su cstado civil ?
Cl Soltero/a Ci Casado/a CI Separado/a
Cl Divorciado/a D Viudo/a D Otro

G. 5Cua1 es su cstado ocupacional? .
C1 Empleado/a Ci Empleado/a de 81 rrnsmo C1 Ama de casa

Cl Retirado/a D Desempleado/a El Otro

 

H. (,Cual es su ocupacibn?

 

1. gCuénto se gano su farnilia en 1998 (antes dc deducir los impuestos)? [Incluye las gananzas de padres e hijos.]

CI Menos de $5,000 0 $15,000 to $24,999 0 $50,000 to $75,000
CI $5,000 to $9,999 CI $25,000 to $34,999 Ci $75,000 to $99,999
D$10,000to$14,999 D$35,000to$49,999 Cl$100,000y tabs

1. LCuél es el grado mas alto que usted ha completado?

Cl Octavo grado o menos Cl Algunos creditos a nivel universitarios Cl Algunos créditos a nivel graduado
CI Algunos creditos dc Escuela Superior 0 Grado Associado Cl Maestrfa
D Diploma dc Escuela Superior CI Bachillerato Cl Doctorado

163

Pagina 5.

Gracias por haber tornado de su tiempo para completar este cuestionario y
por compartir sus opiniones conmigo. [Usted ha sido de gran ayuda!

Por favor devuelva e1 cuestionario en el sobre predirigido.

164

APPENDIX B

Map of Parks in Southbridge and Samples of Correspondence

165

/

 

MAP OF SOUTHBRIDGE, MA AND ITS PUBLIC PARKS

er

"' r'rr

r
{:2
Downtown So’uthbrrdge
/ WavifleDanReaeatianAm(A)w
”H. ' Warflredei)

   
   

/'

 

 

 

 

 

166

SAMPLE LETTER FOR COMMUNITY LEADERS
December 8, 1998

Father Peter Joyce

St. Mary’s Catholic Church
263 Hamilton Street
Southbridge, MA 01550

Hello! I am writing you concerning a research project i am conducting involving
the Puerto Rican community in Southbridge and their use of Southbridge parks.
The study looks at possible things which might influence why Puerto Ricans (or
other ethnic group members) in the United States may or may not use public
recreation spaces. This research is required for the completion of my doctorate
in the Department of Park, Recreation and Tourism Resources and Urban
Studies at Michigan State University.

I am asking for your help in soliciting names for my study. The phone book was
going to be used as a list for mailing, but I realized that many Puerto Ricans are
unlisted. I am approaching different community leaders to obtain names of
people who might be interested in participating in the study.

The participants will be required to give their name, mailing address, and
whether they want the survey in Spanish or English. I would like to obtain
names of both men and women. Please use the enclosed forms to collect the
names. Their names are for mailing purposes only and will be taken off the list
once their questionnaire is returned. This information WILL NOT be given to
anyone. It will be kept extremely confidential. Participation is voluntary. I am
asking that you please announce this to the Puerto Rican congregation at St.
Mary’s.

I will be in touch to follow up and offer any clarification that is needed. If you
have any questions prior to my follow up, please don’t hesitate to call me at the
number printed in the left margin. Thank you for any help you may have given
me.

Sincerely,

Edwin GOmez
Doctoral Candidate
Michigan State University

167

SAMPLE OF SIGN-UP SHEET FOR SURVEY PARTICIPATION

Por favor de escn’bir su nombre y direceidn si quieres participar en la
investigacidn de puertom'queflos y el uso de las parques en Southbridge.
Please write your name and address if you would like to participate in the study
of Puerto Ricans and their use of parks in Southbridge. Gracias. Thank you.

(1) Nombre:

 

Direccién:

 

o rid MA 1
Lenguaje preferido para el
questionario:

 

(2) Nombre:

 

Direccién:

 

Southbridge, MA 01550
Lenguaje preferido para el

 

 

 

 

 

 

 

 

questionario:
(3) Nombre:
Direccibn:

Mam
Lenguaje preferido para el
questionario:
(4) Nombre:
Direccibn:

S hbrid MA 1 50
Lenguaje preferido para el
questionario:
(5) Nombre:
Direccibn:

 

Southbridge, MA 01550
Lenguaje preferido para el
questionario:

 

(6) Nombre:

 

Direccién:

 

Spighprigge, MA 01550
Lenguaje preferido para el
questionario:

 

(7) Nombre:

 

Direccibn:

 

Segthbrigge, MA 01550

Lenguaje preferido para el
questionario:

 

168

(8) Nombre:
Direccibn:

 

 

Southbridge, MA 01550
Lenguaje preferido para el
questionario:

 

(9) Nombre:
Direccion:

 

 

Southbridge, MA 01550
Lenguaje preferido para el

questionario:

 

(10) Nombre:
Direccibn:

 

 

o hbrid
Lenguaje preferido para el
questionario:

MA 01550

 

(1 1) Nombre:
Direccibn:

 

 

Somhbridge, MA 01550

Lenguaje preferido para el
questionario:

 

(12) Nombre:
Direceion:

 

 

Southbridge, MA 01550
Lenguaje preferido para el

questionario:

 

(13) Nombre:
Direccion:

 

 

Southbridge, MA 01550
Lenguaje preferido para el

questionario:

 

(14) Nombre:
Direccion:

 

 

Southbridge, MA 01550

Lenguaje preferido para el
questionario:

 

SAMPLE OF SPEECH FOR SPANISH-SPEAKING PARISHIONERS AT ST.
MARY’S CHURCH DURING CHRISTMAS EVE MASS

Bueno, quiero darle las gracias a Padre Peter por dejanne hablar aqui hoy. Y
primero que nada, quiero desearle a todos una Feliz Navidad. Soy hijo de
Chiqui y Julie Gomez. Estoy haciendo el doctorado en la Universidad Estatal de
Michigan. Yo estoy conduciendo un estudio de los puertorriquenos en cuanto al
uso de Ios parques de Southbridge y la influencia de la cultura puertorriquena.
Este estudio es un requisito para la tenninacibn de mi doctorado en Parque,
Recreacibn y Turismo.

Yo sé que aqui hay otros Iatinos que no son puertorriquefios, y la razbn que
estoy haciendo el estudio de puertorriquefios solamente es porque soy
puertorriquei'io y me exigieron que no haga el estudio con otros grupos Iatinos
porque yo no conozco la cultura de ellos.

Bueno, (,porque estoy aqui hoy?

Pues, porque necesito Ia ayuda y cooperacibn de todos Ios puertorriquel'los aqui
presente. Yo necesito sus nombres porque muchos de ustedes no estan en la
guia telefonica 0 en el censo. Voy a colocanne atras, cerca de la puerta de la
entrada principal. Por favor paren, escriban su nombre y direccién en Ios
papeles que estan en la mesa, o hablen conmigo si tienen algunas preguntas.

Solamente puede dar el nombre si tienes mas de 17 afios. También, les quiero
decir que el cuestionario lo van a recibir a fines de enero 0 al principio de
febrero. Ya varias personas me han dado su informacibn, si algunos de
udstedes estan aqui presente. pues no me tienen que dar Ia informacibn otra
vez.

Finalmente, les quiero decir que el cuestionario les va a coger de 10 a 12
minutos. Yo les voy a enviar el sobre con mi direccién y franque - haci que no
tienen que pagar por mandarlo.

