Perceived and Objective Neighborhood Characteristics Associated with Physical Activity Among Asian Americans Residing in Philadelphia County

PERCEIVED AND OBJECTIVE NEIGHBORHOOD CHARACTERISTICS ASSOCIATED WITH PHYSICAL ACTIVITY AMONG ASIAN AMERICANS RESIDING IN PHILADELPHIA COUNTY

A Dissertation Submitted to the Temple University Graduate Board

In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY

by Aisha Bhimla May 2020

Examining Committee Members:

David B. Sarwer, Ph.D., Advisory Chair, Department of Social and Behavioral Sciences Grace X. Ma, Ph.D., Department of Clinical Sciences Levent Dumenci, Ph.D., Department of Epidemiology and Biostatistics Lois A. Butcher-Poffley, Ph.D., Department of Kinesiology Kevin A. Henry, Ph.D., External Member, Department of Geography and Urban Studies Andrew T. Kaczynski, Ph.D., External Member, University of South Carolina

ABSTRACT

Introduction: Physical activity is an important lifestyle behavior known to improve overall health as well as provide substantial reductions in the risk of chronic illnesses, including obesity, cardiovascular disease, certain cancers, hypertension, stroke, depression, and anxiety. The built environment that surrounds where individuals reside has substantial impacts on physical activity behavior. Neighborhood walkability and the availability of physical activity resources such as parks and facilities within neighborhoods serve as opportunities to engage in leisure-based physical activity and active transportation. The purpose of this study was to determine whether objective and perceived built environmental characteristics were associated with neighborhood physical activity levels among Asian Americans living in Philadelphia County, Pennsylvania.

Methods: Participants were recruited from July to December 2019. A Built Environment and Physical Activity questionnaire that included socio-demographics, the Neighborhood

Environment Walkability Scale (NEWS), and the Neighborhood Physical Activity

Questionnaire (NPAQ) were administered in English, Mandarin, Korean, and

Vietnamese. The Walk Score was obtained online for each participant’s address.

Furthermore, proximity to parks and recreational facilities was measured using

Geographic Information Systems (GIS) to find the network distance to the nearest facility and park. Zero-inflated negative binomial regression models were used while controlling for the clustered sample design and for socio-demographic characteristics. Physical activity was generally examined with respect to active transportation, recreational walking/cycling, and overall physical activity. Additional analyses were also conducted

to examine how perceived versus objective characteristics, and ethnic group differences explained the relationship between built environment predictors and meeting the physical activity guidelines. These findings were assessed through logistic regression models.

Results: In research question 1, greater perceived neighborhood walkability was associated with an increased likelihood of engaging in recreational walking/cycling and overall physical activity, but not active transportation. When examining subscales separately, there were differences in which built environmental factors affected various physical activity domains. Physical barriers to walking and fewer cul-de-sacs predicted fewer hours of engaging in active transportation. Having fewer cul-de-sacs, more physical barriers and higher residential density predicted lower odds of engaging in active transportation. Greater perceived land-use mix, lower perceived crime, and having more cul-de-sacs predicted greater hours and lower odds of engaging in recreational walking/cycling. Greater perceived land use mix, aesthetics, more cul-de-sacs, and lower residential density were associated with greater weekly METs of physical activity and a greater likelihood of engaging in overall physical activity. In research question 2, ethnic identity moderated the relationship between perceived neighborhood walkability for active transportation and moderate-intensity physical activity but not overall physical activity. In research question 3, the Walk Score was not statistically significantly associated with various participation in physical activity domains. Lastly, results of research question 4 illustrated that longer driving time to the nearest park was associated with greater METs associated with MVPA and overall PA, and to a specialty exercise facility was linked to greater METs associated with walking and active transportation.

iii

Furthermore, a shorter walking time to a park was associated with less weekly METs of walking, active transportation, and overall PA, but greater odds of engaging in walking and active transportation. A greater walking time to the nearest specialty exercise facility was associated with greater active transportation and MVPA, but decreased odds of walking and active transportation. Greater driving or walking time/distance to the nearest gym/recreation center were not associated with any type of physical activity.

Conclusions: This study illustrated that the built environment could enhance participation in various forms of physical activity among Asian Americans. Having wide and accessible destinations in one’s neighborhood, known as land-use mix, promotes overall physical activity. Furthermore, greater aesthetically pleasing environments and areas with cul-de-sacs can provide spaces to engage in various forms of physical activity. Perceived neighborhood walkability had a strong impact on physical activity behaviors in this population, while mixed evidence was found for the influence of Walk Score and proximity to parks and recreational facilities as predictors of physical activity. While this study provides unique findings among a predominantly Asian American immigrant population, study limitations and recommendations for research suggest future areas to investigate. Nevertheless, major findings can be used to inform environmental and policy strategies to maximize active transportation and recreational based physical activity in order to sustain the health of Asian-American immigrants living in urban areas.

DEDICATION

To my parents, Ishaq and Estrellita Bhimla

To the study participants and community organizations who dedicated their time to take

part in this study, thank you for all that you continue to do to make this city vibrant,

unique, and an overall better place.

ACKNOWLEDGMENTS

I am deeply indebted to many individuals who have supported me on this Ph.D. journey during my five years at Temple. I would like to thank my advisor/chair, Dr.

Sarwer, for supporting me in the process of completing this project and for drastically helping me improve my scientific writing and critical thinking skills. I am truly grateful to have finished my final year of my Ph.D. under your guidance, expertise, and continuous encouragement to get me to the finish line. Second, I would like to thank Dr.

Ma for her unwavering support and mentoring since the beginning of my Ph.D. A major reason I came to Temple was for the opportunity to be part of the Center for Asian

Health-- thank you for believing in me, for allowing me to be part of your team, and for the immense opportunities you have provided to allow me to grow into a health disparities researcher. I’d like to thank Dr. Butcher-Poffley (Dr. B), for always being a helping hand and all that she has contributed to my learning and growth as a kinesiology professional. Thank you, Dr. Dumenci, for your willingness to spend a significant amount of time with me in enhancing my statistical and data analysis skills over the past years. I am immensely thankful to Dr. Henry and Dr. Kaczynski (We the North!) for providing their expertise and assistance on my dissertation as experts in the field.

This project would not have been possible with the help of several colleagues--I am tremendously thankful for Phuong, Chul, Amy, and Jin, for their support in collecting data for this project. Also, I am enormously thankful to Will Tagliamonte from the

Spatial Analytics Lab at Temple for the time he spent assisting me on the GIS portion of this project. I would also like to express my gratitude towards CPH for awarding me the

Visionary Research Fund and the support of grant funding by Dr. Ma and The Center for

Asian Health to carry out the dissertation study activities.

My professors, classmates, and co-workers in the Department of Kinesiology have supported me since I began the program. Thank you to Dr. Sachs for shaping me into the researcher I am today and the life lessons you have taught me. A big thank you to

Dr. Tierney and Dr. Gehris for always caring, and the great career/life advice you have provided. Sean, Andrea, and Rosalind, thanks for all that you’ve done to support me as a student and TA and for your genuine/kind personalities. I am so thankful for Mr. Harris,

Dr. Phillips, and Dr. Rooney for the incredible opportunities that I was able to have as an

A&P TA and which I hope to apply as a future professor one day. Lastly, thanks to my classmates, Ksenia Power, Allison Casola, Eric Aurelien, Chrissy Garcia, Jen Ciaccio,

Amanda Manu, Gabi Salvatore, and the A&P TA’s who made this experience all the better.

A tremendous thanks to all of my incredible co-workers at the Center for Asian

Health: Lin, Wenyue, Yin, Minsun, Kerry, Tina, Anastasia, Tyrell, Liz, Timmy, Taylor, and the late Joanne. These past five years have been the best learning experience, and I am grateful to have such a passionate, hardworking, and intellectual team to work with every day. A huge thank you for also helping me on this dissertation, from Taylor who proofread my paper, Lin and Timmy who helped me understand my data, and Kerry and

Julia for letting me practice my defense with you on Zoom. You guys are the best!

Lastly, a huge thanks to family and friends for all of their emotional support and love. Thanks to Julia Trout and my Toronto besties Shawn, Lisa, and Steph for the years of loyal and genuine friendships. To my big siblings Amina and Omar, thank you for

vii

raising me to be independent and intelligent. Leila, Gabriel, and Amelina, you are all the cutest and make me a proud aunt. To Victor, thank you for sticking by my side for every good and bad day these past 7+ years and all that you have sacrificed in order for me to accomplish my goals. You have made me happier, hardworking, appreciative, and an overall better person ever since we met on those tennis courts at USF. Lastly, thank you to my parents, who have always strived to provide a better life for me in Canada and here in the US—I could have never achieved this without the strength, hard work, and perseverance that you have instilled upon me.

viii

TABLE OF CONTENTS Page ABSTRACT ...... ii DEDICATION ...... iv ACKNOWLEDGMENTS ...... vi

CHAPTER 1. INTRODUCTION ...... 1 Background ...... 2 Need for Study ...... 8 Statement of the Problem ...... 15 Research Questions and Objectives ...... 17 Delimitations ...... 19 Limitations ...... 21 Definition of Terms...... 23 2. REVIEW OF LITERATURE ...... 28 Introduction ...... 28 The Social Ecological Model and Physical Activity Promotion ...... 28 Psychosocial/Cultural Mediators and Mechanisms of the Built Environment on Physical Activity ...... 31 Physical Activity Domains and Measures ...... 36 Built Environment and Physical Activity ...... 38 Built Environment Components ...... 50 Built Environment and PA among Underserved Communities ...... 63 Built Environmental Predictors of Physical Activity in Asian Americans ...... 73 Determinants of Physical Activity among Asian Americans ...... 77 Summary ...... 79 3. METHODS ...... 81 Research Design...... 81 Research Sites ...... 81 Participants ...... 82 Instrumentation ...... 83 Procedures ...... 92 4. RESULTS ...... 101 Descriptive Statistics ...... 101 Bivariate Analysis ...... 113 Analysis of Research Question 1 ...... 116

Analysis of Research Question 2 ...... 128 Results of Research Question 3 ...... 138 Results of Research Question 4 ...... 143 Additional Analyses ...... 157 5. DISCUSSION ...... 161 Summary ...... 161 Strengths and Limitations ...... 186 Implications...... 190 Conclusion ...... 197 REFERENCES CITED ...... 199 APPENDICES A. ELIGIBILITY SCREENING FORM ...... 231 B. IRB CONSENT FORMS ...... 232 C. NEIGHBORHOOD ENVIRONMENT WALKABILITY SCALE [ENGLISH] .. 239 D. NEIGHBORHOOD PHYSICAL ACTIVITY QUESTIONNAIRE [ENGLISH] . 244 E. NEIGHBORHOOD SELF-SELECTION [ENGLISH] ...... 252 F. MULTI-ETHNIC IDENTITY MEASURE – REVISED (MEIM-R) [ENGLISH] 253 G. SOCIODEMOGRAPHIC AND HEALTH QUESTIONS ...... 254 H. MAP OF DISTRIBUTION OF ASIANS LIVING IN PHILADELPHIA (CENSUS DATA: ARCMAP 10.7.1) ...... 255 I. SUMMARY OF STUDY FINDINGS ...... 256

FIGURES

Page

Figure 1-1. Neighborhood Environment Influences ...... 5 Figure 1-2. Odds of Physical Activity among Asian Subgroups. (Becerra et al., 2015) .. 10 Figure 1-3. Odds of Hypertension among Ethnic Groups by BMI (Wong et al., 2014)... 14 Figure 1-4. Associations between Perceived Neighborhood Walkability and Physical Activity ...... 17 Figure 1-5. Ethnic Identity Interactions with Perceived Neighborhood Walkability ...... 18 Figure 1-6. Associations between the Walk Score and Physical Activity ...... 18 Figure 1-7. Association of Proximity to Physical Activity Resources and Physical Activity ...... 19 Figure 2-1. SEM applied to physical activity behavior (Bauman et al., 2012) ...... 29 Figure 2-3. Measures of accessibility (L. E. Thornton et al., 2011) ...... 45 Figure 2-4. Park accessibility and levels of MVPA. (Schipperijn et al., 2017)...... 59 Figure 2-5. Rate of Park Amenities with Increasing Ethnic Minority Composition (Hughley et al., 2016) ...... 69 Figure 3-1. Philadelphia County Zip Codes (ArcMap 10.4) ...... 82 Figure 3-2. Breakdown of Physical Activity Domains based on the NPAQ...... 87 Figure 3-3. Sample Walk Score Result ...... 90 Figure 3-4. Histogram of Distribution of Outcomes ...... 98 Figure 4-1. Map of Geographic Distribution of Participant's by Zip Code ...... 107 Figure 4-2. IRRs of NEWS Subscales and Active Transportation ...... 119 Figure 4-4. IRRs of NEWS and Recreational Walking and Cycling ...... 122 Figure 4-5. IRR of NEWS and Physical Activity ...... 125 Figure 4-6. Predicted Values of Neighborhood Walkability and Active Transportation by Ethnic Identity ...... 131 Figure 4-7. Predicted Values of Neighborhood Walkability and Moderate Intensity Physical Activity by Ethnic Identity ...... 135

TABLES

Page

Table 3-1. Neighborhood Environment Walkability Scale...... 84 Table 3-2. Activities Converted into METs...... 86 Table 3-3. Walk Score Categories (Hirsch et al., 2013) ...... 90 Table 3-4. North American Industry Classification System Codes for Identifying Physical Activity Resources using InfoUSA ...... 91 Table 3-5. GIS Measures using the Find Nearest Tool ...... 96 Table 3-6. Frequency of Participants Reporting no Physical Activity ...... 98 Table 3-7. Study Variables ...... 99 Table 3-8. Variables by Research Question ...... 100 Table 4-1. Demographic Characteristics of Study Participants ...... 102 Table 4-2. Health Characteristics of Study Participants ...... 104 Table 4-3. Acculturative Characteristics of Study Participants ...... 105 Table 4-4. Geographic Distribution of Participants by Zip Code ...... 106 Table 4-5. Descriptive Built Environment Characteristics ...... 109 Table 4-6. GIS Measures of Park and Physical Activity Facility and Park Proximity ... 110 Table 4-7. Duration of Physical Activity among Participants and by Gender and Ethnic Group ...... 112 Table 4-8. Physical Activity Level of Study Participants in Metabolic Equivalents...... 113 Table 4-9. Correlation Matrix of NEWS Subscales (Spearman Rank) ...... 115 Table 4-10. Correlation Matrix of Built Environment Variables (Spearman Rank) ...... 115 Table 4-11. ZINB Regression Models for Perceived Neighborhood Walkability Subscales and Active Transportation...... 117 Table 4-12. IRRs of NEWS Subscales and Active Transportation ...... 119 Table 4-13. ZINB Regression Models for Perceived Neighborhood Walkability Subscales and Recreational Walking and Cycling ...... 120 Table 4-14. IRRs of NEWS Subscales and Recreational Walking and Cycling ...... 122 Table 4-15. IRRs of NEWS Subscales and Physical Activity ...... 125

xii

Table 4-16. ZINB Regression Models for Overall Perceived Neighborhood Walkability and Physical Activity Subtypes ...... 127 Table 4-17. Moderated ZINB Model of Ethnic Identity and Perceived Neighborhood Walkability in Predicting Active Transportation in Asian American Residents ...... 130 Table 4-18. Moderated ZINB Model of Ethnic Identity and Perceived Neighborhood Walkability in Predicting Recreational Walking/cycling in Asian American Residents 132 Table 4-19. Moderated ZINB Model of Ethnic Identity and Perceived Neighborhood Walkability in Predicting Duration of Vigorous Activity in Asian American Residents 133 Table 4-20. Moderated ZINB Model of Ethnic Identity and Perceived Neighborhood Walkability in Predicting Duration of Moderate Physical Activity in Asian Americans 134 Table 4-21. Moderated ZINB Model of Ethnic Identity and Perceived Neighborhood Walkability in Predicting Overall Physical Activity (METs) in Asian Americans ...... 136 Table 4-22. ZINB Regression Models for the Walk Score and Active Transportation .. 139 Table 4-23. ZINB regression models for the Walk Score and recreational walking/cycling ...... 141 Table 4-24. ZINB Regression Models for the Walk Score and METs of Physical Activity ...... 142 Table 4-25. ZINB Model for Driving Time to the Nearest Physical Activity Facilities and Parks and Association with Physical Activity ...... 147 Table 4-26. ZINB Model for Access to Nearest Drive Time to Physical Activity Facilities and Parks and Association with Overall Physical Activity ...... 149 Table 4-27. ZINB Model of Walking Time to the Nearest Physical Activity Facilities and Parks and Association with Physical Activity ...... 153 Table 4-28. ZINB Model for Nearest Walk Time to Physical Activity Facilities and Parks and Association with Overall Physical Activity ...... 155 Table 4-29. Built Environment Predictors of Park Use ...... 158 Table 4-30. Built Environment Predictors and Meeting the Physical Activity Guidelines ...... 158 Table 4-31. Ethnic Differences in the Relationship between Built Environment Factors and Physical Activity ...... 160

xiii

CHAPTER 1

INTRODUCTION

Physical activity levels are disproportionately low among adults living in the

United States (US), with only one in five adults meeting the Centers for Disease Control

(CDC) physical activity guidelines of 150 minutes of aerobic intensity or 75 minutes of vigorous intensity levels and two days of muscle strengthening activities per week (CDC,

2015). Reduced physical activity due to less time spent engaging in active travel and in leisure time activities have resulted in a rise in risk of developing chronic diseases (Lavie et al., 2019). Consequentially, the cost of physical inactivity on the US health care system is estimated at 131 billion dollars per year, representing approximately 12.5% of total health care expenditures (S. A. Carlson et al., 2015). A multitude of invariable and modifiable factors-sociodemographic, cognitive, cultural, psychosocial as well as aspects of the built environment likely contribute to the low rates of physical activity (Choi et al.,

2017; Maddison et al., 2009). More specifically, there is strong evidence indicating that the quality of the built environment infrastructure, which includes neighborhoods, sidewalks, aesthetics, crime, and physical activity facilities, have a substantial impact on promoting physical activity (Brownson et al., 2009; Ferdinand et al., 2012; Saelens,

Sallis, & Frank, 2003).

Ethnic minority groups in the US are less likely to meet the physical activity guidelines. For example, Asian Americans are less physically active than Non-Hispanic

Whites (Becerra et al., 2015; Kandula & Lauderdale, 2005; Kao et al., 2016; Yi et al.,

2015). A number of studies have investigated the variables related to low levels of physical activity among Asian Americans (Becerra et al., 2015; Kandula & Lauderdale,

2005; Kao et al., 2016; Yi et al., 2015). Several studies have, then explored how psychosocial determinants play a role in physical activity adoption and maintenance.

Fewer studies have focused on understanding built environment impacts. Furthermore, there is a paucity of research that have included objective measures of the built environment and their impact on physical activity (Bungum, Landers, Azzarelli, &

Moonie, 2012; Li, Kao, & Dinh, 2015; Patel et al., 2015).

The purpose of this dissertation was to examine the effects of built environmental factors on physical activity behavior among Asian Americans living in Philadelphia

County, Pennsylvania. This dissertation sheds light on addressing the relationship between the built environment and physical activity among an underrepresented population. Furthermore, study implications for future built environment interventions, from methodological considerations, and increasing education about maintaining a physically active lifestyle in one’s environment to potential improvements in the built environment structures surrounding neighborhoods that target ethnic minorities at low risk of physical activity engagement are addressed.

Background

Physical activity is an important lifestyle behavior known to improve overall health as well as provide substantial reductions in the risk of mental and physical chronic illnesses including obesity, cardiovascular disease, certain cancers, hypertension, stroke, depression, and anxiety (Blair, Goodyear, Gibbons, & Cooper, 1984; Brown, Winters-

Stone, Lee, & Schmitz, 2012; Lee, Folsom, & Blair, 2003; Taylor, Sallis, & Needle,

1985). There are vast benefits associated with being physically active, with an established dose-response relationship between physical activity and premature mortality and chronic

disease morbidity (Warburton & Bredin, 2017). Physical inactivity is estimated to contribute to a striking 8.3.% of total deaths in the United States among adults aged 25 years and older (S. A. Carlson et al., 2018).

The built environment is largely shaped by policies and laws/regulations in promoting long-term physical activity and improving overall public health among the US population through sustainable interventions (Sallis, Floyd, Rodríguez, & Saelens, 2012).

The neighborhood and built environment include how one’s surrounding environment, from access to healthy foods, having a safe place to be physically active, or exposure to crime and violence can impact their health (Marmot et al., 2008; WHO, 2018). Social determinants of health are known as the conditions in which individuals grow up, live, and work which affect their health and quality of life and targets upstream influences

(U.S. Department of Health and Human Services, 2014). The United States Department of Health and Human Services (USDDHS) identified five social determinants of health, and they include health and health care, education, economic stability, neighborhood and build environment, and social/community context (U.S. Department of Health and

Human Services, 2014). Strategies to improve physical activity in the population have been established nationally and through policy and governments affirmations as a form of health promotion and public health initiatives. The Centers for Disease Control and

Prevention (CDC) have established that the physical activity guidelines for American adults in order to exhibit health benefits include obtaining at least 150 minutes of moderate physical activity or 75 minutes of vigorous physical activity per week plus two days of strength exercises (CDC, 2015). Healthy People 2020 objectives include the goals of increasing the proportion of adults who engage in physical activity, but also to

improve the surrounding built environment through policies related to the community, street, and transportation/travel that would increase the availability of physical activity opportunities (U.S. Department of Health and Human Services, Office of Disease

Prevention and Health Promotion, 2014).

Built environmental predictors of physical activity are elucidated through the context of multilevel theories that explain health behaviors. The Social Ecological Model

(SEM) is a multilevel conceptual framework which postulates that individuals are part of a larger social system where multiple aspects of the physical, cultural, and social environment can impact one’s health (McLeroy, Bibeau, Steckler, & Glanz, 1988; Sallis,

Owen, & Fisher, 2008). Within the SEM, individual, interpersonal, neighborhood, social, and cultural environments have been identified as strong indicators of predicting health behaviors, especially physical activity (Brownson et al., 2009; Kaczynski, 2012; Wold &

Mittelmark, 2018). The interplay of multiple factors at all of the mentioned hierarchical levels contribute to the decision to engage in leisure-time physical activity (LTPA) as well as engaging in active transportation such as walking and cycling (Burbidge &

Goulias, 2009). The SEM is more comprehensive than traditional health behavior models when aiming to understand multilevel factors influencing physical activity.

Figure 1-1. Neighborhood Environment Influences on Transport and Recreation/Exercise (Saelens, Sallis, & Frank, 2003).

Historically, getting enough physical activity was far less an issue, as neighborhoods were built for high accessibility to places by walking, known as active transportation (Ewing et al., 2003; Koohsari et al., 2013). The introduction of cars and technology within the last 50 years has driven changes known as urban sprawling, defined by increased dispersion of a population into lower-density residential areas, increased separation of homes, shops and workplaces (Ewing et al., 2003).

Consequences associated with urban sprawling are poor accessibility, sedentary, and driving dependent lifestyles (Ewing et al., 2003). American Community Survey reports on commuting methods indicates the highest method is driving in personal vehicles

(76.4%), followed by carpooling (9.4%), public transportation (5.2%), walking (2.8%) and biking (0.6%) (McKenzie, 2015). In the last four decades, advances have been made in understanding built environmental influences on transport behavior and physical activity, especially active transportation such as walking and cycling, known as non- motorized forms of transport (Brownson et al., 2009; Ferdinand et al., 2012; Saelens,

Sallis, & Frank, 2003; Sallis et al., 2012). Studying the built environment has stemmed from multiple fields, including public health, physical activity and transportation and city planning, to examine how neighborhoods can be maximized to increase physical activity participation (Brownson et al., 2009; Ferdinand et al., 2012; Saelens, Sallis, & Frank,

2003; Sallis et al., 2012).

The built environment has differential effects on several domains of physical activity that extend beyond participating in LTPA (Sugiyama et al., 2009). Four different exercise domains were established and include: leisure/recreation/exercise, occupational, transportation, and household based (Pratt et al., 2004). After recognizing the importance of differentiating between physical activity domains, the Behavior Risk Factor

Surveillance Survey (BRFSS) began to include transportation, household, and occupational-related physical activity in their national surveys beginning in 2001 (Ewing,

2005). The built environment has the ability to impacts physical activity behavior in all of these settings.

Walking is the most common physical activity. (Boarnet et al., 2008; Nagel et al.,

2008). Neighborhood walkability is indicated by higher residential density, mix-land use

(a combination of residential and commercial areas), street connectivity, and increased availability of facilities for walking and biking (e.g., presence of bike lanes) (Ferdinand et al., 2012; Saelens, Sallis, Black, et al., 2003; Saelens, Sallis, & Frank, 2003; Zuniga-

Teran et al., 2017). Increased neighborhood walkability has been linked to increased number of trips made by walking and cycling, and overall increased levels of active transportation among residents living in those neighborhoods (Frank et al., 2010; Saelens,

Sallis, & Frank, 2003).

In addition to having walkable neighborhoods, the availability of physical activity resources such as parks and facilities within neighborhoods serve as opportunities to engage in leisure physical activity and potential routes of increasing active transportation

(Diez Roux et al., 2007; Kaczynski et al., 2014; Leggett et al., 2018; Parker et al., 2016).

Density of resources have shown to be positively associated with physical activity levels

(Diez Roux et al., 2007). For example, individuals who had greater than four exercise facilities within 1000 meters of their home were engaging in 5.4 more minutes of MVPA per day and were more likely to be meeting the physical activity guidelines (Eriksson et al., 2012). Green spaces, such as parks, additionally serve as important resources for engaging in physical activity and the diverse availability of facilities within parks can promote park use and park-based physical activity (Kaczynski et al., 2014; Kaczynski &

Havitz, 2009; Schipperijn et al., 2013; Wen et al., 2013). Recreational physical activity facilities located in one’s neighborhood can strongly encourage one to maintain a physically active lifestyle.

Disadvantaged communities with respect to having poor walkability features and poor access to recreational facilities presents an environmental justice issue especially when linked to socioeconomic and ethnic minority status (Taylor, Poston, Jones, & Kraft,

2006). Unjust environmental conditions have been evident in neighborhoods where predominantly ethnic minorities reside, specifically African Americans and

Hispanic/Latino Americans, and low-income individuals, presenting more challenges to overcome in order to live a healthy lifestyle (Conroy et al., 2017; Giles-Corti & Donovan,

2002b; Sallis et al., 2009, 2011; Taylor et al., 2006). There are few studies investigating whether there are disparities in physical activity among Asian Americans that are due to this population residing in low walkable neighborhoods or a built environment not conducive to engaging in regular physical activity (Bungum et al., 2012; Kelley,

Kandula, Kanaya, & Yen, 2016; Li et al., 2015). Even less is known about how cultural factors may play a role in affecting this relationship, and whether Asian Americans who reside in neighborhoods that are ethnically dense exhibit different physical activity behaviors. The current study incorporated cultural related measures such as ethnic identity as important variables to consider when investigating this relationship.

Need for Study

There are approximately 20.4 million Asian Americans in the US (López et al.,

2017). Asian Americans represent the fastest growing population of individuals in the

US, experiencing a 72% population increase between 2000-2015 (López et al., 2017).

Despite this population growth and presence, they remain an underrepresented ethnic minority, exhibiting higher rates of chronic disease and facing disparities in engaging in

health behaviors. (Becerra et al., 2015; Kandula & Lauderdale, 2005; Palaniappan et al.,

2010; Yi et al., 2015).

A general misconception historically established is that Asian Americans have been portrayed as the ‘model-minority’ (Islam et al., 2010). This includes misconceptions that all Asian Americans belong to a higher socioeconomic status, undergo upward mobility and have high levels of education (Islam et al., 2010). However, Asian

Americans are highly diverse, representing several ethnicities, socioeconomic statuses, cultures, customs, and immigration histories. Regrettably, several research studies on

Asian Americans have aggregated them into one group, masking the differences in chronic disease morbidity and engagement in health behaviors (Islam et al., 2010). It is critical to investigate health and health behaviors across subgroups.

Ethnic minorities are less likely to be physically active, and Asian Americans are less physically active than Non-Hispanic Whites (Becerra et al., 2015; Kandula &

Lauderdale, 2005; Kao et al., 2016;Yi et al., 2015). According to the National Health

Interview Survey in 2012, 17.1% of Asian adults in the United States met the CDC physical activity guidelines of 150 moderate physical activity or 75 minutes of vigorous physical activity and two days of muscle strengthening activities, compared to 23% of non-Hispanic white adults per week (Blackwell et al., 2014). In a population based study on physical activity determinants in Asian American subgroups using the California

Health Interview Survey (CHIS), Koreans and Vietnamese Americans living above the poverty level, and Chinese adults of younger age were less likely to meet the American

College of Sport Medicine physical activity guidelines (Becerra et al., 2015).

Additionally, Korean and Vietnamese adults were less likely to meet the guidelines than

Chinese adults (Becerra et al., 2015). However, this study only compared the physical activity levels of Asian subgroups without comparing them to other ethnic groups (e.g., non-Hispanic Whites). Figure 2 illustrates the moderating effect of age on physical activity levels among different Asian ethnic groups.

2.5

2.03 2

1.5 1.25

0.97 0.99

Odds Ratio Odds 1 0.69 0.59 0.5

0 Chinese Filipino South Asian Japanese Korean Vietnamese Subgroup (Ref. Group=>45)

Figure 1-2. Odds of Physical Activity among Asian Subgroups. (Becerra et al., 2015)

Among this population, most research has explored mainly how psychosocial factors play a role in physical activity adoption and maintenance without considering impact of built environment structures (Becerra et al., 2015). Among ethnic minority populations, and specifically Asian Americans, physical activity determinants impacting physical activity have included socioeconomic status, acculturation, gender, education level, and perceived environmental barriers and facilitators (Becerra et al., 2015).