Gracias por su tiempo y cooperacibn y [Feliz Navidadl

169

SAMPLE LETI' ER FOR ADMINISTRATION AT THE HIGH SCHOOL

February 9, 1999

Mr. Walter Gosk
Southbridge High School
25 Cole Avenue
Southbridge, MA 01550

Dear Mr. Gosk:

Greetingsl I would like to take this opportunity to thank you once again for
agreeing to help me distribute the questionnaire to the Puerto Rican students at
Southbridge High (SHS). In our previous conversation you had suggested that I
include a letter stating what the study is about. This letter is enclosed in the box
with the surveys.

In addition to the letter of introduction, 1 have included a brief autobiography of
specific events after I left SHS. l have provided you with 130 English surveys
and 20 Spanish surveys. This was the number of surveys recommended by Mr.
Bailey. The letters of introduction are also in English and Spanish. Please give
the Spanish versions to the bilingual teacher.

I understand that you may run into some difficulty in expediting the survey
because of winter break. If this is the case, conduct the surveys alter break and
get them back to me at the earliest convenient time. I have already spoken with
Graciela, and she knows she will be receiving the box of surveys from you or
your wife.

If there are any questions, please don’t hesitate to call me at 517-353-5190. If
you have access to e-mail, my address is gomezedw@pilot.msg.edu. Thanks
again!

Sincerely,

Edwin Gemez
Project Coordinator
Michigan State University

170

SAMPLE LETTER FOR ADMINISTRATION AT HEAD START PROGRAM

February 9, 1999

Mrs. Evelyn Velez
Head Start Program
115 Marcy Street
Southbridge, MA 01550

Dear Evelyn:

Greetings! I would like to take this opportunity to thank you once again for
agreeing to help me distribute the questionnaire to the Puerto Rican parents in
the Head Start Program. In our previous conversation you had suggested that I
include a letter stating what the study is about. This letter is enclosed in the
envelope with the surveys.

In addition to the letter of introduction, l have included a brief autobiography of
specific events after I left Southbridge High School. I have provided you with the
surveys and the list of people you said you would administer the questionnaire
to. The letters of introduction are also in English and Spanish.

The surveys are being sent out next week. I hope that you will be able
administer all the surveys before the end of February. If you cannot, then have
them back for me as early as possible in March. Please contact my mother
(Julie) so that she can pay to have the surveys sent to me. This way it does not
cost your program any additional cost. I have already spoken with my mother,
and she knows she will be receiving the surveys from you. Her number is XXX-
XXXX..

If there are any questions, please don’t hesitate to call me at 517-353-5190. if
you have access to e-mail, my address is gomezedw@pilot.msu.edg. Thanks
again!

Sincerely,

Edwin Gbmez
Project Coordinator
Michigan State University

Enc (2)

171

SAMPLE PRE-SURVEY LETTER: ENGLISH

(OTRO LADO PARA ESPANOL)

February 5, 1999

«First» «Last»
«Nums «Streets
Southbridge, MA 01550

Dear «First»:

Hello! I am writing you in reference to a study I am conducting that involves
Southbridge Puerto Ricans and their use of Southbridge public parks. The study takes
into consideration different things which may influence the use of public areas by
Puerto Ricans in the United States. This research is the final requirement for my
doctorate in the department of Park, Recreation and Tourism Resources and Urban
Studies at Michigan State University.

I am asking for your help. I have contacted different leaders in the Puerto Rican
community in order to obtain names of people who might be interested in participating
in my study. You (or someone who knows you) gave me your name so that you may
participate in my study.

Information provided by participants WILL NOT BE GIVEN to anyone. The information
will be kept extremely confidential. Your participation is completely voluntary, and it will
be greatly appreciated.

Within a few days, you will be receiving the questionnaire. Until then, I have provided
an autobiography. This introductory letter offers a brief history of my time away from
Southbridge, far away from my Puerto Rican community. I hope that you find it
interesting. Many thanks in advance!

Sincerely,

Edwin Gemez
Doctoral Candidate
Michigan State University

172

SAMPLE PRE-SURVEY LETTER: SPANISH

(OTHER SIDE FOR ENGLISH)

5 de febrero del 1999

«First» «Last»
«Num» «Street»
Southbridge, MA 01550

Querido «First»:

jHolaI Yo te estoy escribiendo en referencia a una investigacibn que estoy
conduciendo a cerca de los puertorriquefios en Southbridge y el uso de los parques
pr'rblicos de Southbridge. El estudio considera cuales son las cosas que infl en la
utilizacién de sitios publicos por los puertorriquel‘los en Ios estados unidos. sta
investigacibn es un requisito para la terminacibn de mi doctorado en el departamento
de Recursos de Parque, Recreacién y Turismo y Estudios Urbanos en la Universidad
Estatal de Michigan (Michigan State University).

Yo le estoy pidiendo su ayuda. He contactado a diferentes llderes de la comunidad
puertorriquefia para obtener nombres de personas que estarian. interesado en
participar en mi estudio. Usted (o alguien que Ie conoce) me did su nombre para que
participe en el estudio.

La informacibn de los participantes NO SE VA A DAR a nadie. lnfonnacibn se va a
mantener sumamente confidencial. Su participacibn es completamente voluntaria, y
sera bien apreciada.

Dentro de unos pocos dias, le va a llegar el cuestionario. Hasta que llegue el
cuestionario, Ie he provisto unos datos autobiograficos. Esta carta de presentacibn
ofrece una breve historia de mi tiempo fuera de Southbridge, lejos de mi comunidad
puertorriquei'la. Espero que la eneuentre interesante. Hasta luego y imuchas gracias
de antemanol

Sinceramente,

Edwin Gemez
Candidato Doctoral
Michigan State University

173

SAMPLE PRE-SURVEY AUTOBIOGRAPHY: ENGLISH
(OTRO LADO PARA ESPANOL)
Edwin Gbmez: From Southbridge, Massachusetts to East Lansing, Michigan

l was bom in Southbridge on March 24, 1968: the third son of Natividad (Chiqui) and Julia
(Julie) Gemez. Some of you may know my brothers Gerardo (Gery) and Julio (Cano) and
my sister Graciela (Chela). My parents are both from Puerto Rico. Therefore, Spanish was
always spoken at home. I have very fond memories of Southbridge. Perhaps my best
memory is growing up in a community of Puerto Ricans. Regardless of where you went
there was always a "bendicibn' given, an 'hola compai", or the pride one felt when hearing
the word: Boricual

Being the son of parents who emigrated from Puerto Rico I was raised in the old style. What
my parents most wanted for us was a good education. I am happy to say that Southbridge’s
school system prepared me very well for my college years. But, I get ahead of myself...

After graduating fiom Southbridge High, Iwent to Bany University in Miami, Florida. This
was my first time away from home, and I admit that I was a little scared. But, I did well at the
university. After my scholarships ran out, I did not have the money to continue my studies. I
went back to Southbridge and worked at the American Optical for a while.

I worked in the same area of the factory where my mother used to work. I finally got to see
and appreciate what my parents went through to provide for our family. I had a great time
there. I met people from Gmece, Poland, and Puerto Rico. The job was respectable, but I
still wanted to go to college. I decided to join the Anny.

l was in the active Army for three years! i started as a private and left as a sergeant. The
hours were long, but it was worth it The best thing about the Army was that I met people
from all over the United States, and I Ieamed about different cultures. I had a lot of prestige
and responsibility at the age of 22. The Army teaches you to stay focused on a goal. My
goal was to go to a university and get a college degree. In the Army I earned the College
Fund and GI Bill so I could go to college.