Knowledge about physical activity patterns of Asian Americans, specifically time spent

engaging in active transportation versus recreational physical activity, is lacking. Having destinations (e.g., library, shops, cafés) within one’s neighborhood is closely linked to higher amount of active transportation, less is known regarding how built environment feature can potentially promote recreational physical activity among this population

(Kelley et al., 2016). There is a paucity of research conducted on assessing objective measures of the built environment and their impact on physical activity among Asian

Americans (Bungum et al., 2012; Li et al., 2015; Patel et al., 2015). Active living research examining built environmental impacts on physical activity in the general population have only included small study samples on Asian Americans (Wen et al.,

2007).

Nonetheless, built environment studies have suggested that a good majority of exercise and recreational related physical activity occurs in neighborhood streets near one’s home and serve as a crucial resource for regular physical activity participation

(Giles-Corti & Donovan, 2002b; Huston et al., 2003; Sugiyama et al., 2009). For example, Sugiyama and Colleagues (2009) examined neighborhood street use and found that most recreational based physical activity occurred near one’s home but was highly influenced by one’s perception of their neighborhood being aesthetic. Furthermore, having outdoor recreational facilities such as parks and green spaces as well as stores, can further promote the use of neighborhood streets for physical activity for activities such as walking, biking, jogging, and running (Giles-Corti et al., 2005; Hoehner et al., 2005;

Huston et al., 2003; Sugiyama et al., 2009). Few studies have examined how both parks and recreational facilities affect physical activity levels, a more comprehensive measure

of resource availability (Jones et al., 2015). These relationships will be further investigated in the proposed study.

The relationship between the built environment and overall physical activity have been investigated in a number of studies (de Sa & Ardern, 2014; Ferdinand et al., 2012;

Kärmeniemi et al., 2018; McCormack & Shiell, 2011). However, most studies fail to classify built environmental effects by physical activity domain and make conclusions based on overall general physical activity level (Giles-Corti et al., 2006). Few studies have examined whether walking, cycling, and physical activity occur within the context of one’s specific neighborhood. The present study will aim to contribute to this literature gap by investigating neighborhood specific physical activity.

Health Status of Asian Americans

Chronic diseases such as hypertension, diabetes, overweight/obesity, hyperlipidemia and cancer, disproportionately affect ethnic minorities in the US

(Commodore-Mensah et al., 2016, 2018; Staimez et al., 2013). The most representative sample of Asian residents at the national level is the National Health Interview Survey

(NHIS), which includes disaggregated Asian groups. Other studies have highlighted disparities also being measured in regional, state, and community based studies but in the past has largely been masked due to aggregating Asian Americans into one group, and due to small sample sizes (Holland & Palaniappan, 2012; Srinivasan & Guillermo, 2000).

Among Chinese, Korean and Vietnamese male and female adults, cancer and malignant neoplasms were reported to be the leading cause of mortality, followed by diseases of the heart (Hastings et al., 2015). Rates of overweight/obesity were reported to range from 15-

28% among Chinese for overweight and 2-27% for obesity, 38.4% of overweight and

7.5% obesity among Koreans, and 17.1% among overweight and 2.1% among

Vietnamese adults (Staimez et al., 2013). Furthermore, rates of diabetes reported among these ethnic groups, 2.2-28% among Chinese, 18.1% among Koreans and 5.3-15.6% among Vietnamese (Staimez et al., 2013).

Although the general US population faces high chronic disease burden due to overweight and obesity, Asian Americans are found to be at a higher risk of chronic conditions including metabolic disease, hypertension and diabetes with lower BMI values compared to African Americans, Hispanics, and non-Hispanic Whites (Lee, Brancati, &

Yeh, 2011; Wong, Chou, Sinha, Kamal, & Ahmed, 2014). Wong and colleagues (2014) found that each one-point increase in BMI was associated with a higher risk of hypertension, wish Asians experiencing a much higher risk with lower levels of BMI.

Lower BMI cut-offs have been developed specifically for Asian Americans to capture the true impact of overweight/obesity on chronic disease due to higher percentage body fat present in this ethnic group, as the obesity risk is underestimated with normal BMI values

(Jih et al., 2014; Staimez et al., 2013; WHO Expert Consultation, 2004). Moreover, Asian

Americans exhibit differing body fat distributions, such higher central adiposity that is strongly correlated with chronic disease (Liu et al., 2017; Wong et al., 2014).

Figure 1-3. Odds of Hypertension among Ethnic Groups by BMI (Wong et al., 2014)

Immigration among Asian Americans

Asian Americans have unique immigrant experiences with recent calls to action to address health among this group due to heterogeneity in immigration histories and cultures experienced (Palaniappan et al., 2010). With respect to history of Asian immigration, the liberalized policy period between 1965 to currently has brought

Chinese, Korean and Vietnamese groups to the US in large numbers (Trinh-Shevrin et al.,

2009). Chinese immigrants arrived with higher social disadvantage than Korean and

Vietnamese immigrants, having high occupational and economic standing (Trinh-Shevrin et al., 2009). During this time period, Vietnamese have resettled to due to US political responsibility and consequences of civil and socioeconomic hardships within their country, thus mainly arriving under refugee status (Trinh-Shevrin et al., 2009). Varying immigration histories accompany distinct socioeconomic adjustment and migration

experiences among these ethnic groups, which directly influence mental and physical health.

The ‘Immigrant paradox’ or ‘healthy immigrant effect’ is characterized as new immigrants exhibiting better health outcomes than US born counterparts that degrade the longer their residence in the US due to the change in environment (Commodore-Mensah et al., 2016; McDonald & Kennedy, 2004; Osypuk & Alonso, 2015; Sohail et al., 2015).

NHIS data from 2010-2016 found that Southeast Asian immigrants, including individuals from Thailand, Vietnam and the Philippines reported the highest risk of hypertension and second highest prevalence of diabetes mellitus compared to all other immigrants

(Commodore-Mensah et al., 2018). Increased acculturation and barriers to care among immigrants especially are known to be associated with the development of poor lifestyle behaviors and chronic disease development in Asian Americans and stand as a rationale for developing culturally relevant interventions among different Asian subgroups.

Community-based studies in Philadelphia on Asian Americans, specifically Chinese,

Korean and Vietnamese adults, have reported rates of overweight to be 65% and obesity to be 71% with random intermediate or poor blood glucose levels, and 34.25% with those who were pre-hypertensive or hypertensive (Patterson et al., 2016).

Statement of the Problem

The purpose of this study was to determine whether objective and perceived built environmental characteristics, specifically perceived neighborhood walkability, the Walk

Score, and physical activity resources are associated with neighborhood physical activity levels among Asian Americans living in an urban environment. This study addressed psychological, sociodemographic, cultural, and built environment characteristics and

physical activity behavior stemming from the SEM. This study addressed whether the built environment affects Asian Americans’ participation in physical activity, from active transportation to leisure physical activity behaviors based on their surrounding urban neighborhood, while taking socioeconomic status, culture, and other important variables into consideration. This study contributes to the overall literature of the few existing studies examining the impact of the built environment on physical activity levels among

Asian Americans using objective measures and geographic information systems.

This study is unique in comparison to existing studies as it will examine the built environment and physical activity in Asian American populations. The study sample includes residents from urban neighborhoods and mainly first generation immigrants.

Existing studies on perceptions of the built environment have only assessed incivilities, crime, aesthetics, diversity of destinations in the neighborhood, and objective neighborhood walkability measured using the Walk Score (Bungum et al., 2012; Hirsch et al., 2013; Kao et al., 2016; Kelley et al., 2016; Y. Li et al., 2015). However, the current study adds to the literature by examining a more comprehensive measure of built environment perceptions, such as infrastructure for walking/cycling, aesthetics, residential density, street connectivity, and mixed land use, which have not been routinely included in previous studies. Furthermore, no existing studies have examined proximity to physical activity facilities using GIS. This study also will incorporate the influence of ethnic identity with the built environment and its impact on physical activity behaviors in a novel population.

Research Questions and Objectives

The questions and objectives examined in this study were:

Research Question 1: Do Asian residents with higher perceived neighborhood walkability report higher levels of active transportation, recreational walking/cycling, and overall physical activity?

Hypothesis 1: Higher perceived neighborhood walkability will be associated with increased active transportation, recreational walking, and leisure physical activity levels.

Control Variables

Perceived Neighborhood Physical Activity1 Walkability

Figure 1-4. Associations between Perceived Neighborhood Walkability and Physical Activity

1Separate models for active transportation, recreational walking/cycling, and overall physical activity within and outside of the neighborhood

Research Question 2: Does level of ethnic identity impact the relationship between perceived neighborhood walkability and active transportation, recreational walking/cycling, and overall physical activity?

Hypothesis 2: High ethnic identity and high perceived neighborhood walkability will have the strongest impact on physical activity levels

Control Variables

Perceived Neighborhood Physical Activity1 Walkability

Ethnic Identity

Figure 1-5. Ethnic Identity Interactions with Perceived Neighborhood Walkability

1Separate models for active transportation, recreational walking/cycling, moderate activity, vigorous activity, and overall physical activity

Research Question 3: Do Asian residents living in neighborhoods of greater walkability measured objectively report higher levels of active transportation, recreational walking/cycling, MVPA, and overall physical activity?

Hypothesis 3: Higher Walk Scores will be positively associated with increased active transportation, recreational walking/cycling, and overall physical activity levels.

Control Variables

Objective Neighborhood Physical Activity1 Walkability s

Figure 1-6. Associations between the Walk Score and Physical Activity

1Separate models for active transportation, MVPA, and overall physical activity within and outside of the neighborhood. 18

Research Question 4: Do Asian residents living in neighborhoods with higher accessibility to physical activity resources report higher physical activity levels, specifically walking in the neighborhood, active transportation, MVPA, and overall physical activity?

Hypothesis 4: Increased proximity physical activity resources and parks will be associated with higher physical activity levels.

Control Variables

Proximity to

Physical Activity Physical Activity1 Resources

Figure 1-7. Association of Proximity to Physical Activity Resources and Physical Activity

1Separate models for walking in the neighborhood, active transportation, MVPA, and overall physical activity within and outside of the neighborhood.

Delimitations

The following delimitations were present in this study:

1. This study did not conduct a comprehensive environment assessment using area based sampling in specified neighborhoods that traditional built environment studies use as a main approach with strict sampling procedures. Asian Americans are generally dispersed among numerous Philadelphia neighborhoods, with around 30 out of 384 census tracts having an ethnic density of Asian Americans at 20% or above (United States Census

Bureau, 2015). Due to study restrictions in being able to capture a representative sample

of Asian Americans within these neighborhoods, the study used convenience sampling to recruit participants to make comparisons between their perceptions of their built environment and determine how their immediate surrounding environment affects their physical activity levels. Participants were recruited through identified Community Based

Organizations in Philadelphia County that are representative of Chinese, Korean, and

Vietnamese residents.

2. This study recruited adults in the age range of 20 to 69 years. Those under 20 were not recruited as their behaviors may differ and be affected by their parent’s behaviors, beliefs, and perceptions, thus introducing parental influences on behavior. In addition, those over 65 years old represent a unique population due to having different lifestyles

(e.g., retired) and/or having conditions (e.g., cardiovascular disease, cognitive impairment), which can potentially impact their ability to be physically active. Previous studies on older Asian Americans have found them to have higher physical activity levels compared to younger adults (August & Sorkin, 2011; Becerra et al., 2015; Kao et al.,

2016; Y. Li et al., 2015)

3. The study was conducted among Asian Americans, thus having to take into

Consideration associated cultural beliefs and values that impact health. Level of ethnic identity and acculturation are both known to influence health behaviors and overall health. These variables will be used as important variables when observing the relationship between the built environment and physical activity among this population.

4. Although numerous Asian countries comprise the Asian American ethnicity, this study included only individuals of Chinese, Korean, and Vietnamese ethnicity. These ethnic groups are the most prevalent in the Philadelphia region, and present unique physical

activity participation that introduces heterogeneity of Asian American subgroups.

Furthermore, these groups are within the top five largest Asian subgroups in the United

States (Lopez et al., 2017).

5. Data was only collected from one member of a household to capture more geospatial areas among participants. This inclusion criteria was addressed during the data collection process.

Limitations

This study was limited in the following ways:

1. There are several factors that can affect the consistency of the built environment. If a specific region where a participant lives is undergoing major environmental changes, this can significantly affect the participants’ perceived environment and influence the study findings. The Walk Score and physical activity resource proximity measures may not include updates to these environmental changes occurring.

2. There is evidence to suggest that weather and climate impact physical activity levels.

Due to the nature of when the data collection for measuring physical activity will take place, in the summer and fall seasons of 2019, this may impact participants’ physical activity levels, which can be higher than averaged out in a whole year.

3. Physical activity levels will be measured by self-report, which has been shown to lead to response bias based on social desirability and potential overall reported levels (Adams et al., 2005).

4. Although objective measures are being used to measure physical activity

resources and built environment walkability, measurement error can occur by the researcher when conducting the online environmental audit, including collecting Walk

Scores, and information about physical activity resources.

5. The study is a cross-sectional design, which will not allow for causal inferences between the built environment and impact on physical activity levels.

6. Data sources may come from different years. Participants’ survey responses, walk scores, and physical activity resource density were collected in 2019. Shapefiles may be provided from previous years.

7. Some measures that contribute to built environment walkability measured using GIS will not be measured in this study. Mixed-land use and street connectivity are important predictors of physical activity. Additionally, microscale features such as sidewalk quality, which are known predictors of physical activity, were not examined.

8. The Walk Score is limited in that information about pedestrian infrastructure and quality, aesthetics, and traffic are not incorporated into this measure, and are important indicators of built environment as well. Thus, it does not provide a comprehensive measurement of total neighborhood walkability.

9. Since the Walk Score is a global measure of neighborhood walkability that encompasses mainly accessibility to a wide range of amenities, the study findings cannot determine specifically which built environmental feature has a significant impact on physical activity levels in the study sample.

10. Geocoding was utilized to match raw address information and physical activity resources in order to measures resource proximity. Several errors in this process can occur, such as: 1) unsuccessful geocoding of addresses due to inaccurate data, 2)

geocoding to the incorrect location due to inaccurate spatial digital files or addresses, and

3) addresses that have not been updated.

11. GIS Buffers were identified to estimate proximity and density of physical activity facilities and can be conceptualized differently by individuals with respect to what they perceive as within their environment.

12. The generalizability of data may be limited as the study sample includes a few hundred Asian participants, which represent Philadelphia County. Thus, out of the total population of Asian residents in Philadelphia County, this study only obtained a small sample.

13. Self-selection bias may have occurred due to the nature of the study and that it is a voluntary survey. Information bias, specifically self-reported recall bias may have occurred since it required individuals to recall their physical activity behavior in the past week. More validated measures such as accelerometers have been used to measure physical activity but are high costing and as a result, not feasible for this study.

Definition of Terms

Acculturation: a process that takes place because of contact with different people, groups, and social influences, mainly studied among individuals from countries different from where they were born or grew up in.

Active transportation: walking or cycling associated with travelling to a specific destination.

Asian American: a racial category used in government classification to identify individuals of Asian descent. Although numerous Asian countries comprise the Asian

American group, this study will include only individuals of Chinese, Korean, and

Vietnamese ethnicity. These ethnic groups are the most prevalent in the Philadelphia region.

Built environment: all of the physical parts of where we live and work (e.g., homes, buildings, streets, open spaces, and infrastructure) (CDC, 2018).

Census data: information about the members of a given population.

Connectivity: directness of travel between two points that is related to street design

(Saelens, Sallis, & Frank, 2003).

Covariates/control variables: predictive or explanatory variable of the dependent variable.

Demographics: statistical data of a population.

Ethnic density: residential concentration of a racial/ethnic minority group.

Ethnic identity: a measure of self-concept, characterized as how individuals strongly associate themselves with the behaviors, knowledge, and awareness of cultural beliefs and traditions of one’s ethnic group (Iwamoto & Liu, 2010).

Exercise: A type of physical activity that involves planned, structured, and repetitive bodily movement in order to maintain or improve one or more components of physical fitness.

Geographic Information Systems (GIS): computer-based systems designed to integrate different types of spatial and attribute information. Data relevant to physical activity can include topography, existing land uses, geological features (e.g., hills), infrastructure systems (streets, mass transit), recreation facilities, and residences. Each spatial feature can be linked with attributes (e.g., size, shape, and amenities for a park and demographic variables of a person living near the park) (Saelens, Sallis, & Frank, 2003).

Incivilities: elements that would reduce the pleasure associated with using a certain physical activity resource (Lee et al., 2005).

Land use mix: level of integration within a given area of different types of uses for physical space, including residential, office, retail/commercial, and public space. Land use is controlled by zoning ordinances that reflect political decisions most often made at the local level (Saelens, Sallis, & Frank, 2003).

Moderate to vigorous physical activity (MVPA): engaging in non-walking/cycling activities associated for recreation, health, and fitness purposes

Leisure-time physical activity (LTPA): activities that individuals engage in during free time, which are not work oriented or do not involve life maintenance tasks such as house cleaning or sleeping (Pratt et al., 2004).

Neighborhood: The environment surrounding one’s residential address.

Neighborhood walkability: a composite measure comprised of proximity to facilities, intersection density and block length as measured by the Walk Score.

Neighborhood self-selection: bias that results from health behaviors predicting neighborhood choice or one’s decision to live in a certain neighborhood (James et al.,

2015).

Perceived neighborhood walkability: how one perceives different aspects of one’s surrounding environment to be walkable based on several characteristics (e.g., proximity to services, land-use mix, crime, and graffiti).

Physical activity: any movement of the body that requires energy expenditure, including walking, cycling, and other moderate to vigorous intensity activities

Physical activity guidelines: obtaining 150 minutes of moderate intensity physical activity or 75 minutes of vigorous intensity physical activity or a combination of the two

(aerobic), and two days of muscle strengthening exercises per week (CDC, 2015).

Physical activity resource/facility: settings used for engaging in physical activity (e.g., parks, fitness club, courts, fields).

Proximity: The straight-line distance between different land uses such as residential, office, retail, and commercial activities (Saelens, Sallis, & Frank, 2003).

Recreational walking/cycling: engaging in walking and/or cycling for the purposes of recreation, health, or fitness within and outside of one’s neighborhood

Socioeconomic status: a composite measure of economic and sociological standing.

Utilitarian walking: walking as a form of transport in order to reach destinations.

CHAPTER 2

REVIEW OF LITERATURE

Introduction

This literature review provides a background for the research study, including a review of the SEM applied to physical activity. Psychosocial and cultural determinants of physical activity including behavioral theory, psychological effects of green exercise, and ethnic identity are discussed in detail. Physical activity domains and measures are reviewed. Perceived and objective built environment measures and built environment components will be discussed further in depth, including urban sprawl, neighborhood walkability, neighborhood parks, green spaces, and recreational facilities. Environmental justice issues are discussed with respect to how the built environment affects physical activity among ethnic and low-income minority groups in the US. Lastly, a review of physical inactivity predictors, health disparities in physical activity and chronic disease among Asian Americans are presented.

The Social Ecological Model and Physical Activity Promotion

The theoretical model often used to address the etiology of complex behaviors is the Social Ecological Model (SEM) (McLeroy et al., 1988). The SEM was introduced by

McLeroy and Colleagues (1988) to elucidate microlevel to macrolevel factors which influence disease and/or behavior in order to promote health (Golden & Earp, 2012;

McLeroy et al., 1988). The multiple levels of the SEM proposed initially include intrapersonal, interpersonal, institutional, community, and policy (Golden & Earp, 2012;

McLeroy et al., 1988). An adapted SEM was developed to explain correlates of physical

activity behavior among children, adolescents, and adults (Bauman et al., 2012; Stokols,

1992). These factors included individual, interpersonal, environment, policy, and global determinants illustrated in Figure 2-1 (Bauman et al., 2012; Stokols, 1992).

Figure 2-1. SEM applied to physical activity behavior (Bauman et al., 2012)

Within this framework, genetic, physiological, psychosocial and behavioral factors serve as individual/intrapersonal level factors that influence physical activity behavior (Bauman et al., 2012; Stokols, 1992). Among adults, high self-efficacy and perceived overall health were significant individual factors associated with engaging in more physical activity (Bauman et al., 2012). Factors such as attitudes, beliefs, values, behaviors, and knowledge have been highly suggested to influence physical activity

(Bauman et al., 2012). Furthermore in high-income countries, those of male gender, higher socioeconomic status, a younger age, and white ethnicity were more likely to engage in regular physical activity (Bauman et al., 2012). In terms of biological factors,

twin and family based studies demonstrate that physical activity engagement is highly heritable, specifically engaging in moderate and leisure activity levels (Moore-Harrison

& Lightfoot, 2010). Specific mechanisms leading to genetic influences on physical activity are related to brain structures that control personality, mood, and genetic predispositions to exercise responsiveness and weight loss (Stubbe et al., 2006). Although these individual level characteristics are important physical activity determinants, increasing knowledge and awareness about benefits are shown to be ineffective components of interventions targeted to increase physical activity and should consider environmental components (Bornstein & Davis, 2014).

At the interpersonal level, factors related to one’s social contacts, including family members, friends, partners/spouses, and co-workers can impact health behaviors, including physical activity participation (Bauman et al., 2012). Environmental factors that influence physical activity relate to social, built, and natural features are important and include: design features, neighborhood walkability, transit accessibility, parks, recreational facilities, and safety (Bauman et al., 2012). Natural environment includes presence of green space, vegetation, weather/climate considerations, and presence of natural physical activity spaces such as parks and trails (Bauman et al., 2012). Policy level includes the government policies and legislations surrounding that health behavior or disease at the local, regional and national levels (Bauman et al., 2012). Such policies aid to regulate resource investments and regulations to enhance physical activity participation and are typically addressed within multiple sectors (Bauman et al., 2012).

For example, The National Physical Activity Plan was developed to address land use, community design, and transportation in the context of improving access to physical

activity (Bornstein & Davis, 2014). Although macrolevel factors can have population wide effects, these are more difficult to measure and require extensive complex analyses.

Lastly, global patterns of change occurring within the past few decades such as urbanization and economic growth have shifted lifestyle behaviors including reductions in physical activity and increase in chronic diseases (Bornstein & Davis, 2014).

The SEM was developed to promote health and well-being by focusing on dynamic interactions between personal and environmental level factors instead of focusing on only one aspect (Stokols, 1992). Thus, theories that only target one level of effect, mainly intrapersonal, only explain a fraction of physical activity behavior. The proposed study heavily emphasizes how individual level factors such as demographics, ethnicity and behavior as well as the surrounding built environment impact physical activity behavior among Asian Americans living in an urban environment.

Psychosocial/Cultural Mediators and Mechanisms of the Built Environment on Physical Activity

Psychosocial factors are strong predictors of physical activity behavior, with few studies providing support for its interactive effects with the built environment and have mainly constituted mixed findings (Rhodes et al., 2018). The Theory of Planned Behavior

(TPB) constructs have been shown to impact the relationship between the built environment and physical activity levels among adults and youth (Maddison et al., 2009;

Rhodes et al., 2007). Intention and perceived behavioral control are a proximal determinant of physical activity (Maddison et al., 2009). A systematic review demonstrated overall positive interactive effects between social cognitive factors and built environment (Maddison et al., 2009). Specifically, having higher subjective or

objective accessibility to recreation facilities and higher exercise intention and self- efficacy predicted higher LTPA (Maddison et al., 2009). For other social cognitive factors, there were no interactive effects between accessibility of resources and attitudes, perceived barriers, subjective norms, and social support to explain LTPA (Maddison et al., 2009).

More specifically, one’s perception of the built environment is potentially explained by social cognitive factors that ultimately predict physical activity (Rhodes et al., 2018). The interactive effects between the built environment and physical activity may be two-fold; individuals with high versus low motivation to exercise may engage in more exercise because they live in a highly walkable neighborhood, or individuals who have lower levels of motivation to exercise may benefit more greatly from living in a favorable built environment versus those who do not, as they will face fewer environmental barriers. Figure 9 describes these two hypotheses developed in a systematic review by Rhodes and colleagues (Rhodes et al., 2018).

When testing these hypotheses, different results were found based on physical activity domain; interactions between social cognitive factors and the built environment failed to explain total physical activity, moderate-to-vigorous physical activity and transportation related physical activity (Rhodes et al., 2018). LTPA was predicted by both environmental accessibility through a positive interaction with physical activity intention and environmental aesthetics through a negative interaction with affective judgments (Rhodes et al., 2018).

In a multiethnic diverse sample of adults in six urban cities, walking behavior of adults was influenced by attitudes towards walking and perceived neighborhood

environment and neighborhood walkability (Yang & Diez-Roux, 2017). More specifically, overall minutes spent walking in a week was influenced by both positive attitudes towards walking and having greater neighborhood walkability measured by the

Walk Score and overall perceptions of pleasantness, convenience, and safety in one’s environment (Yang & Diez-Roux, 2017). Thus, strong interplays are existent between psychosocial predictors in explaining the built environment impact on physical activity.

Psychological and Cognitive Markers associated with ‘Green Exercise’

‘Green exercise’ is defined as physical activity which takes place in a natural environment and constitutes both urban and natural green spaces (Flowers et al., 2017).

Experimental studies in the area of ‘green exercise’ have illustrated that combining physical activity with nature has several psychological and physiological benefits as well as the ability to enhance the level of participation (Akers et al., 2012; Barton & Pretty,

2010). Extensive research on the impact of green exercise on mental health has found immediate psychological impacts on self-esteem and mood when controlling for exercise intensity, duration, and sociodemographics including age and gender (Akers et al., 2012;

Barton & Pretty, 2010; Flowers et al., 2017; Gladwell et al., 2013; Wooller et al., 2016).

Greenspace typologies in which benefits were most salient included ones that were on a waterside, and were equally beneficial among urban spaces, countrysides, forest/woodlands, and wild habitats (Barton & Pretty, 2010).

To understand potential cognitive mechanisms to explain this relationship and the visual sensation associated with green exercise, simulated studies comparing green, gray and red conditions while exercising revealed that the green condition that mimicked a green exercise environment resulted in lower perceived exertion, lower anger, and lower

total mood disturbance experienced among participants (Akers et al., 2012). Furthermore, the impact of outdoor versus laboratory environments on psychological outcomes were assessed in a sample of 35 women who walked for 10 minutes on a treadmill indoors versus an outdoor environment. When comparing these two conditions, the outdoor environment was associated with more positive affect, more enjoyment, and higher intentions to exercise in the future (Focht, 2009). Thus, outdoor walking appears to facilitate individuals to be more physically active in their daily lives (Flowers et al., 2017;

Focht, 2009).

The Theory of Planned Behavior (TPB) has been used as a theoretical construct to determine psychosocial factors that measure knowledge, attitudes, behaviors and intentions addressed within the TPB towards green exercise (Flowers et al., 2017). Three self-reported instruments were developed by researchers at the University of Essex, including the Indirect Beliefs About Green Exercise questionnaire (BAGE-ID), the

Beliefs about Green Exercise Questionnaire (BAGE), and the Intention to Perform Green

Exercise Questionnaire (INT-GE), showing high validity (Flowers et al., 2017). From these measures, several advantages reported with participating in green exercise that included fresh air, general positive affect, having a change of scenery, openness/freedom, and affordable cost/resources (Flowers et al., 2017) .

Ethnic Identity

Due to the presence of wide cultural diversity in the US, one’s identity to their own cultural values and beliefs versus those of American identities have been an important consideration for understanding health behaviors and outcomes (Brown et al.,

2014; Yancey, Siegel, & McDaniel, 2002). Ethnic identity is a socially constructed

concept defined as the extent to which an individual is attached to their own racial/ethnic groups’ values, cultures, language, and traditions, reflected in their self-concept and is internalized (Charmaraman & Grossman, 2010; Phinney & Ong, 2007). In the psychology literature, ethnic identity has been used to try and understand where there are health disparities in engaging in health behaviors as well as increased chronic disease risk. Early studies focused on ethnic identity among adolescents and health/well-being (J.

W. Berry et al., 2006; Charmaraman & Grossman, 2010; Yancey et al., 2002). Ethnic identity is distinct from an ethnic related concept known as acculturation, which is defined as the process of change that occurs with adapting to a new environment (Liem et al., 2000). The main difference lies in the process of each, where in acculturation, individuals adopt behaviors associated with the host country, whereas ethnic identity is an internalization that occurs (Phinney & Ong, 2007).

A few studies have looked at the relationship between ethnic identity and physical activity behavior or the mediating effect of ethnic identity on physical activity behavior

(Brown et al., 2016; Smalley, Warren, McClendon, Peacock, & Caro, 2016). Among a culturally diverse study of pregnant women with gestational diabetes, higher ethnic identity was positively associated with physical activity behavior, with a one-unit increase in score leading to an 11% increase in total amount of activity and increase in walking among the study sample (Brown et al., 2016). Furthermore, in a study among

African American rural women, stronger ethnic identity was associated with a higher likelihood that women were in a later stage of change in exercise behavior based on the

Transtheoretical Model, thus exhibiting higher motivation to exercise (Smalley et al.,

2016). Another reason that ethnic identity is related to physical activity can be that

individuals perception of engaging in health behaviors such as physical activity and perceiving it as something that is more associated with predominantly white middle class individuals (Oyserman et al., 2007; Smalley et al., 2016). This addressed the need for incorporating culturally tailored interventions in order to address culturally specific barriers that are experienced by ethnic minority groups in being physically active

(Smalley et al., 2016).

Some of the notable measures used to characterize level of ethnic identity include the psychological scales and qualitative studies conducted in ethnic groups (Iwamoto &

Liu, 2010; Yancey et al., 2002). The Multigroup ethnic identity measure (MEIM) is comprised of 12 items that assess commitment or sense of belonging and exploration

(Phinney, 1992). Exploration refers to learning about one’s ethnic group and their participating in ethnic practices (Phinney, 1992). A revised version of the MEIM was developed and tested on a diverse sample of university students from a predominantly minority serving institution and included the same two factor model with only six items, exploration and commitment (Phinney & Ong, 2007). Confirmatory factor analyses indicated that this model was a strong fit with Cronbach’s alphas of 0.76 for the exploration factor and 0.78 for the commitment factor with a combined alpha of 0.81

(Phinney & Ong, 2007).