This money allowed me to get my baccalaureate degree in Geographyfl’ourism at Salem
State College. I went from there to Rochester, New York for my master’s in Tourism and
Hospitality Management at the Rochester Institute of Technology. I am currently at Michigan
State University working on my doctorate in Park, Recreation and Tourism Resources and
Urban Studies. The scariest part of all my colleges was that I didn’t know what to expect.
However, it was also the best part! I never thought I would get this far, but with love, faith
and perseverance, I will make itl

After this study is completed, I will finish my doctorate! My plans are to teach at a university.
I was very fortunate that I had excellent professors to mentor me and encourage me. I would
like to continue that tradition (perhaps with your children). I owe great thanks to God, my
family and friends, my Puerto Rican community, and Southbridge schoolteachers. You have
all helped me in many ways. Thank you!

174

SAMPLE PRE-SURVEY AUTOBIOGRAPHY: SPANISH

(OTHER SIDE FOR ENGLISH)
Edwin Gomez: Desde Southbridge, Massachusetts a East Lansing, Michigan

Nacl en Southbridge el 24 de Marzo del 1968: B! tercer hijo de Chiqui y Julie Gomez.
Algunos de ustedes conocen a mis herrnanos Gery, Cano, y Chela. Mis padres son de
Puerto Rico. Por eso, en casa siempre se hablaba espaflol. Estoy aficionado de mis
memorias de Southbridge. Ouizas mi mejor memoria es criandome en una comunidad de
puertorriquel‘ios. Sin preocuparse de donde estaba, siempre habia una "bendicion", un 'hola
compal”, 0 el orgullo que uno sentia cuando ola Ia palabra: jBon’cual

Siendo hijo de padres que emigraron de Puerto Rico, me criaron al estilo viejo. Lo que mas
querian mis padres para nosotros era una buena educacion. Siento alegrla en decir que las
escuelas de Southbridge me prepararon muy bien para mis afios en la universidad. Pero,
me estoy adelantando...

Despues de graduarme desde Southbridge High, me ful a Barry University en Miami,
Florida. Esta era la primera vez fuera de casa, y admito que estaba un poquito asustado.
Pero, hice muy bien en la universidad. Cuando se acabaron mis becas, no tenia suficiente
dinero para continuar mis estudios. Regrese a Southbridge y trabajo en la American Optical
por un tiempo.

Resulta que estaba trabajando en BI mlsmo sitio donde trabajo mi madre. Por fin pude ver y
apreciar por lo que pasaron mis padres para mantener nuestra familia. Alli lo pase muy
bien. Conocl a gente de Grecia, Polonia, y Puerto Rico. El trabajo era respetable, pero yo
todavia queria ir a la universidad. Decldi alistarme al ejército.

[Estuve en ei ejorcito por tres allos! Empezé como privado y sail sargento. Las horas
fueron Iargas, pero valla la pena. Lo mejor del ejército fue que conocl a mucha gente desde
diferentes partes de Ios Estados Unidos, y aprendi de diferentes culturas. Tenia mucho
prestigio y responsabilidad a Ios 22 altos. El ejército te ensel‘ia cOmo mantenerse en foco a
una meta. Mi meta era Ir a la universidad y lograr una educaciOn unlversitaria. En el ejército
obtuve eI “College Fund" y '6! Bill" para Ir a la universidad.

Este dinero me ayudo obtener mi bachillerato en GeograflalTurIsmo en Salem State College.
Desde ahl me ful a Rochester, Nueva York para mi maestria en Manejo de Turismo y
Hospitalidad en el Rochester Institute of Technology. Corrientemente, estoy en Michigan
State University terrninando mi doctorado en Recursos de Parque, Recreacion, y Turismo y
Estudios Urbanos. Lo que mas me asusto en todas las universidades fue que no sabla que
me esperaba. Pero, [era también la mejor partel Jamas crela que hubiera alcanzado a
tanto, pero con amor, fe y perseverancia, jiO hare!

Después de completar esta estudio, terminare con todos los requisitos de mi doctorado.
Pienso dar closes en una universidad. Afortunadamente, tenia profesores excelentes que
fueron ejemplares y me apollaron. Yo quisiera continuar esa tradiciOn (quizas con sus hijos).
Le debo muchas gracias a Dios, mi familia y amigos, mi comunidad puertorriquel'ia, y los
maestros de las escuelas de Southbridge. Todos me han ayudado en diferentes maneras.
[Graciasl

175

SAMPLE INTRODUCTION LETTER FOR SURVEY: ENGLISH

(OTRO LADO PARA ESPANOL)
February 9, 1999

«First» «Last»
«Num» «Street»
Southbridge, MA 01550

Dear «First»:

Greetings! I am a Ph.D. student in the Department of Park, Recreation and Tourism
Resources at Michigan State University. You are being asked to participate in a study.
The attached questionnaire concerning participation in Southbridge public parks is part
of a research project for my Ph.D. at Michigan State University. This study is
concemed specifically with determining how residents in Southbridge view their public
parks.

The results of this study will help to provide crucial data to academicians and
professionals interested in providing public leisure and recreation services to a diverse
population. The United States (and Southbridge) is changing in terms of its ethnic and
racial composition. Because of this, we need to better understand how people from
different cultural backgrounds interact; how they feel about the environment they live
in; and how this affects the use of public recreation settings.

You or your friends or family have indicated that you would be interested in
participating in this study. Only by obtaining the opinions of residents, such as
yourself, can I accurately determine the Puerto Rican community’s point of view on
Southbridge's public recreation areas. I guarantee that all information will be held in
the strictest of confidence.

The number at the lower, right hand corner on the front of the questionnaire is an
identification number for mailing purposes only. This number allows me to identify
who has responded to the survey so that I do not send a second survey. Once you
have returned the questionnaire, your information is kept, but your name is removed
from the database.

Your participation in this study is voluntary. The survey should take about 10-12
minutes of your time. It will be appreciated if you can complete the questionnaire
before February 28, 1999, and return it in the enclosed, self-addressed, stamped
envelope.

My sincerest thanks for your participation!

Edwin Gomez
Project Coordinator
Michigan State University

176

SAMPLE INTRODUCTION LETTER FOR SURVEY: SPANISH

(OTHER SIDE FOR ENGLISH)
9 de febrero del 1999

«First» «Last»
«Num» «Street»
Southbridge, MA 01550

Estimadola «First»:

Saludosl Soy estudiante de doctorado en el departamento de Parque, Recreacion y
Turismo en la Universidad Estatal de Michigan. Le solicito que participe en éste
estudio. El cuestionario adjunto acerca de la participacion en los parques publicos de
Southbridge es parte de un trabajo de investigacion para mi doctorado en Michigan
State University. Este estudio intenta detenninar como Ios residentes puertorriquefios
de Southbridge perciben sus parques pr’rblicos.

Los resultados de este estudio ayudaran a proveer informacion para académicos y
profesionales interesados en proveer servicios de recreacion pt’rblica para una
poblacion multicultural. Los Estados Unidos (y Southbridge) estan cambiando en
tenninos de su composicion étnica y racial. Debido a ésto, necesitamos entender
como la gente de diferentes culturas interaccionan; como se sienten del ambiente en
el que viven; y como esto afecta el uso de sitios de recreacion publica.