Physical Activity Domains and Measures

Physical activity is multidimensional construct measured by various factors including intensity, duration, frequency, and activity types (Sylvia et al., 2014). Activity types established include leisure, home-based, occupational, transportation, and structured physical activity (Pratt et al., 2004; Sylvia et al., 2014). Existing methods for

measuring physical activity includes the use of self-report questionnaires, activity diaries/logs, direct observations by observers, and devices such as accelerometers, pedometers, and armbands (Sylvia et al., 2014).

Active transportation is defined in the urban planning and transportation literature as behaviors associated with walking or biking to destinations or to participate in an activity rather than for the mere reason of travelling itself (Handy et al., 2002).

Within other types of activity, utilitarian walking is defined as walking for transport in order to reach destinations (Handy et al., 2002; Sugiyama, Neuhaus, Cole, Giles-Corti, &

Owen, 2012; Yang, 2016). Travel behavior models emphasize dynamic factors such as personal decision making, variables related to the built environment, sociodemographics

(including vehicle ownership), and alternative transportation modes with travel time

(Yang, 2016). Travel diary surveys are a common method of collecting data on trips taken by walking (Handy et al., 2002).

The environment can act to facilitate and inhibit physical activity behavior and having access to various physical activity facilities are important for certain types of activity as well as meeting the weekly recommendations (Heinrich et al., 2017). As expected, meeting physical activity guidelines through engaging in LTPA to meet the recommended levels of moderate and/or vigorous intensity activity is statistically significantly associated with having access to a regular facility (Cerin et al., 2008;

Eriksson et al., 2012; Heinrich et al., 2017). Furthermore, vigorous intensity activity has been strongly correlated to exercising in a specialized physical activity facility involving equipment such as a gym/fitness club rather than reporting exercising in a park or at home (Heinrich et al., 2017). Behavioral Risk Factor Surveillance System (BRFSS) has

illustrated increases in levels of LTPA at the national level from 1984 to 2015, with rates being generally steady in Pennsylvania (An et al., 2016). In addition to physical activity and sedentary behavior, unstructured activities can be captured through accelerometers and provide minute by minute counts (Evenson et al., 2015).

Built Environment and Physical Activity

The built environment comprises land-use patterns, transportation systems, and natural or built design features which provide opportunities or limitations for physical activity (Brownson et al., 2009; Handy et al., 2002). More specifically, it is often described as neighborhood features such as sidewalks, trails, parks, recreational facilities, traffic safety, and other factors that can potentially promote active transportation/utilitarian travel and LTPA among children, adults and older adults

(Ferdinand et al., 2012). Physical activity can be broken down into leisure, transportation and household related activities (Brownson et al., 2009). The built environment contributes 3-28% of the variance among studies examining physical activity participation (Li, Fisher, Brownson, & Bosworth, 2005; Nagel et al., 2008). Studying the built environment has involved a multidisciplinary approach involving researchers from the fields of public health, physical activity, and transportation/city planning, who examine how neighborhoods can be maximized to increase physical activity participation

(Brownson et al., 2009). A supportive built environment is seen as a sustainable strategy to encourage individuals to adopt or continue a physically active lifestyle (McCormack &

Shiell, 2011).

Generally, the built environment is divided into macro and micro -level measures

(Bauman & Bull, 2007; Bronfenbrenner, 1979; Brownson et al., 2009; Cain et al., 2017;

King & Clarke, 2015; McMillan, Cubbin, Parmenter, Medina, & Lee, 2010). Within the built environment field, macro environments include the overall community design reflective of walkability, such as street connectivity, residential density, and mixed land use (e.g. retail stores, food stores, parks, private recreation facilities). Microscale environments are smaller scale factors characteristic to neighborhoods and includes pedestrian facilities, street-crossings, traffic volume, crime, and incivilities (e.g., graffiti)

(Sallis et al., 2011). Other examples of microscale features include the quality of sidewalks, landscaping, adequate lighting, and others (Brownson et al., 2009). Micro scale features are known to be more changeable for interventions, since they are more cost-effective to implement (Cain et al., 2014; Millstein et al., 2013). Most research has been conducted on macroscale features and how they impact physical activity levels among both adults and adolescents (Brownson et al., 2009; Frank et al., 2010). This review will further investigate details regarding measuring the macrolevel built environment.

Measuring the Built Environment

The built environment can be studied through both subjective and objective measures (Adams et al., 2009; Brownson et al., 2009; Li et al., 2005; Nagel et al., 2008).

Objective measurements use direct observation audits or spatial data analysis, whereas subjective measurements are based on self-report surveys of residents regarding the perceived environment (Adams et al., 2009). For example, spatial data can be used to determine land use, the presence of recreational facilities, and information about crime, traffic, or land topography using Geographic Information Systems (GIS) (Frank, Schmid,

Sallis, Chapman, & Saelens, 2005). Earlier studies have used GIS to map out built

environment features, avoiding the need to collect data within the field and the ability to capture features from several neighborhoods (Frank, Schmid, Sallis, Chapman, &

Saelens, 2005). However, this presents a limitation as several built environmental factors can be missed, such as infrastructure and business establishments within neighborhoods

(L. Ma & Dill, 2015). More contemporary research on built environment has introduced built environment scales that researchers, community members/advocates, and policy makers can use to conduct an audit within neighborhoods of interest (Brownson et al.,

2009). Most studies before 2010 have been cross-sectional, making it difficult to determine a causal relationship between built environmental effects on PA levels; however, more recent studies have been conducting longitudinal designs and natural experiments to be able to infer these relationships (Frank, Schmid, Sallis, Chapman, &

Saelens, 2005). This study utilized both perceived and objective built environment measures to capture how the built environment affects physical activity levels among

Asian Americans.

Objective and Perceived Built Environment

The objective and perceived environment surrounding individuals are evaluated through different methods. Perceived versus objective measures of the built environment, including land use, the presence of recreational facilities, and different types of active transportation have been shown to be positively associated with overall physical activity

(Hoehner et al., 2005). Studies suggests that the objective built environment versus an individual’s perception of the built environment are not equivalent to each other and are distinct measures that should both be incorporated in assessing physical activity levels

(Gebel et al., 2009; Hinckson et al., 2017; Hoehner et al., 2005; Koohsari et al., 2015; L.

Ma & Dill, 2015; McGinn et al., 2007; Van Dyck et al., 2012). Several issues to consider that relate to distinguishing between perceived and objective measures, which include how neighborhoods are defined (e.g., buffer, census tract, etc.), the type of physical activity (active travel or leisure activity), and the conditions of recreational facilities that may be perceived differently than its overall presence (Hoehner et al., 2005). Ma and Dill

(2015), studied 616 adults in Portland, Oregon and found that the perceived and objective built environment were independently associated with bicycling behavior and frequency.

Furthermore, there was a higher combined variance explaining the model when both objective and perceived built environment were included in the analysis (L. Ma & Dill,

2015). In one study, Walk Score did not correlate with participant’s perceived access to having physical activity facilities (recreational centers and parks) nearby, thus indicating a lack of awareness regarding environmental access to physical activity facilities especially among more sedentary individuals (Carr et al., 2010).

Several factors can affect concordance between perceived and objective neighborhood environments. Perception of the built environment can be affected by psychological factors related to individuals past experiences, personality, and others

(John, 1987). Two people living in the same environment can perceive their built environment in different ways. There is little knowledge on what factors influence the ability of individuals to accurately report the features of their neighborhood; some studies have looked at how physical activity levels can lead to greater exposure and more accurate reporting (Adams et al., 2009; Leslie et al., 2007). For example, individuals who are more physically active within their environment can better recall the built environment in comparison to less physically active individuals. Furthermore, there is

ambiguity in determining what one’s neighborhood is defined as, as some may view it as a closer distance to their home while others can view it as a distinct feature such as the name of their city or town.

Perceived Built Environment and Physical Activity

The perceived built environment has been measured through several psychometric scales. In a review by Brownson and colleagues (2009), there were over 100 studies conducted measuring the perceived built environment, incorporating questionnaires ranging from 7 to 68 items. A meta-analysis of the perceived built environment on physical activity found that the presence of facilities including physical activity facilities, shops and services, and features including sidewalks were positively associated with physical activity (Duncan, Spence, & Mummery, 2005). In this analyses, the perceived built environment only counted for a small portion of explaining physical activity behavior; the meta-analysis indicated this variance to be from 4-7% to explain total physical activity behavior (Duncan et al., 2005). With respect to a national study, having higher perceived access to parks and trails had a positive association with physical activity among US adults (Brownson et al., 2009).

Self-reported knowledge and perception about the physical activity environment was mainly a method to collect environment influences prior to the introduction of GIS technology for objectively measuring the environment (Forsyth et al., 2006). However, with the introduction to novel software for measuring the built environment, there is a potential to expand predictors of engaging in physical activity and the amount of variance that can explain this behavior due to objective features of the built environment.

Objective and GIS Built Environment Measures and Concepts

Several objective methods have been identified for measuring the built environment, which include: 1) primary spatial data and field observations, 2) geospatial analyses and 3) online and commercial sources. Primary spatial data is collected by fieldwork auditors who directly make environmental observations and secondary spatial data are derived from external sources (commercial, administrative, internet sources, phone directory) that are more cost-effective to obtain (Thornton, Pearce, & Kavanagh,

2011). Some methods are more cost-effective than others are and there are several drawbacks reported with their accuracy and validation, strengths, and weaknesses of each method are discussed below.

Field observations through the use of built environmental audit tools are one of the main methods used in physical activity research and are considered a ‘gold standard’

(Brownson et al., 2009; Han, Powell, Slater, & Quinn, 2012). Fieldwork audits consist of community and built environment audit instruments, such as the Physical Activity

Resource Assessment Tool (PARA) and Environmental Assessment of Public Recreation

Spaces (EAPRS) which measure following components: land use, streets and traffic, bicycling facilities, public spaces and amenities, parking, maintenance, and safety indicators (Brownson et al., 2009). Researchers studying park features and characteristics have created tools such as the System for Observing Play and Recreation in

Communities, which aim to identify physical activity levels of park users (SOPARC)

(McKenzie, Cohen, Sehgal, Williamson, & Golinelli, 2006). Furthermore, audit tools have been more recently developed to measure microscale features of one’s neighborhood, such as the Microscale Audit of Pedestrian Streetscapes (MAPS)

(Millstein et al., 2013). These audit tools are laborious and costly to conduct, often

requiring several resources in order to conduct the audits and are often limited to geographic locations (Carr et al., 2011; Evenson et al., 2016; E. Han et al., 2012).

A second common method of objective measures of the built environment are through geospatial analyses with the use of GIS. GIS allows researchers to utilize spatial data for describing the environment and determining patterns between environmental features and health outcomes or behaviors (Thornton et al., 2011) GIS has been used to assess the built environment and physical activity through utilizing existing data comprising spatial characteristics, such as addresses or census tracts. The main GIS- based measures include: population density, land-use mix, recreational facilities, street pattern, sidewalk coverage, vehicular traffic, and crime (Brownson et al., 2009).

However, several limitations of using GIS measures include challenges in measuring the quality of built environment features (e.g., conditions of a trail, graffiti on playground equipment), restricting neighborhoods to be measured by census tract or other boundaries that don’t match one’s actual neighborhood boundaries, and the inaccurate combining of measures. In addition, studies have shown high disconcordance between secondary spatial data. Overall, using GIS and secondary spatial data that has already been collected

(e.g., market research companies, phone directories, etc.) is more cost-effective than field observations, and can be analyzed over larger geographic areas (Thornton et al., 2011).

Accessibility is the ease of accessing built environment features and sometimes defined as the presence of certain features and include several components (Thornton et al., 2011). A buffer is a metric used to measure accessibility, which is defined as boundaries placed surrounding at a single point, buffers are defined along with the scale

(e.g., 100 meters to 1 mile), type of buffer (Euclidian or network), and the origin that the

researcher is measuring from (e.g., a home or another single point) (Thornton et al.,

2011). Figure 6 describes the buffers applied to an individual’s address. Furthermore, density is also a metric used to measure accessibility, which is defined as the count of features within a specific area, or the relative number of features per population or geographic area (Thornton et al., 2011). Kernel Density is a type of density measure which converts point data into a continuous density surface map, providing a density value for each grid cell to calculate density values across the whole study region

(Thornton et al., 2011). Lastly, proximity is also a component of accessibility, which is characterized by the distance in which a feature is closest to a particular point. Proximity can be measured by Euclidian or network distance (Thornton et al., 2011).

Figure 2-2. Measures of accessibility (L. E. Thornton et al., 2011)

With respect to measuring accessibility to physical activity resources, GIS has evolved to provide enhanced metrics to better define accessibility such as density within buffers, kernel density estimation, and proximity based on network distance or activity- spaces which represent locations that individuals visit (Thornton et al., 2011). A strength of this was study was the use of the network distance of physical activity resources and the impact it has on each individual’s physical activity level.

Other objective measures of the built environment include online sources that have been developed to provide a convenient and simple way to measure aspects of neighborhood walkability. The Walk Score provides a valid proxy measure of neighborhood walkability that is publicly available, based on three indicators: 1) shortest distance to equally weighted category of amenities, which include grocery stores, coffee shops, restaurants, bars, movie theatres, schools, parks, libraries, book stores, fitness centers, drug stores, hardware stores, and clothing/music stores; 2) block length; and 3) intersection density (Carr et al., 2010, 2011; Duncan, Aldstadt, Whalen, Melly, &

Gortmaker, 2011; Hall & Ram, 2018). Any amenities located within a one-quarter mile gets assigned the maximum number of points for that category and the value declines as the distance approaches 1-mile, administering zero points for greater than 1-mile (Carr et al., 2010, 2011; Duncan et al., 2011; Hall & Ram, 2018). The more amenities are in close proximity to one’s home, the more walkable a neighborhood is and thus the higher the score could be on a scale from 0-100. The Walk Score for large scale studies are obtained from a Google AJAX Search application program interface (API), which provides periodically updated geographic data (Walk Score, 2019a). However, other studies

directly input the address into the website interface in order to determine the associated

Walk Score.

Carr, Dunsiger and Marcus (2011) assessed the validity of an older version of the

Walk Score prior to 2010 by replicating the methodology used by researchers to match them to Walk Scores determined by GIS metrics. Validity testing established that the

Walk Score exhibited good reliability and high agreement, with strong correlations for the following amenities present: Food vendors (restaurants, coffee shops, bars) = 0.71; grocery stores = 0.68; parks = 0.61; movie theatres = 0.56; schools = 0.70; libraries =

0.57; fitness facilities = 0.66; drug stores = 0.72; retail (clothing, music, book, hardware stores) = 0.73, and a total amenity correlation of 0.74 (Carr et al., 2011). Thus, it has been justified as a valid tool to use as a built environment measure. Additionally, it is noted that the

Walk Score is a strong measure to utilize for cross comparisons because it is standardized and ubiquitous across different countries (Owen et al., 2004; Saelens &

Handy, 2008; Thielman et al., 2015). In another validation study, the Walk Score showed strong correlation with several GIS measures when a 400, 800, and 1600 meter buffer were applied surrounding the individual’s home (Duncan, Aldstadt, Whalen, & Melly,

2013). Thus, it was shown to be a strong measure generalizable to different geographic locations across a wide sector of scales (Duncan, 2013). Systematic reviews on the built environment have expressed that inconsistent measures across research studies makes it difficult to make conclusions about the findings, thus the Walk Score helps to make standardizations across studies (Thielman et al., 2015).

The Walk Score has been shown to have mixed findings with physical activity, including purposive walking, transport mode, and in predicting health related outcomes mainly when comparing very low and very high walkable neighborhoods (Hall & Ram,

2018). A large multiethnic study including Chinese, Hispanic and African Americans across six cities (Los Angeles, Chicago, Baltimore, St. Paul, New York, and Forsyth

County) found a 12% lower odds of not walking with every 10 point increase in Walk

Score (Hirsch et al., 2013). Furthermore, Walk Score was associated with more minutes of walking for transport and leisure walking (Hirsch et al., 2013). Among a study of recent Cuban immigrants who had moved to the US in 90 days or less, researchers found a significant relationship between neighborhood walkability measured by the Walk Score and purposive walking to amenities (Brown et al., 2013). This study highlights that participants had no selection of their residence in the US, removing neighborhood selection bias from the study (Brown et al., 2013).

Although it provides an accurate measure of neighborhood walkability, it has some limitations. It is advised that the Walk Score be used as a proxy for estimating accessibility to amenities and neighborhood density rather than a global measure of neighborhood walkability, as other features such as crime, were incorrectly correlated with the Walk Score (Carr et al., 2010). Some limitations of the Walk Score are they do not provide all neighborhood walkability measures, it does not take into consideration the area of size of an amenity, and for example, how the size of parks may influence the

Walk Score.

Commercial data sources that compile business lists from multiple sources are also utilized in the field and integrated with GIS to measure accessibility and availability

of resources in the environment (Hoehner & Schootman, 2010). Commercial data sources in research studies have mainly been used to identify retail food outlets and physical activity facilities to determine the impact of health outcomes and behaviors through examining food and physical activity environments (Han et al., 2012; Hoehner &

Schootman, 2010). Among those commercial databases, InfoUSA and Dun and

Bradstreet (D&B) are two commonly used sources in research (Hoehner & Schootman,

2010). However, few validation studies conducted have illustrated that secondary sources for identifying physical activity facilities and food environments can have reported errors and lower accuracy, thus, should be used with caution (Thornton et al., 2011).

Benefits of using these commercial databases are that they are cost-effective, but there are concerns regarding how valid they are because they’re not naturally designed for conducting research and is more tailored for market purposes, thus this can introduce error and bias (Thornton et al., 2011). Thus, the accuracy and completeness of results may be of concern (Hoehner & Schootman, 2010). Several studies have determined the validity and concordance of this method when comparing them to field audits or comparisons among different types of commercial databases. Hoehner and Schootman

(2010) compared the concordance between InfoUSA and D&B; among studies that compared the commercial databases to field audits and/or governmental records, a moderate agreement was found for physical activity facilities (Boone, Gordon-Larsen,

Stewart, & Popkin, 2008; Paquet, Daniel, Kestens, Léger, & Gauvin, 2008). Boone and

Colleagues (2008), for example, compared the concordance between commercial databases and field audits by researchers among 80 census blocks in an urban and nonurban community. On the contrary, potential limitations and errors that can occur with

the commercial database and GIS geospatial analyses include data that is incomplete, facilities that are not correctly classified or are outdated, and locations that are not geocoded properly (Boone et al., 2008; Hoehner & Schootman, 2010). Extra caution should be taken to ensure the data is complete through incorporating other data checks such as online audits or use of additional commercial databases.

Built Environment Components

Urban Sprawl

As a result of changing land development patterns, one of the consequences of degrading health and lowered physical activity participation is urban sprawl (McCormack & Shiell,

2011). Urban sprawl is defined as the increase in dispersion of a population into lower- density residential areas (Ewing, 2005; Ewing et al., 2003). Implications of urban sprawling have created increased separation of homes, shops and workplaces, lack of distinct or thriving centers (e.g. down town or suburban town centers), large block densities between roads and poor accessibility from one place to another (Ewing, 2005;

Ewing et al., 2003). This development pattern of cities is directly related to the creation of metropolitan areas that have emerged in the US and have become more car dependent as a means of transport rather than engaging in active transportation such as walking or biking (Ewing, 2005). Metropolitan sprawling has contributed to less physical activity related to active travel, which includes walking to work or taking public transportation which still includes some aspect of walking (Ewing, 2005).

Few urban sprawl metrics were developed by several researchers to quantify urban sprawl and its impact on several health and environmental related outcomes

(Laidley, 2016). Among them, an urban sprawl index was developed and then updated a decade later to provide improved metrics (Ewing et al., 2003, 2014). The urban sprawl index consisted of four factors, including development density, land use mix, activity centering and street accessibility (Ewing et al., 2014). Using this metric, relative to different counties, the Philadelphia-Camden-Wilmington metropolitan area ranked as the

5th most compact city, thus indicating lower urban sprawl following New York, Kings,

San Francisco, and Bronx counties (Ewing et al., 2014). However, in more recent urban sprawl index developed, Philadelphia County was not listed as one of the top 10 least sprawling cities (Laidley, 2016).

A study using the Behavioral Risk Factor Surveillance System (BRFSS) data to measure the effect of urban sprawl on health behaviors found that a higher sprawl within counties was associated with lower levels of physical activity, including less leisure walking and associated with higher levels of obesity (Ewing et al., 2003). Thus, physical activity is a mediator in the relationship between urban sprawl and obesity rates (Ewing et al., 2003). However, the revised version of the urban sprawl index and physical activity did not find any significant associations due to the poor measures utilized for physical activity and only used one item and not specific domains (Ewing et al., 2014).

Residential self-selection has been shown to be a strong predictor indicating that individuals choose to move to less sprawling cities and thus leading them to engage in more physical activity and have a lower body mass index (BMI). The Harvard Alumni

Health Study conducted a longitudinal analyses on the impact of urban sprawl changes measured by tract level to their physical activity levels (Lee, Ewing, & Sesso, 2009).

Three categories of urban sprawling changes between 1988 to 1993 were created,

including those who had moved to a less sprawling area, those who moved to a more sprawling area and no changes in sprawling (Lee et al., 2009). In this longitudinal analysis there were no statistically significant changes in physical activity based on urban sprawling changes due to residental self-selection and study power to detect significance

(Lee et al., 2009).

Neighborhood Walkability Measures

The concept of neighborhood walkability has been significantly expanded over the last three decades to encompass more measures. Originally, the main metrics for neighborhood walkability included street connectivity, residential density, and land use mix (Hajna et al., 2015). Nonetheless, neighborhood walkability is a strong promoter of walking as a means for transportation and thus a more active lifestyle (Adams et al.,

2009; Cerin et al., 2017; Sallis et al., 2012). Connecting residential areas with work and leisure based destinations have a relationship with active transportation levels (Sallis et al., 2012). Improved access to public transit (bus and rail stops) are directly related to increased active transportation, and those who use public transportation are more likely to meet physical activity guidelines. Furthermore, transit accessibility is measured in studies by the density of transit stops that are available within a certain identified buffer (M. A.

Adams et al., 2014).

Street connectivity is defined as the amount of three or more-way intersections per square kilometer within a neighborhood buffer and an indicator of shorter, direct and various routes to destinations (Hajna et al., 2015; Jáuregui et al., 2017; Zuniga-Teran et al., 2017). The number of intersections as well as 4-way intersections within a census block is found to be associated with increased walking distance (Boarnet et al., 2008; Cao

et al., 2006; Chatman, 2009; McCormack & Shiell, 2011). The more intersections present, the easier it is to get around between the individual’s origin and destination

(Saelens, Sallis, & Frank, 2003). Neighborhoods that have slow traffic indicative of many intersections is more appealing for pedestrians because it allows them to reach their destinations through many routes (Leslie et al., 2007).

Land use mix is defined as the types of land uses within a neighborhood (Saelens,

Sallis, & Frank, 2003). There is a larger land use mix in downtown neighborhoods compared to suburban neighborhoods, which makes it more convenient for active travel

(Hajna et al., 2015; Saelens, Sallis, & Frank, 2003). Land types include residential (e.g. apartments) and commercial land (e.g., stores, churches, schools, other services) (Leslie et al., 2007). A greater heterogeneity in land use mix indicates a higher score (Saelens,

Sallis, & Frank, 2003).

Residential density is defined as the number of residences per square kilometer of residential land area within the home buffer (Adams et al., 2014; Frank, Andresen, &

Schmid, 2004). The higher the residential density, the more conducive the environment is to non-motorized forms of transport (Saelens, Sallis, & Frank, 2003). Gross and net residential density are the two main measures used in built environment studies and is usually measured derived from census data (Frank et al., 2010).

Perceived built environment related to walkability has been investigated among adults in the US and globally (Cerin, Conway, et al., 2013; Hinckson et al., 2017;

Jáuregui et al., 2017; Maddison et al., 2009; Saelens, Sallis, Black, et al., 2003). The

Neighborhood Environment Walkability Scale has been the most commonly used scale across diverse settings (Brownson et al., 2009; Cerin, Conway, et al., 2013). Originally

developed by Saelens, Sallis and Colleagues in 2002, the measurement was first validated and tested in two San Diego neighborhoods to examine their association with physical activity among residents, consisting of 8 neighborhood environment factors/subscale

(Saelens, Sallis, Black, et al., 2003). Overall, there was a strong parallel between high neighborhood walkability and perceived high residential density, mixed land use, street connectivity, aesthetics and traffic related safety [F1,105>9.69 for each subscale, P<.003]

(Saelens, Sallis, Black, et al., 2003). Test-retest reliability analyses indicated intraclass correlation values of greater than or equal to 0.75 for most of the subscales (Saelens,

Sallis, Black, et al., 2003). In another reliability study, intraclass correlation coefficients

(ICC) obtained from a US sample illustrated moderate to high test-retest reliability with

ICC values of the following: Land use mix-diversity=0.93, residential density=0.78, land use mix-access=0.77, street/walking environment=0.41, infrastructure for walking/cycling=0.76, neighborhood aesthetics=0.66, neighborhood safety=0.69, and neighborhood satisfaction=0.65 (Brownson et al., 2004).

The NEWS has also been tested among a multi-country pooled analyses and offered in several languages and culturally adapted in other countries within Europe,

Africa, Asia and South America, illustrating strong comparability across 12 different countries (Cerin, Conway, et al., 2013). As an example, the NEWS has highly been regarded in several environments and adapted to several cultures, including Chinese

Seniors (NEWS-CS) utilized in China and Hong Kong, taking into consideration several differences in environments and cultural characteristics that Chinese seniors are accustomed to (Cerin et al., 2010, 2011). For example, some of the modified environmental settings included the addition of a land use diversity subscale that

included: ‘noodle shops’, ‘Chinese fast food restaurants,’ ‘religious practices’ and a social disorder/littering subscale that included: ‘objects dropping from high-rise buildings’ (Cerin et al., 2010, 2011). Findings vary based on country and whether they are high or middle-low income as well as where the study is conducted. Among Mexican adults living in a large urban area perceived high neighborhood aesthetics was the most important correlate for leisure physical activity (Jáuregui et al., 2017). In Hong Kong, where the population and residental density are much higher, considerations should be made about certain subscales.

An abbreviated NEWS scale was developed as well and showed strong reliability measured through confirmatory factor analyses among 1286 adults (Cerin et al., 2006).

Based on the confirmatory factor analyses of the NEWS, measures for the abbreviated version were selected based on high criterion validity with respect to the individual measures being strongly correlated with walking for transport and recreation, strong construct validity measures by factor loading and strong test-retest reliability (Cerin et al.,

2006). Those measures which did not fit the criteria mentioned were not included in the

NEWS-A (Cerin et al., 2006).

There has been mixed findings between identifying which environmental factors relevant to neighborhood walkability are associated with physical activity levels and this differs by physical activity domain. Neighborhood walkability has been linked to higher levels of transportation related walking (e.g. walking to the grocery store) and cycling, and is weaker/inconclusive for its association with leisure-time walking (Cerin et al.,

2007; Hajna et al., 2015; Jack & McCormack, 2014; McCormack & Shiell, 2011;

Zuniga-Teran et al., 2017). A meta-analysis investigating neighborhood walkability

found that higher street connectivity, land use mix, and residential density as indicators of higher walkability derived by GIS was associated with 766 steps per day as measured by biosensors (Hajna et al., 2015). One of the first novel built environmental studies conducted in the US, specifically on walkability, was the Neighborhood Quality of Life

Study (NQLS) (Frank et al., 2010; Sallis et al., 2009). In this study, a composite GIS walkability index was created to categorize various neighborhoods into high and low walkability in order to determine physical activity levels of residents with the King

County-Seattle and Baltimore-Washington regions. Four quadrants were created based on the walkability index and income level to control for income within 32 neighborhoods.

This study was one of the first to demonstrate that neighborhood walkability is an important environmental correlate of physical activity (Frank et al., 2010). Across both counties, the rate of walking to work was 4-6% higher among those living in a high- income/high walkability versus a high-income/low walkability neighborhood based on county (Frank et al., 2010)

Neighborhood Parks and Green Spaces

Green Exercise is a term used to describe natural spaces in which exercise can occur (Gladwell et al., 2013). Parks, trails and playgrounds are part of the infrastructures found within green spaces that promote physical activity and serve as main facilities to engage in leisure physical activity as well as destinations for walking (Dahmann et al.,

2010; Kaczynski et al., 2014). These facilities promote physical activity within surrounding communities and with the ability to reach wider populations. Especially in urban areas, parks and other green spaces can serve as an important place for promoting community physical activity and may serve as the only settings in which residents can

engage in outdoor physical activity (Baran et al., 2014; Douglas et al., 2018; Hughey et al., 2016; Joseph & Maddock, 2016; Kaczynski et al., 2014; Schipperijn et al., 2013). The presence of public parks can promote leisure physical activity, especially among ethnic minority and underserved populations, as they offer free and low-cost activities while reaching larger populations (Floyd et al., 2008; Moody et al., 2004). Several considerations exist for determining what factors enhance park use and physical activity that deal with not only park design/characteristics but issues of safety, temperature/weather, and overall aesthetic appeal (Baran et al., 2014; Joseph &

Maddock, 2016; Kaczynski et al., 2008).

Methods to measure park accessibility have included GIS measures and direct observation using audit tools such as SOPARC and EAPRS, in addition, surveying participants living close to parks assessing their physical activity behavior (Cohen et al.,

2007; Evenson et al., 2016; Joseph & Maddock, 2016). Despite the well-known physical and mental health benefits of exercising in green spaces, there is a lack of consensus on whether presence and/or access to parks and other urban green spaces are associated with increased physical activity (Schipperijn et al., 2017). Conflicting findings are mainly due to the large heterogeneity in defining accessibility of parks and the wide range of measures that have been used among research (Schipperijn et al., 2017).