Usted o sus amigos y familiares han indicado que estaria interesado en participar en
este estudio. Con el analisis de sus opiniones podré deterrninar el punto de vista de
la comunidad puertorriquefla acerca de Ios sitios de recreacion publica en
Southbridge. Garantizo que toda la informacion sera confidencial.

El numero en la esquina derecha del cuestionario es un numero de identificacion
usado solamente para esta correspondencia. Esto sirve para poder identificar quién
me ha contestado el cuestionario, para no enviarle otro. Tan pronto usted conteste el
cuestionario, su nombre sera eliminado del banco de datos.

La participacion en el estudio es completamente voluntaria. Te tomara alrededor de
10 a 12 minutos para completar eI cuestionario. Apreciaré si usted completa eI
cuestionario antes del 28 do febrero del 1999. Por favor devuelva el cuestionario en
el sobre predirigido que se Ie provee.

[Mis mas sinceras gracias por su participacion!

Edwin Gomez
Coordinador de Proyecto
Michigan State University

177

SAMPLE INTRODUCTION LETTER FOR FOLLOW-UP SURVEY: ENGLISH

(OTRO LAoo PARA ESPANOL)
March 19, 1999

«First» «Last»
«Num» «Street»
Southbridge, MA 01550

«First»:

Hello! About a month ago, a questionnaire was distributed seeking your opinions on
Southbridge parks and other Information. Unfortunately, there have been problems with the
post office. Therefore, i know that you have not received the original survey, or have
received it late and you were not able to answer It on time. As of today, I have not yet
received your questionnaire.

It is important that I receive your questionnaire In order to assure the success of this study
and public recreation efforts. I obtained your name from friends and relatives because they
thought that you would be interested in participating in this study. In order for the results of
this study to be truly representative of the opinions of Puerto Ricans in Southbridge, it is
essential that each person chosen for the study return their questionnaire.

l have conducted this study because of the belief that Puerto Ricans in the Southbridge
community could provide a very unique point of view on how different ethnic groups in the
United States view their identity, community, and parks. The results of this study will help to
provide information to professionals Interested in offering services related to public
recreation for a diverse population.

If you have not completed me questionnaire, I ask that you please take a few minutes to
complete it. In the event that your questionnaire has been misplaced, a replacement is
enclosed. Your participation in this study is completely voluntary and confidential. The
survey should take about 10-12 minutes of your time. It will be appreciated if you can
complete the questionnaire before April 4, 1999 (Easter Sunday). Please return the
questionnaire in the self-addressed, business reply envelope (put it in your mailbox or drop
it off at the post office). Postage is free!

If you have already completed and returned the survey, please disregard this letter. Thank
you for your cooperation in this very Important matter.

Have a happy Easter and may God bless you!

Sincerely,

Edwin GOmez
Project Coordinator
Michigan State University

178

SAMPLE INTRODUCTION LETTER FOR FOLLOW-UP SURVEY: SPANISH

(OTHER SIDE FOR ENGLISH)
19 de marzo del 1999

«First» «Last»
«Num» «Street»
Southbridge, MA 01550

«First»:

[Holal Hace un mes atras, te envie un cuestionario pidiendo sus opiniones de los parques
de Southbridge y otra informacion. Desafortunadamente, han habido problemas con el
correo. Por eso, se que no he recibido el cuestionario, o lo recibio tarde y no ha podido
contestarlo a tiempo. Hasta hoy, yo no he recibido su cuestionario.

Es importante recibir su cuestionario para asegurar el exito de este estudio y los esfuerzos
de la recreacion publica. Obtuve su nombre atraves de sus amigos y familiares porque
ellos pensaron que usted estaria interesado en participar en este estudio. A fin de que Ios
resultados de este estudio representen verdaderamente las opiniones de los
Puertorriquenos en Southbridge, es esencial que cada persona escogida para el estudio
devuelva su cuestionario.

Hice este estudio con la cmenzia de que Ios Puertorriquenos en la comunidad de
Southbridge pueden proveer un punto de vista unico acerca de distintos grupos étnicos en
Ios Estados Unidos y como estos perciben sus identidades, comunidades, y parques. Los
resultados de esta investigacion ayudaran a proveer informaciOn a profesionales
interesados en ofrecer sus servicios en areas relacionadas a la recreacion publica para una
poblacion diversa.

En caso de que usted no haya llenado su cuestionario, Ie pido que por favor tome unos
pocos minutos para completarlo. Si usted perdio el cuestionario, aqui le envio otro. La
participacion en el estudio es completamente voluntaria y confidencial. Te tomara
alrededor de 10 a 12 minutos para completar el cuestionario. Apreciaré si usted completa
el cuestionario antes del 4 de abril del 1999 (“Domingo do Resurreccion”). Por favor
devuelva el cuestionario en el sobre predirigido que se Ie provee (ponlo en su buzon o
dejalo en el correo). No se necesita sellos, les gratis!

Si completo y envio el cuestionario anteriorrnente, favor de no prestarle atencibn a esta
carta. Gracias por su cooperacion en este asunto tan importante.

Espero que tenga unas felices Pascuas y ique Dios le bendigal

Sinceramente,

Edwin GOmez
Coordinador de Proyecto
Michigan State University

179

APPENDIX C

Conflrmatory Factor Analysis Concepts

180

CFA Concepts

The purpose of this section is to provide the reader with an
understanding of general confirmatory factor analysis concepts. It is meant to
be as a conceptual overview, and as such, the mathematics will not be
illustrated. The researchers primary concern is with measurement. The
researcher has an idea of what constructs are to be measured, and this is the
basis for the theoretical model. Theory indicates which constructs to measure.

Before subjecting the data to analysis, a multiple factor measurement
model was established in Chapter 3 (see Figure 10). Chapter 3 explained the
indicators which were used for each of the theoretical constructs in the model
on the basis of face validity. If the variables/Items do not meet face validity then
it is unreasonable to expect the constructs to pass test of lntemal consistency
and parallelism.

The second step is to test the internal consistency for all constructs that
have four or more indicators. The third step is to test for parallelism for all sets
of indicators. In and example of three factors, T, U, and V, one can test the
indicators of T with respect to U and V, one can test the indicators of U with
respect to T and V, and one can test the indicators of V with respect to T and U.

Discard indicators that do not comply with the internal consistency
theorem, or the parallelism theorem, or both. Repeat steps two and three
again, until a set of indicators that meet these three criteria have been
developed. And when the set of indicators pass those tests, then and only then

does it make sense to test the theoretical model. Implicit in these steps is an

181

understanding of the measurement model underlying the confirmatory factor
analysis.
The Measu ment M I

In Figure 18, the underlying constmct is called Tme Score, usually
abbreviated T. The word "True" in True Score is used in the sense that it is a
faithful representation of a constmct. This is often referred to as a latent or
unmeasured variable (Loehlin, 1987). The measurement model states that
Tme Score drives various indicators (which are the X’s) and each one of those
indicators is an imperfect indicator, that is, there is some error associated with

it. This model assumes a linear scaling model.

Figure 19. One Factor Measurement Model

 

T (True Score)“

l/xl/Xl‘xx

X1 X2 X3 Xk(lndicators)
i i i
e1 e2 e

3 ek (Errors)

' - True Score drives any number of indicators.