In a large global-based study, The International Physical Activity and

Environment Network (IPEN) examined if certain park characteristics were associated with physical activity in different urban cities across 8 different countries in the world

(Schipperijn et al., 2017). Their study incorporated several measures for park accessibility, including the number of parks by size and buffer, area of parks by buffer

size, and distance to parks within 100 meters based on varying total square foot area of parks (Schipperijn et al., 2017). Having one more park in one’s neighborhood was associated with engaging in 1.8 more minutes of moderate-to-vigorous physical activity

(PA) each week (Schipperijn et al., 2017). Furthermore, those living in neighborhoods with the highest number of parks within a kilometer radius of their neighborhood engaged in 24 more minutes of moderate-to-vigorous physical activity weekly measured by accelerometer (Schipperijn et al., 2017). Figure 7 illustrates the different park accessibility measures and total moderate-to-vigorous physical activity; positive associations with perceived proximity to parks, number of parks within a 1km buffer, number of parks within a 0.5km buffer and negative association with greater distance from parks that are at least 1 acre in size were observed (Schipperijn et al., 2017).

Figure 2-3. Park accessibility and levels of MVPA. (Schipperijn et al., 2017).

In contrast, a systematic review of observational studies found a wide range of percentage of moderate-to-vigorous physical activity performed in parks, from 31-85%.

The majority of studies being conducted in the southern or western USA (Joseph &

Maddock, 2016). Han and colleagues (2013) estimated moderate-to-vigorous physical activity using a model for measuring the impact of parks on contributing to daily physical activity. Parks accounted for 14% of moderate physical activity among those who lived within 0.5 miles of the park, and contributed to 4% among those living within 1.0 miles of the park (Han, Cohen, & McKenzie, 2013).

One study which assessed proximity and facilities associated with parks found that the number but not proximity of parks found within one mile of residents’ homes were associated with higher levels of park based physical activity (Kaczynski et al.,

2014). Furthermore, a large systematic review on park proximity and density and objectively measured physical activity found mixed findings in studies measuring park density and proximity with objective physical activity, but a stronger relationship was observed when there was a smaller buffer applied to park distance from the participants’ homes (Bancroft et al., 2015). In this review, park density and proximity was both subjectively and objectively measured, which may have contributed to the heterogeneity in study findings. Also, although physical activity level was objectively measured, there was little indication of specific park related physical activity or GPS data to identify activity in parks among the studies (Bancroft et al., 2015). They suggested that when measuring density of parks, that 0.25-1 mile buffers should be applied (Bancroft et al.,

2015).

Researchers have also tried to delineate which components of parks are the most beneficial for promoting physical activity (Cohen et al., 2007; Kaczynski & Henderson,

2007). Park features such as trails, basketball courts, playgrounds, baseball fields, fitness stations and skate parks are associated with more park use and park based physical activity (Baran et al., 2014; Cohen et al., 2009; Kaczynski et al., 2008, 2014; Kaczynski

& Havitz, 2009). Specific park features including paved and unpaved trails as well as wooded areas (dense growth of trees, shrubs or bushes) were more associated with reported park based physical activity participation (Kaczynski et al., 2008).

Although the literature on park accessibility has been mixed, several implications for park based physical activity have been identified. More longitudinal studies are needed to investigate behavior following changes in built environment interventions, as most studies in this area are of cross-sectional nature. Gender differences in park use were found with males engaging in moderate-to-vigorous physical activity in parks to a greater extent than females (Cohen et al., 2007; Evenson et al., 2016; Joseph & Maddock,

2016; Kaczynski et al., 2008). Thus, interventions to address gender specific built environmental impacts are needed. Overall, recommendations towards living at least within one mile of a park is suggested among study findings on the relationship between park accessibility and enhancing physical activity levels (Cohen et al., 2007).

Recreational Facilities Accessibility

Research on recreational facilities within neighborhoods and their impact on physical activity have been mixed (Diez Roux et al., 2007). Several of these studies are older (between 1990-2005) and used simple measures such as perceived accessibility to address recreational facility access (Humpel et al., 2002). The use of GIS measures to illustrate recreational facility accessibility can help to validate the findings from these studies that used self-report measures (Humpel et al., 2002). Having access to a mix of local recreational and non-recreational destinations is positively associated with leisure walking (Hoehner et al., 2005; McCormack & Shiell, 2011). However, regarding the access and availability of recreational facilities for moderate and vigorous physical activity, systematic reviews have identified mixed results, generally showing a moderate positive relationship between increased physical activity among those who have higher

access and convenience to facilities (Brownson et al., 2009; Humpel et al., 2002;

Kaczynski & Henderson, 2007).

An earlier study conducted in 1990 postulated that proximity to closer facilities would impact individuals through serving as visual stimuli for promoting exercise behavior in which others would view individuals exercising in facilities and internalize this behavior as the norm (Sallis et al., 1990). Other benefits of being in close proximity to exercise facilities is that it reduces psychological and physical related barriers associated with travelling and traffic, as well as encourages walking to these facilities

(Sallis et al., 1990). Statistically significant differences were found when comparing exercisers versus sedentary participants, with vigorous exercisers living closer to pay facilities even after controlling for sociodemographic characteristics (Sallis et al., 1990).

However, they failed to determine a relationship between participation in certain exercises provided by these facilities and being in closer to proximity to those relevant facilities (Sallis et al., 1990).

Similar to parks, newer studies on recreational facilities have investigated recreational density based on the type of buffer applied (e.g., 0.5, 1.0 and 5.0 mile buffers) with respect to resident’s address (Diez Roux et al., 2007). Diez Roux and colleagues (2007) examined density of recreational facilities in New York (NY),

Baltimore (MD), and Forsyth County (NC), and found that increased density of recreational resources within a one mile radius was associated with a prevalence ratio of

PR=1.14 (95% CI: 1.03-1.26) among high densely populated neighborhoods and a

PR=1.06 (95% CI: 0.95-1.18) among mid-densely populated neighborhoods. Thus, high recreational resource density was positively associated with physical activities including

engaging in sports, conditioning and other individual activities (e.g., golf, bowling, yoga, etc.) in this study (Diez Roux et al., 2007).

Physical Activity Resource Assessment (PARA) has been used to identify the availability and systematically assess the quality of physical activity resource features, amenities, and incivilities in urban neighborhoods (Lee et al., 2005). Types of physical activity resources that are audited include a fitness club, park, sport facility, trail, community center, school, or combination facility (e.g., park and trail combined) within a participant’s neighborhood. Incivilities are defined as elements that would reduce the pleasure associated with using that physical activity resource, which include auditory annoyances, broken glass, dog refuse, unattended dogs, evidence of alcohol and substance abuse, graffiti, litter, not enough grass or overgrown grass, sex paraphernalia, and vandalism located within recreational facilities (Sallis et al., 2012). Although direct observations will not be made for this dissertation, importance of incivilities exist in built environment studies. Furthermore, pay versus public facilities are important to take into consideration and not necessarily the quantity of facilities within neighborhoods.

Built Environment and PA among Underserved Communities

Environmental Justice Issues in the Built Environment and Physical Activity

Environmental justice is defined as a disproportionate exposure to harmful environmental conditions in low-income and racial/ethnic minorities (Taylor et al., 2006). It aims to determine inequalities in access to quality built environment characteristics, such as environmental amenities, that persist among low-income and ethnic communities of color

(Boone, Buckley, Grove, & Sister, 2009; Wolch, Wilson, & Fehrenbach, 2005). Although

less studied, these inequities are embedded within historical and political influences as a consequence of racial segregation/discrimination, unequal exposure to harmful industrial activities, and housing and employment discrimination, to name a few (Moore, Diez

Roux, Evenson, McGinn, & Brines, 2008; Williams, Mohammed, Leavell, & Collins,

2010; Wolch et al., 2005) Past environmental justice issues placed a heavy emphasis on how low SES and minority groups were exposed to environmental hazards (Mohai &

Bryant, 1991). Researchers started to identify that these inequities permeated into other aspects of environmental inequalities, such as a lack of accessibility to parks and recreational facilities (Wen et al., 2013). The association between the built environment and differences in physical activity levels among race/ethnicity and income have shown mixed results due to potential lack of generalizability of findings across different cities and urban areas and methodology used to measures accessibility to resources (Sallis et al., 2011). Generally, environmental justice studies conceptualized these groups as the level of ethnic minority composition within an area (e.g., census tract, zip code) (Gordon-

Larsen et al., 2006; Hughey et al., 2016; Vaughan et al., 2013).

In the past decade, studies examining the relationship between the built environment and physical activity have been conducted in ethnic minority groups and examined several built environment characteristics (Ferdinand et al., 2012). Studies have identified apparent disparities in the built environment based on ethnicity when it comes to the spatial accessibility and availability of parks and recreational facilities (Dahmann et al., 2010; Gordon-Larsen et al., 2006; Kaczynski et al., 2014; Osypuk, Roux, Hadley,

& Kandula, 2009). Environmental inequities in green space may be potentially present among communities, and measured by size of vegetation and amount of space present in

neighborhoods, which GIS studies have aimed to determine (Wen et al., 2013; Wolch,

Byrne, & Newell, 2014).

Inequalities in park accessibility measures among ethnic minorities

Parks are community infrastructures that promote physical activity and healthy being of individuals (Hughey et al., 2016). Not only do parks provide benefits to enhance physical activity and mental/physical health, but also benefits that are economically and ecosystem related (Boone et al., 2009). Parks were largely studied within the environmental justice realm as parks were viewed as a positive example of how to enhance society and get rid of social problems that existed within certain communities.

Substantial research has been conducted on the relationship between parks with physical activity levels among ethnic urban communities, which has for the most part illustrated evident disparities in accessibility (Bancroft et al., 2015; Boone et al., 2009; Cohen et al.,

2007; Estabrooks, Lee, & Gyurcsik, 2003; Hughey et al., 2016; Jones et al., 2015;

Kaczynski et al., 2014; Vaughan et al., 2013; Weiss et al., 2011; Wen et al., 2013; Wolch et al., 2014). In the park literature, findings illustrate that the number of parks and proximity of parks within ethnically dense neighborhoods are not necessarily undistributed among ethnic neighborhoods in comparison to NHW neighborhoods

(Bruton & Floyd, 2014; Dahmann et al., 2010; Taylor et al., 2006; Vaughan et al., 2013;

Wen et al., 2013). Rather, an uneven spatial accessibility and quality of parks and associated facilities, including features and amenities among ethnic urban neighborhoods are what contributes to these environmental injustices (Bruton & Floyd, 2014; Dahmann et al., 2010; Taylor et al., 2006; Vaughan et al., 2013; Wen et al., 2013).

There is little evidence to suggest that the number of parks are unevenly distributed among ethnic versus NHW neighborhoods but varies among cities and urban environments (Vaughan et al., 2013; Weiss et al., 2011; Wen et al., 2013). Some studies have found that neighborhoods with predominantly African Americans and Hispanics showed they had more number of parks compared to White neighborhoods while others have found no difference (Vaughan et al., 2013; Weiss et al., 2011; Wen et al., 2013).

Thus, number of parks may not necessarily present an environmental justice issue, but other park characteristics such as size of parks, presence of incivilities and park quality may explain these inequities exist. Some reasons for this may be that ethnic minorities still reside in the older historical city areas that underwent large development and had established parks/recreational spaces, despite poor quality of the resources (Vaughan et al., 2013).

Park accessibility among Ethnic Minorities

With regards to the relationship between how park accessibility influences physical activity levels in ethnic minorities, a study conducted in Los Angeles among 8 parks in ethnic communities found that proximity to parks predicted higher physical activity levels but there were large unmet needs that communities had (Cohen et al.,

2007). In observing environmental injustice at in two urban cities, Baltimore and Los

Angeles, two studies discovered that although there were more parks in neighborhoods of predominantly ethnic minorities that were within a quarter-mile distance within census block groups (African American in Baltimore, Latinos in Los Angeles), the total park acreage was much less, and that white neighborhoods predominantly had more acres of parks (Boone et al., 2009; Wolch et al., 2005). Less park acreage was also found among

Los Angeles census tracts with more foreign born residents and high levels of ethnic minority groups (Weiss et al., 2011).

In order to address environmental justice issues beyond accessibility to parks, evaluating how parks contribute to the quality of life and health are needed (Boone et al.,

2009). For example, it’s less clear whether immigrants utilize parks, school playgrounds and active travel within their neighborhood regardless of the presence of park space as those spaces may be unsafe or that the space may belong to another community

(Ferdinand et al., 2012; Stodolska et al., 2011). Thus, the issue of social accessibility and whether this impacts ethnic minorities utilization of those spaces needs to be also taken into consideration (Wen et al., 2013). Participation in physical activity in parks has been linked to better psychological and physical health (Barton & Pretty, 2010).

Park Incivilities among Ethnic Minority Neighborhoods

Few studies have argued that although environmental justice research has focused on identifying how park presence and accessibility differ by ethnicity and socioeconomic status, less is known regarding how there are disparities in incivilities that can also and more importantly can impact physical activity levels (Weiss et al., 2011). More specifically, exposure to higher number of incivilities in parks was found to be negatively associated with physical activity levels among ethnic neighborhoods (Weiss et al., 2011) .

A recent study directly targeting ethnic communities of color assessed the physical and social quality of parks along with characteristics of park users in Latinos, Asians and

African Americans in San Jose, California (Douglas et al., 2018). It was found that physical disorder characterized by park incivilities (e.g., broken glass, litter, vandalism) negatively impacted physical activity levels among park users (Douglas et al., 2018).

Although Latinos were the most frequent park users, they engaged in less physical activity compared to Asians when visiting the park, as measured by the SOPARC

(Douglas et al., 2018). Studies that utilized built environment in-person audits found that there were more incivilities in lower income and ethnic minority neighborhoods. Among one study in Greensboro North Carolina, researchers used the EAPRS tool to analyze 21 parks on several characteristics (Bruton & Floyd, 2014). They found that minority neighborhoods had fewer wooded areas, less tree cover (lower tree canopy), less clean sitting and rest areas, indicating a lack of park maintenance among these parks (Bruton &

Floyd, 2014).

Furthermore, a study conducted in Kansas City, Missouri census tracts conducted audits of parks across the county to determine how the quality of parks and park features were distributed based on the percentage of residents within census tracts that were ethnic minorities and of low-income (Vaughan et al., 2013). The Community Park Audit Tool

(CPAT) was utilized to assess these outcomes; there were no significant differences in park availability or number of park features, but higher minority census tracts had different amenities such as more basketball courts and less trails, and lower income census tracts had more incivilities and aesthetic features within each park compared to medium income level census tracts (Vaughan et al., 2013).

Ethnic Minority Composition and the Built Environment

Despite the existence of environmental injustices based on park accessibility, socioeconomic disadvantage, which incorporates multiple socioeconomic indicators (high poverty, low education, renting housing occupancy, and employment rate) is proposed as an important composite measure as well (Hughey et al., 2016). Higher disadvantaged

neighborhoods had twice as many incivilities classified as poorly maintained and high presence of trash, graffiti, and vandalism within their parks. Furthermore, ethnic minority composition was an important moderator in this relationship, where there was an increased incident rate of incivilities and decreased incident rate of amenities with increased levels of neighborhood racial/ethnic minority composition (Hughey et al.,

2016).

Figure 2-4. Rate of Park Amenities with Increasing Ethnic Minority Composition (Hughley et al., 2016)

Examining park quality and features such as incivilities that are present and what types of amenities exist in ethnic minority neighborhoods should be assessed in conjunction with the availability of these resources to fully comprehend how environmental injustices may exist. The impact of park maintenance, and presence of incivilities can serve as barriers to participation in physical activity which has been previously documented (Suminski et al., 2012). Thus, the effects of these characteristics rather than generally accessibility to parks in general may be of larger importance in predicting physical activity (Suminski et al., 2012). Ensuring equitable quality of these resources for ethnic minorities is important for promoting healthy life and reduce

challenges and barriers to overall health and well-being. Future studies can also examine how these environmental justice characteristics in turn impact physical activity levels directly among those ethnic groups who live within these neighborhoods.