 

 

 

If X1 were a perfect indicator, the correlation between True Score and X1

would be 1.0. The extent to which this correlation (between item and True

182

Score) is not 1.0, there is error. This is known as error of measurement. It is
random response enor. Random response enor is some sort of non-systematic
perturbation that affects the response. It keeps the correlation between item
and True Score from being 1.0 (Frank J. Boster, personal communication,
March 2, 1999).

According to the classical theory of errors, if the (scaling) measurement
model is correct, any observed score (any X) has two components: it is affected
by True Score and it is affected by error (x = T + e ). For example, there are
two reasons for a respondent to check ”agree“ on an acculturation item. One is
his/her level of acculturation. The second is how much measurement error
there is associated with that item. These two things determine where the check
mark went on that item.

The error is defined by whatever is not defined by True Score. The
relationship between the indicator and True Score have a name. They are
called factor loadings or parameters. The first step in a CFA is to estimate the
parameters.

Estimating Parameters in a Measurement Model: The Centroid Method

This study uses Program CFA as the software program for the CFA
portion of this study. Program CFA was initially developed by John E. Hunter at
Michigan State University's Department of Psychology (April 20, 1985), and was
revised by Hunter and Mark A. Hamilton of the Department of Communication at
the University of Connecticut (May 11, 1992). Program CFA is a least squares,

oblique multiple groups program that performs confirmatory factor analysis. It

183

uses the Centroid Method to estimate the parameters (factor loadings) for the
measurement model. Factors are defined using communalities so that the
factors are the actual constructs; i.e. defined without enor of measurement.

Hunter’s Program CFA iterates three times using the Centroid Method.
The factor solution stabilizes using communalities (item reliabilities) in the
correlation matrix diagonal instead of ones in the diagonal. The use of
communalities in the diagonal corrects for attenuated correlations due to
measurement error by not assuming perfect reliability (i.e., an item correlating
with itself perfectly) (Hunter & Gerbing, 1982).
lntemal Consisteng Theorem

The internal consistency theorem is employed to test the validity of the
model. The internal consistency theorem states that the correlation between
any two indicators of the same underlying True Score is the product of the

correlations with True Score:

 

It”: 17ro (3)

 

The relationships that one observes in the model are all spurious
because they are driven by their common cause: T. This model, a partial
correlation model, predicts the correlation between the two variables that are
related spuriously (X1 & X2) as the product of their correlations with their cause.

So, the X1/X2 correlation should be, if the model is correct, equal to rm. rm.

184

Another way of saying this is that any two items (x3) should be correlated
so that the correlation is equal to the product of the factor loadings. If the
theorem fails, it suggests that "i" and "j" are not indicators of the same
underlying factor, and to sum them would be a case of invalid measurement.
Summation of the indicators can only occur if in fact this theorem held.
Multi-factor Model
Figure 19 illustrates a causal string with three constructs: T, U and V. There
are three measurement models, one for each constmct. There is error in
predicting U and error in predicting V. For T, the indicators are the X’s, and the
errors associated with those indicators. For U, the indicators are Y’s. and the
errors associated with those indicators. For V, the indicators are 23 and their

errors.

Figure 20. Multi-Factor Measurement Model
3“ iv
T\ > u \~ > v
AA?» [WEQ‘AY l/VN A

ii’fifvfi Z'iii

 

M

82 63 9; 31

 

 

 

The multi-factor model is an extension of the one factor model. In the
multi-factor model case, there are more than one factor, and it is necessary to

test the internal consistency theorem for each of the constructs (T , U, V). But

185

the multi-factor model alsomakes predictions about the indicators of the

different factors. This prediction is derived from the parallelism theory.

Parallelism Theorem

The parallelism theorem states that the correlation between ultimate
indicators of different factors (e.g., X’s are indicators of T, and Y's are indicators
of U) would be the triple product of the correlation of X with its True Score T (its
factor loading) and Y with its Tme Score U (its factor loading) times the

correlation between the two true scores:

 

r = rr
xy xTyUTU (4)

In a multi-factor model, a parallelism test is conducted for each pair of

 

factors. That is to say, one can test the indicators of one factor and test if they
are parallel with respect to indicators of another factor. If the items are parallel,
this is further indication that the items measures their respective construct, and
not any other construct. Another name for parallelism is external consistency.
By testing parallelism, a more rigorous test of the content validity of ones
measures is achieved.

The test of parallelism is a more stringent test than internal consistency.
It is important to impress upon the reader that a lot of the invalidity of an item
does not show up until the parallelism test. If an item correlates higher with
another factor than its own, then regardless of whether it is internally consistent,

it is an invalid indicator because it correlates higher on another construct.

186

It is critical that the indicators are ultimate measures of the same
factor/construct because if they are not, they should not be summed and
correlated. If items pass both the internal consistency and parallelism tests,
then it makes sense to sum the items. The factor/construct consisting of the
summed items is correlated with the other constructs (factor correlation matrix).
These are the correlations (corrected for attenuation) used in the path analysis.
Confirrnatogy Versus Explorateu Facter Analysie

CFA begins with a measurement model. It states that certain indicators,
and only certain indicators, are indicators of a specific factor, and the other
indicators are indicators of another specific factor. In an exploratory factor
analysis (EFA) model, every indicator is driven by every factor (see Figure 20).
Fundamentally, this is a major difference concerning how a factor is defined.
Conceptually, in terms of partial correlations, in the CFA model, the correlation
between X, and X2, controlling for T, should equal 0 (because it’s driven by a
common cause). In a CFA, the Y,/Y2 correlation would also be zero when
controlling for U, and so on. But when X and Y are correlated, the correlation
will not be zero. A CFA is reflective of Tyson and Bailey's (1970) clustering
approach to non-orthogonal solutions.

In an EFA, the correlation between X’s and Y’s will be zero. With EFA,
you correlate any X with any Y and control for all factors (F1, F2, F3...Fk). The
partial correlation between the items, controlling for all the factors is zero. It is
necessary to control for all factors in an EFA. If, for example, there were six

indicators and two factors, the math would imply the model presented in Figure

187

21. Each factor is a linear combination of all the indicators, not just those that

are said to be indicators of that factor.

Figure 21. EPA Model

 

F1 F2
/ o’o’q
I i” i I I I

at e! 3

 

 

 

In CFA, a factor was defined as tme score. True score (an estimate) is
defined as the sum of all its indicators that were valid measures of it. Theory
guides the analysis. But, in an EFA, factors are summary variables. These
summary variables are created in a non-theoretical fashion. There is no
measurement model with which to start. These variables, in fact, are created to
summarize all of the information in the correlation matrix. CFA starts with a
model that makes substantive sense, and tests the notion that each set of
indicators is driven by only one factor. Kline (1994) suggests that ”many
psychologists believe that confirmatory factor analysis is in principle a superior
method to exploratory factor analysis because it tests hypotheses, which is

fundamental to the scientific method” (p. 80).

188

APPENDIX D

Data Analysis Matrix and Chart

189

CORRELATION MATRIX OF 22 ITEMS USED IN
FINAL CONFIRMATORY FACTOR ANALYSIS 3 b M °

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BEN ACC DIS SID REC SES

3 4 5 e 3 11 12 14 19 21 22 23 24 25 29 27 so 31 32 34 35

3 .47 .51 .43 .15 .15 1e .13 —.01 .01 -.01 -.01 .05 .05 .07 .03 11 .17 .17 -13 -.13

4 .01 .53 .51 17 .13 1a .20 -.01 .01 -.01 ..01 .06 05 .09 .07 13 .19 .19 -15 -15

5 .01 .01 .55 .13 .20 .20 .22 -.01 .01 -.01 -.01 .06 09 .03 .03 14 .20 .20 —1e -13

e .01 .01 .01 .13 .13 .13 .2o -.01 -01 -01 -.01 .03 05 .03 .07 12 .13 .13 -14 -14

a .02 .01 .01 .03 .30 .30 .37 .25 21 .31 .29 .23 .25 37 .35 -.03 -.04 .-04 -15 -15

11 .03 .05 .02 .02 .02 .97 73 .27 23 .34 .32 .30 27 .40 .33 .03 -.05 —.05 -16 -13
12 .03 .02 .01 .05 .03 .01 .73 .27 .23 34 .32 .30 27 .40 .33 -.03 -05 -.05 .15 -16
14 01 .02 .05 .03 &&£_ .30 20 .33 .35 .34 .30 45 .42 -.04 -.03 -.06 -13 -1a
19 10 .04 .01 .03 .04 .04 .09 .07 41 .62 .57 .20 .13 .27 .25 ~.04 -.00 -.00 -.03 -03
21 .01 .04 .01 .01 .04 .03 .01 03 .03 .52 .43 .17 15 .22 .21 -.05 -.05 -.05 -.07 -.07
22 02 .05 .09 .03 .03 .03 .03 .oo .oo .02 .71 .25 23 .34 .32 -.05 -.07 -.07 .10 -10
23 04 .03 .04 .02 .02 .05 .01 .02 £$ 03 _ .23 21 .31 .29 -.05 -07 -.07 -.10 -.1o
24 .02 .02 .03 .03 13 .03 .02 .oo .02 04 .04 .03 27 .39 .37 01 .02 .02 -.11 ..11
25 .04 .01 .05 .01 .1o .05 .01 .03 .13 13 .13 1e .02 .35 .33 01 .02 .02 -10 -.1o
26 11 .00 .02 .07 .03 .04 .06 .05 .12 .03 12 .11 .01 .00 .49 02 .02 .02 —14 .14
27 .02 .03 .02 .03 .04 .00 .11 .03 04 .04 .08 io_1.o=2_ .02 .02 .02 —.13 -.13
30 .01 .05 .oo .02 .03 .01 .02 .oo .03 03 .02 .01 .05 I: 04 .02 _3—13—1 —.03 -.03
31 .07 .oo .07 .01 .02 ca .03 .01 .02 02 .03 .01 .01 02 .02 .03 .00 .43 -.05 -95
32 .09 .07 .03 .01 .01 .oo .01 .04 .04 .04 97 .02 .00 02 .07 .14 £é= .95 —.05
34 .05 .03 .oo .10 .02 .1o .05 .07 .04 .03 11 .04 .03 oo .03 .03 01 .09 .15 —:
35 .03 .02 .03 .06 .10 .02 .02 .07 .03 07 .02 .01 .14 02 .oo .03 .02 .09 .03 &=
3 4 5 0 a 11 12 14 19 21 22 23 24 25 29 27 30 31 32 34 35

BEN 35 .73 .79 .70 .25 .21 .23 14 -.09 -02 03 -.01 14 .05 .10 .03 .20 .31 .17 -.1e -.25
too 19 .22 .22 .27 .74 .31 .31 .90 .29 31 .45 .37 .43 39 .46 .40 -.02 —.09 —.05 -.14 -.27
01s -.00 -.01 .02 ..02 .33 .41 .36 .41 7o 59 .33 .31 .30 .44 .24 .29 -.09 .07 —.05 -.03 -.15
510 03 1o .13 17 .31 .53 .51 .57 .33 30 .43 .43 .54 .49 72 .93 -02 .07 .04 .13 -.23
age 24 .27 .29 23 -.07 -12 -.07 -.03 -.00 -14 -.06 -09 -.02 03 .04 .14 .52 .33 .31 —.04 .10
as —.23 —15 .25 -13 -32 -13 -19 -24 —.1o -.19 -.07 —.10 -23 .15 -13 .09 -.06 -.19 .07 .75 75

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

' - Numbers in the first row and column (of upper matrix) represent the items.

° - Numbers above the diagonal (in upper matrix) are the PREDICTED correlations.

° - Numbers below the diagonal (in upper matrix) are the errors/residuals.

‘ - Numbers in the double line squares on the diagonal are inter-item correlations. and
were used to assess internal consistency.

° - The lower matrix is the item by factor correlation matrix (used to assess parallelism), with
bold numbers representing the factor loadings of each item on its factor.

190

LOCAL AND GLOBAL TESTS FOR THREE EPRP MODELS

Individuel Link AnafgsLs

Nen-Conetrainfi Link Em .z-v_alue mtg
Original Model
SID—>SES -.07 -.64 .522
ACC—*BEN -.20 m .027
ACC—>DIS -.10 -1.17 .242
DIS-+BEN -.10 -1.08 .280
ACC-+REC .16 1 .49 .136
Revised Model
SID—>SES -.07 -.64 .522
SES-*DIS -.02 -.17 .863
SES—*REC -.06 -.55 .584
Final Revised Model
SID—>SES -.07 -.64 .522
SES—vDIS -.02 -.17 .863
SID—>BEN -.02 -.19 .848
SES—>REC -.07 -.55 .581
DIS—>REC -.03 -.30 .764

Overall Chi Square for Each Model

Model Chi Sggare g1 tvalue
Original 10.09 5 .073
Revised .74 3

 

01

Final Revised .87

191

REFERENCES

192

REFERENCES

Anderson, J.C., & Gerbing, D.W. (1988). Structural equation modeling in
practice: A review and recommended two step approach. Psychological Bulletin, '
fl (3), pp. 411-423.

Allison, M. T. (1988). Breaking boundaries and barriers: Future
directions in cross-cultural research. Mel of Leieure Reseerch, 10, pp.
247-259.

Baldwin, J. A., & Bell, Y. R. (1985). The African self-consciousness
scale: An africentric personality questionnaire. The Weetem Jeumel ef Black

Studiee, 9 (2), pp. 61—68.
Baron, R.B., & Kenny, DA. (1986). The moderator-mediator variable

distinction in social psychological research: Conceptual, strategic and statistical

considerations. Joumal ef Personalny ang Seciel Psyeholegy, 51 (6), pp.
1 1 73-1 1 82.

Barona, A., & Miller, J. A. (1994). Short acculturation scale for Hispanic
youth (SASH-Y): A preliminary report. Hiseenic Joumel of Behevieral Sciences,
1_e (2), pp. 155-162.

Burch, W. R., Jr. (1969). The social circles of leisure: Competing
explanations. Jegmel of Leisure Reseerch, 1, pp. 125-147.

Burlew, A. K., & Smith, L. R. (1991). Measures of racial identity: An
overview and a proposed framework. The Journal of Black Psycholegy, 17 (2),

pp. 53-71.

193

Brown, RE. (1982). I he new New Englanders. Worcester, MA:
Commonwealth Press.

Carr, D. S., & Williams, D. R. (1993a). Understanding diverse
recreationists: Beyond quantitative analysis. In P. H. Gobster (Ed.), Managing
grben end high-ea fireation settings (pp. 101-106). Chicago, IL: USDA
Forest Service North Central Forest Experiment Station.

Carr, D. S., 8 Williams, D. R. (1993b). Understanding the role of

 

ethnicity in outdoor recreation experiences. Joumel of Leieure Research 25 (1 ),
pp. 22-38.
Chavez, D. J. (1993). Vieiter Qrceetions of crewding and diecrimination

at two nationel feaete in Segthern Celifomia (Research Paper PSW-RP-216).
Albany, CA: USDA Forest Service Pacific Southwest Research Station.

Christensen, J. E., & Dwyer, J. F. (1994, April 10-12). Examining
African American and White outdoor recreation participation after demographic
standardization on selected characteristics. In G. A. Vander Stoep (Ed.),
Proceedings of the 1994 Northeestem Recreation Symmsium (pp. 159-161).
Sarasota Springs, NY: USDA Forest Service, Northeastern Forest Experiment
Station.

Craig, W. (1972). Recreational activity patterns in a small Negro urban
community: The role of the cultural base. Emnemic Geegraehy, 48, pp. 107-

115.

194

Cortes, D. E., Rogler, L.H., & Malgady, R. G. (1994). Biculturality
among Puerto Rican adults in the United States. Amerign Journal of
Communfiy Peycholegy, 22 (5), pp. 707-721.

Cox, T., Jr. (1990). Problems with research by organizational scholars
on issues of race and ethnicity. The rn lofA lied B havioral Science 26,
pp. 5-23.

Datz, B. (1998, December 13). Jobs fueled migration: Industrial
Southbridge drew Puerto Ricans. Sunday Telegram, p. A23.

Davis, J. A. (1985). The legic of gggl order. Newbury Park, CA:
SAGE Publications, Inc.

Dillman, D. A. (1978). ll and t I h n rv : Th Ides n
methfi. New York, NY: John Wiley & Sons.

Duncan, 0.0. (1975). lntmdggien te etmctural eguation models. New
York, NY: Academic Press.

Dwyer, J. F. (1993, April 18-20). Customer diversity and the future
demand for outdoor recreation. In G. A. Vander Stoep (Ed.), Proeeedings of
the 1993 Northeastem RecLeetien Sympeeigm (pp. 59-63). Sarasota Springs,
NY: USDA Forest Service, Northeastern Forest Experiment Station.

Dwyer, J. F. (1994, August). Qgetemer givefi'fl and the future demand
for outdmr reegetion (General Technical Report RM-252). Fort Collings, CO:
USDA Forest Service, Rocky Mountain Forest and Range Experiment Station.

Dwyer, J. F., & Gobster, P. H. (1992). Recreation opportunity and

cultural diversity. Park& Recreation 27(9), pp. 22-33,128.

 

195

Echegaray, C. (1998, December 13). Commonwealth or State?: Island’s
vote divides Puerto Ricans in Worcester County. Sunday Telegram, pp. A1,
A22.

Edwards-Hewitt, T., 81 Gray, J. J. (1995). Comparison of measures of
socioeconomic status between ethnic groups. Psycholegical Remus, 77, pp.
699-702.

Falk, J. H. (1995). Factors influencing African American leisure time
utilization of museums. Mm_al of Leieure Reeearch. 27 (1 ), pp. 41- 60.

Feagin, J. R. (1994). The continuing significance of race: Antiblack
discrimination in public places. Amerign Sociolegigl Review, 56, pp. 101-116.

Floyd, M. F. (1998). Getting beyond marginality and ethnicity: The

challenge for race and ethnic studies in leisure research. Journal of Leisure

 

Research 30(1), pp. 3-22.

Floyd, M. F., Gramman, J. H., 81 Saenz, R. (1993). Ethnic factors and
the use of public outdoor recreation areas: The case of Mexican Americans.
Leisure Seienees, 15, pp. 83-98.

Floyd, M. F., McGuire, F. A., Shinew, K. J., & Noe, F. P. (1994). Race,
class, and leisure activity preferences: Marginality and ethnicity revisited.
Journal of Leisure Research, 26 (2), pp. 158-173).

Frazier, E. F. (1957). The black bogrgmisie. Glencoe, IL: Free Press.

Gordon, M. M. (1964). Assimilatien in American life. New York, NY:

Oxford University Press.

196

Gramman, J. H. (1996). Ethnic'mr, race. and outdoor recreation: A
revi w of trends Ii 11 rch (Miscellaneous Paper R-96-1). Vicksburg,
MS: Army Corps of Engineers, Waterways Experiment Station.

Heise, DR. (1975). Wis. New York, NY: Wiley.

Hendee, J. C. (1969). Rural-urban differences related in Outdoor
recreation participation. The Jeurnal ef Leisure Reeearch, 1 (4), pp. 333-341.

Henderson, K. A. (1998). Researching diverse populations. Joumel of

Leisure Research 30(1), pp. 157-170.

 

Hunter, J.E., & Gerbing, D.W. (1982). Unidimensional measurement,
second order factor analysis, and causal models. Research i_n Organizational

Behavior 4 pp. 267-320.

 

Hutchinson, R. (1988). A critique of race, ethnicity, and social class in
recent leisure-recreation research. Jeume! ef Leieure Research, 20 (1), pp.
10-30.

Hutchison, R. (1987). Ethnicity and urban recreation: Whites, Blacks,
and Hispanics in Chicago’s public parks. Journal of Leisrure Research. 19(3),
pp. 205-222.

lso-Ahola, S. E. (1980). The geie! geyeholegy ef leisure end recreation.
Dubuque, IA: William C. Brown.

 

Jackson, R. G. (1973). A preliminary bicultural study of value
orientations and leisure attitudes. Joumel of Leiegre Research, 5 (1 ), pp.

1 0-22.

197

James, L.R., 8. Brett, J.M. (1984). Mediators, moderators, and tests for
mediation. Joumel ef Apelifi Peyehelegy, ee (2), pp. 307-321.

Johnson, C. Y., Bowker, J. M., English, D. B. K., & Worthen, D. (1997).
Theoreti l rs iv f thnic’ n t r r re ti n A review and
eynthesis of Afrign-American partieipetion (General Technical Report SRS-11).
Asheville, NC: USDA Forest Service, Southern Research Station.

Johnson, C. Y., Bowker, J. M., English, D. B. K., & Worthen, D. (1998).
Wildland recreation in the rural South: An examination of marginality and
ethnicity theory. Jeumal of Leieure Research, 30(1), pp. 101-120.

Kelly, J. R. (1980). Outdoor recreation participation: A comparative
analysis. Leisure Scienge, 3 (2), pp. 129-154.

Kenny, DA (1979). Qerrelatien eng eegselhy. New York, NY: Wiley.

Kline, P. (1994). An eesy ggige te fa_cter enalysis. New York, NY:
Routledge.

Kline, RB. (1998). Prin i n ractice of tru le uation
modeling. New York, NY: Guilford.

Klobus—Edwards, P. (1981). Race, residence, and leisure style: Some
policy implications. Leisgre SeienmeA (2), pp. 95-111.

Lindsay, J. J., &Ogle, R. A. (1972). Socioeconomic patterns of outdoor
recreation use near urban areas. Jouma‘l of Leisure Research, 4 (2), pp. 19-24.

Loehlin,J.C. (1987). Letent veh'eble mflels. Hillsdale, CA: Erlbaum.

Massey, D.S., & Denton, N. A. (1993). Amerign agertheid; Segregation

and the making of the underclass. Cambridge, MA: Harvard University Press.

198

McKee, J. B. (1993). Seciolegy and the race problem: The fajure of e
perspective. Urbana, IL: University of Illinois Press.
McMillan, D. W., 8 Chavis, D. M. (1986). Sense of community: A

definition and theory. Jeumel of Cemmpnhy Psycholegy, 14, pp. 6-23.

McMillan, J. B. (1983). The social organization of leisure among
Mexican-Americans. Jppmal of Leiegm Reeeamh, 15 (2), pp. 164-173.

Meeker, J. W., Woods, W. K., 8 Lucas, W. (1973, Fall). Red, white,
and black in the national parks. North Amerign Revifl, pp. 3-7.

Murdock, S. H., Backman, K., Colberg, E., Hoque, MD. N., 8 Hamm, R.
R. (1990). Modeling demographic change and characteristics in the analysis of
future demand for leisure services. Leieug Servige, 12, pp. 79-120.

Murdock, S. H., Backman, K., Hoque, MD. N., 8 Ellis, D. (1991). The
implications of change in population size and composition of future participation
in outdoor recreation activities. Joprnel of Leieure Reeeerch, 23 (3), pp. 238-
259.

Myrdal, G. (1944). An American dilemma. New York, NY: Harper and
Brothers.

Neulinger, J. (1981). The peyehelggy pf leisure (2"‘1 ed.). Springfield, IL:
Charles C. Thomas, Publisher.

Phillip, S. F. (1994). Race and tourism choice: A legacy of
discrimination? Annale ef Touriem Reeearch, 21 (3), pp. 479-488.

Phillip, S. F. (1997). Race, gender, and leisure benefits. Leisure

 

Science, 19, pp. 191-207.

199

Puerto Ricans say ‘no’ to statehood. (1998, December 14). Southbridge
Evening Newe, p. 13.

Rabinowitz, J. (1998, December 15). Southbridge residents respond to
Puerto Rican vote. Seuthhngge Evening Newe, pp. 1, 16.

Resnicow, K., 8 Ross-Gaddy, D. (1997). Development of a racial identity
scale for low-income African Americans. Johmel of Black Studies, 28 (2), pp.
239-254.

Rodriguez, C. E. (1996). Puerto Ricans: Between Black and White. In
C. E. Rodriguez, 8 V. Sénchez Korrol (Eds.), Historigel perepeetry‘ ee on Pperto
Rican eurvivel in the Unitfl Statee (pp. 25-35). Princeton, NJ: Markus Weiner
Publishers.

Salant, P., 8 Dillman, D. A. (1994). Hew _te mngpet yeer ewn eurvey.
New York, NY: John Wiley and Sons, Inc.

Saldafla, D. H. (1994). Acculturative stress: Minority status and distress.
Hispenie Jeumel ef Behaviohel Seiencee, 16 (2), pp. 116-128.

Sanchez, J. l., 8 Brock, P. (1996). Outcomes of perceived
discrimination among Hispanic employees: ls diversity management a luxury or
a necessity? Acagemy of Menagement Jeumel, 39 (3), pp. 704-719.

Schumen, H., 8 Gruenberg, B. (1972). Dissatisfaction with city services:
ls race an important factor? In Harlan, H. (Ed.), PflQIQ end mljtig in urban

sociehr (pp. 369-392). Beverly Hills, CA: Sage Publications, Inc.

200

Shaull, S. L., 8 Gramman, J. H. (1998). The effect of cultural

assimilation on the importance of family-related and nature-related recreation

among Hispanic Americans. Jopmel of Leisure Research, 30 (1 ), pp. 47-63.

Shavelson, R. J. (1992). Statistical reasoning for the behavioral
sciences (3" ed.). Needham Heights, MA: Allyn 8 Bacon (A Simon 8 Schuster

Company).

Shinew, K. J. (1995), Gender, race, and subjective social class and their
association with leisure preferences. Leisum Scienge, 17, pp. 75—89.

Singh, B. (1981). Race, ethnicity and class: Clarifying relationships and
continuous muddling through. Jeumal of Ethnie Stpgiee, 9 (2), pp. 1-19.

Stamps, S. M., 8 Stamps, M. B. (1985). Race, class and leisure
activities of urban residents. Journal pf Leieuge Reeearch, 17 (1 ), pp. 40-56.

Stodolska, M., 8 Jackson, E. L. (1998). Discrimination in leisure and
work experienced by a white ethnic minority group. Jopmel ef Leisure
Research 30 (1 ), pp. 23-46.

 

Stokes, J. E., Murray, C. B., Peacock, M. J., 8 Kaiser, R. T. (1994).
Assessing the reliability, factor structure, and validity of the African Self-

Consciousness Scale in a general population of African Americans. Journal of

Black Psyeholegy, 20(1), pp. 62-74.

Sugme, T. J. (1996). The Origine ef the Men ch’sis: Race and
ingualm in mstwer Det_reit. Princeton, NJ: Princeton University Press.

201

Tanaka, J.S. (1987). "How big is big enough?": Sample size and
goodness of fit in stmctural equation models with latent variables. Child

Development, 58, pp. 134-146.

Taylor, D. E. (1992). Urban park use: Race, ancestry, and gender. In P.
H. Gobster (Ed.), Mehaging urban end high-pee aereatien settings (pp. 82-86).
Chicago, IL: USDA Forest Service North Central Forest Experiment Station.

Toth, J. F., Jr., 8 Brown, R. B. (1997). Racial and gender meanings of
why people participate in recreational fishing. Leieure Sciences, 19, pp. 129-
146.

Tyson, RC, 8 Bailey, DE. (1970). Clueter Anelyeie. New York, NY:
McGraw-Hill.

Wagner, J., 8 Soberon-Ferrer, H. (1990). The effect of ethnicity on
selected household expenditures. The Smial Scieng Journal, 27(2), pp. 181-
198.

Washbume, R. F. (1978). Black under-participation in wildland
recreation: Alternative explanations. Leisure Scienge, 1 (2), pp. 175-189.

Watson, M. A. (1992). Researching minorities. Journal of the Market

Research Society, 34(4), pp. 337-344.
West, C. S. (1994). Afro-centricity: Moving outside of the comfort zone.

Journal of Physieel Egugtien, Rareetien, and Dance, 65(5), pp. 28-30.
West, P. C. (1989). Urban region parks and Black minorities: Subculture,
marginality, and interracial relations in park use in the Detroit metropolitan area.

Leisure Sciences 11, pp. 11-28.

 

202

Wilson, W. J. (1996). When work disappeegs: The worid of the new
urban poor. New York, NY: Alfred A. Knopf, Inc.

Winter, P. L. 8 Chavez, D. J. (1998, May 29). Conducting research in
the Hispanie commpnhy. [Handout]. Paper presented at The 7‘h lntemational
Symposium on Society and Resource Management, Columbia, MO.

Witt, P. A., 8 Ellis, 6.0. (1989). Th l i re ia nosti batte : Users

Manual. State College, PA: Venture Publishing, Inc.

 

Woodard, M. D. (1988). Class, regionality, and leisure among urban
Black Americans: The post-civil rights era. Jopmal pf Leiepre Reeeerch, 20 (2),
pp. 87-105.

Xie, Y., 8 Goyette, K. (1997). The racial identification of biracial children
with one Asian parent: Evidence from the 1990 Census. Seejel Forcee. 76 (2),
pp. 547-570.

Young, J.W. (1977). The function of theory in a dilemma of path analysis.
Journal pf Applim Peychelegy, 62, , pp. 108-110.

203