Inequitable Distribution and Accessibility of Recreational Facilities

One of the most comprehensive studies that represented almost one-tenth of the

US population was conducted in 7139 census tracts in six U.S. regions through kernel density measures to determine resource availability (Jones et al., 2015). Jones and colleagues (2015) found that there were lower numbers of recreational facilities in ethnic minority tracts versus non-Hispanic White tracts. In another study, neighborhoods that were predominantly Hispanic/Black and Hispanic were 7-9 times more likely to completely lack at least one recreational facility in their neighborhood (measured via census tract) in comparison to non-Hispanic White neighborhoods (Moore et al., 2008).

Furthermore, these recreational facilities had fee-based resources that were less likely to be present in ethnic neighborhoods (Moore et al., 2008). They found that there was an equitable distribution of neighborhood parks but that recreational facilities were less equitably distributed among low income and ethnic minority neighborhoods (Moore et al., 2008).

In addition to lower quantities of recreational facilities, lower access to environmental amenities for physical activity in low income and ethnic minority compared to white communities in urban areas has been documented (Dahmann et al.,

2010; Moore et al., 2008; Sallis et al., 2011; Taylor et al., 2006; Thornton et al., 2016). A comprehensive study in Southern California found that cities within Los Angeles characterized as predominantly Asian and Hispanic had lower access to public

recreational facilities in comparison to white communities (Dahmann et al., 2010).

Additionally, in an analysis of physical activity facilities and information about participants in the National Longitudinal Study of Adolescent Health (1994-1995), individuals who lived in high-minority and low-educated block groups were half as likely to have at least one physical activity facility in comparison to low-minority and higher- educated block groups, measured by the number of recreational facilities within a 5-mile buffer (Gordon-Larsen et al., 2006).

Several of these studies did not assess the quality of recreational facilities or did not link how accessibility to resources impacts physical activity levels directly among residents living within those neighborhoods. Among ethnic minority women participating in an intervention, regardless of socioeconomic status measures by household income, the quality of physical activity resources was still directly related to physical activity level, indicating the need to invest in improving park quality among ethnic minority neighborhoods (Lee, Mama, Adamus-Leach, & Soltero, 2015).

Disparities in Microscale Built Environment Features

Examining microscale features of the environment can also help to understand the links between the built environment and physical inactivity and chronic diseases based on socioeconomic disadvantage (Lovasi, Neckerman, Quinn, Weiss, & Rundle, 2009; James

Sallis et al., 2011). However, macrolevel features that support walkability and physical activity in low income neighborhoods may not be strong indicators of high levels of chronic disease, but microscale features are more indicative of a poor built environment (

Sallis et al., 2011; Thornton et al., 2016; Yu, 2014). Microscale features related to neighborhood walkability, were found to be worse among low-income and high

concentrations of racial/ethnic minority neighborhoods even though individuals lived in highly walkable neighborhoods (Sallis et al., 2011; Thornton et al., 2016; Yu, 2014). In a study that examined microscale features in three different cities (Seattle/King County,

Baltimore, San Diego), the effect of pedestrian features varied by city and regions of the

US (Sallis et al., 2011; Thornton et al., 2016; Yu, 2014). Across all three regions, there were worse aesthetic and social features in low-income and high racial/ethnic minority neighborhoods in comparison to high income and White neighborhoods (Thornton et al.,

2016). Specifically, in Seattle and San Diego, there was less presence of aesthetic features such as art, fountains, sculptures and landscaping in high racial/ethnic minority neighborhoods (Thornton et al., 2016). Lower income neighborhoods were characterized as having more negative aesthetic and social elements that may hinder physical activity participation such as the presence of graffiti, litter, drug paraphernalia, broken windows, and poorly maintained buildings (Thornton et al., 2016). In addition, there were less trees present in Seattle and Baltimore neighborhoods among low-income neighborhoods

(Thornton et al., 2016). Among parks, incivilities that were related to less physical activity among ethnic minorities included dog refuse, litter and vandalism/graffiti

(Douglas et al., 2018).

Interestingly, other microscale features were found to be better in low-income and racial/ethnic minority residential neighborhoods (Thornton et al., 2016). For example, there were more sidewalks, fewer walking obstacles (inadequate street lighting, roll over curbs, absence of public transit stops) and more intersections among these neighborhoods in comparison to high income and low racial/ethnic minority neighborhoods (Thornton et al., 2016). A plausible reason for this finding may be that those residents were in older

parts of the city, thus, these areas were originally built around walking as the main source of transportation (Thornton et al., 2016). This study did not measure the relationship between these microscale features and physical activity levels of residents, which future studies should aim to conduct.

In summary, there is strong evidence to suggest that lower levels of physical activity may be influenced by environmental determinants of health among urban communities of color. These inequities present a social and environmental justice issue, directly affecting disadvantaged groups and can lead to disparities in health behaviors engagement and chronic disease development (Gordon-Larsen et al., 2006). Thus, the research among ethnic minorities such as Asian immigrants, is important to look further into to determine the social injustices that may exist within these communities.

Built Environmental Predictors of Physical Activity in Asian Americans

The relationship between built environment and physical activity is less clear among Asian Americans, as only a handful of studies have been conducted. Older built environment studies had included Asian Americans as an ethnic group, but many did not include a representative sample. Furthermore, studies that investigated the impact of the built environment in Asian Americans also looked at different physical activity correlates outside of environmental factors, such as demographic, psychological, social/cultural including neighborhood social cohesion within the same study (Bungum et al., 2012).

Among those studies assessing the built environment among Asian Americans, few used validated measures or scales utilized widely in the field. Nonetheless, existing studies have demonstrated large heterogeneity in findings based on ethnic group, study location

and built environment components (Bungum et al., 2012; Kao et al., 2016; Kelley et al.,

2016; Li et al., 2015; Osypuk et al., 2009; Wen et al., 2007).

Perceived neighborhood environment was assessed in a study conducted in 14 different Asian Pacific Islander (API) groups residing in Las Vegas, Nevada. In this study, 8-items were used to assess environmental factors, including presence of sidewalks, facilities and parks, crime, motorized traffic, trust of neighbors, unleashed dogs, and closeness to a grocery store and overall neighborhood pleasantness (Bungum et al., 2012). These items were then loaded on two different components: neighborhood safety and environmental physical activity supports (Bungum et al., 2012).

Environmental physical activity supports contributed to 7% of the variance in physical activity level. These components of the built environment that suggest high mixed land use and high walkability of neighborhoods can help to promote physical activity among

Asian Americans but may be heavily influenced by policies to change the built environment (Bungum et al., 2012).

The Multiethnic Study of Atherosclerosis (MESA) is the largest cohort study conducted in ethnic minorities that examined how living in ethnic enclaves not only affects dietary and physical activity behaviors of Chinese immigrant adults, but also how features of these enclaves are related to the built environment (Osypuk et al., 2009).

Living in ethnic enclaves can directly affect lifestyle behaviors, such as having access to healthier foods which leads to better quality diets that are low in fat and processed food, and culturally specific physical activity facilities which can promote higher levels of physical activity (Osypuk et al., 2009). They recruited 711 Chinese adults from the New

York and Los Angeles metropolitan areas. Chinese immigrants living in higher

concentration immigrant neighborhoods reported the environment for physical activity to be worse, including having lower safety, walkability and lower presence of recreational exercise facilities defined as facilities within one mile of their home (Osypuk et al.,

2009).

A similar study, known as the MASALA study, investigated how neighborhood walkability impacted walking for transport among individuals of South Asian ancestry in

Chicago and San Francisco, which includes individuals from India, Pakistan, Bangladesh,

Nepal or Sri Lanka (Kelley et al., 2016). Higher neighborhood walkability, as measured by the Walk Score, was positively associated with increased walking for transport in men but not women (Kelley et al., 2016). More specifically, in men, every 10 point increase in

Walk Score had increased the number of minutes walking for transport by 13.2 minutes, even after controlling for important sociodemographic and clinical variables (Kelley et al., 2016). It appears gender may be a powerful moderator in the relationship between the built environment and physical activity level and may have to do with varying physical activity level in men versus women due to individual and cultural related factors (Kelley et al., 2016). Nonetheless, this was one of the first studies in an Asian American population to use objective built environment measures, the Walk Score, versus perceived self-report of the built environment by participants.

Li and colleagues (2015) conducted a study to understand neighborhood factors and their impact on walking behavior in older Asian Adults from the 2003 California

Health Interview Survey (CHIS) (Li et al., 2015). One factor they examined was the availability of recreational facilities, in which they found that the association between proximity to a park or playground was positively associated with walking among Chinese

and Koreans but not Vietnamese Americans (Li et al., 2015). They also examined whether neighborhood safety affected walking but did not find it to have an impact on individuals in their sample. This study did not differentiate between walking for transport and walking for leisure, which can strongly impact one’s behavior, and they only used single-item measures for assessing neighborhood safety and resource access. A study of the same CHIS data conducted in Asian American adults found no associations between availability of parks, playgrounds or open spaces and walking (Wen et al., 2007).

In general, study findings investigating aesthetic related built environment characteristics find that individuals who perceive their environment as aesthetically pleasing and as having less incivilities (e.g., attractive scenery, less crime, vandalism, etc.,) are more likely to exhibit higher physical activity levels. Kao and colleagues (2016) compared Asian Americans to other ethnic groups in Houston, Texas to investigate how perceived neighborhood incivilities impacted physical activity levels. Neighborhood incivilities, including violence/crime, stray animals, waste, or fumes showed minimal evidence in affecting physical activity levels among the diverse Asian American sample in this study (Kao et al., 2016). The presence of recreational programs and facilities as part of the built environment for leisure related physical activity tailored to immigrants has rarely been studied. Lee and Stodolska (2017) suggest that socioeconomic status and ethnicity should also serve as identification of a marker of disadvantage, in which Asians living in low-income neighborhoods will experience poorer access to the availability and accessibility of recreational resources (Lee & Stodolska, 2017).

Determinants of Physical Activity among Asian Americans

Sufficient physical activity is especially important as a modifiable risk factor for chronic disease for Asian American groups as there is clear evidence to suggest that they are at a higher risk of certain chronic diseases compared to other ethnic groups. Physical activity rates are low and rates are mainly derived from state or local data with a lack of national level data to illustrate disparities in physical activity among Asian Americans.

Several studies on this disparity have been mainly from the western US, with some studies in states, such as Texas, New York, Nevada and Michigan with major cities that constitute a significantly high population of Asian Americans. A recent study conducted using National Health and Nutrition Examination Survey (NHANES) found a decline in physical activity levels in Asian Americans (Kalra et al., 2019). Among one of the largest population based metropolitan studies was one that compared physical activity levels among Asian Americans in Los Angeles, California and New York City, New York utilizing the CHIS and New York City Community Health Survey (NYC CHS), respectively (Yi et al., 2015). Non-Hispanic Asians exhibited the lowest odds of meeting the CDC Physical Activity Guidelines in NYC (OR=1.0) compared to non-Hispanic whites (OR=1.35, p<0.01), Non-Hispanic blacks (OR=1.61, p<0.001), and Hispanics

(OR=2.14, p<0.001) and Los Angeles (OR=1.0) compared to non-Hispanic whites

(OR=1.45, p=0.03), non-Hispanic blacks (OR=1.28, p=0.34; , and Hispanic (OR=1.71, p=0.03) groups (Yi et al., 2015). Within the metropolitan Philadelphia and southern new

Jersey areas, the rate of physical activity among these three groups was only 19.4%

(Patterson et al., 2016). There is a lack of awareness regarding low physical activity levels among this population.

Physical activity behavior among Asian Americans has mainly focused on elucidating predictors of sociodemographic, cultural and immigration patterns (Kandula

& Lauderdale, 2005). Higher education level, income level, and male gender are associated with higher engagement in physical activity and LTPA among Asian

Americans (Kandula & Lauderdale, 2005; Yi et al., 2015). With regards to acculturative factors, studies have found that Asian immigrants increase their levels of LTPA and physical activity in general as they have lived longer in the US and become more acculturated and if they experience upward social mobility (Kandula & Lauderdale, 2005;

Kao et al., 2016; Yi et al., 2016). Furthermore, foreign Born Asians are less likely to engage in regular LTPA compared to US born Asians (Kandula & Lauderdale, 2005).

When comparing the selected Asian American groups, Chinese, Korean and Vietnamese, evident differences exist in predictors of physical activity behavior. For example,

Vietnamese Americans report more occupational based physical activity than Chinese,

Korean and South Asian adults (Kandula & Lauderdale, 2005; Kao et al., 2016). A population based study disaggregated Asian subgroups, including Chinese, Filipino,

South Asian, Japanese, Korean and Vietnamese groups (Becerra et al., 2015).

Bilingualism, older age were associated with higher physical activity among Chinese

(Becerra et al., 2015). Among Koreans, female gender was significantly associated with less odds of meeting the ACSM physical activity guidelines (Becerra et al., 2015).

Among Vietnamese, bilingualism and living above the poverty level were associated with meeting the guidelines (Becerra et al., 2015). Several of these studies face limitations as they are based on self-report, are non-comparison between Asian and Non-Asian groups and not incorporating Chinese, Korean and Vietnamese ethnic groups warrant further

study into interactive effects between demographics, cultural factors and immigrant related factors on physical activity in addition to how the built environment affects their physical activity.

Summary

The literature review presented is an overview of built environment components and physical activity behavior. Strong cross-sectional evidence in built environmental predictors for physical activity behavior are illustrated among studies presented. The

SEM is a multifaceted model proposed to explain multilevel factors associated with physical activity, made up of individual, interpersonal, environment, policy, and global levels. There are vast psychological impacts associated with green exercise specifically versus exposure to other physical activity environments.

The built environment within the realm of physical activity constitutes neighborhoods conducive to physical activity such as walkability and parks and recreational facilities. The literature review also touches upon psychological and built environmental measures that are highly validated. The NEWS is a measure of perceived neighborhood walkability and how one perceives density of resources, mixed-land use, crime, traffic, and other measures of walkability. Objective built environment measures have primarily constituted primary spatial data and field observations, geospatial analyses and online and commercial sources to compute accessibility to physical activity resources and neighborhood walkability. Online sources that have been developed to provide a convenient and simple way to measure aspects of neighborhood walkability include the

Walk Score. Several differences lie in strengths and limitations among these methods.

Inequalities in the built environment that persist among low-income and ethnic communities of color, known to be an issue of environmental justice. The quality of parks and park size/acreage have been documented to be lower in ethnic minority neighborhoods, rather than the number of parks. On the contrary, lower number of recreational facilities for engaging in physical activity as well as lower quality resources are evident in these neighborhoods. Higher incivilities were reported among neighborhoods with higher ethnic minority composition.

Few studies have been conducted on the impact of the built environment among

Asian Americans on physical activity level. Thus, further studies are needed in this area of research. Asian Americans are the largest growing ethnic group in the US, experiencing biculturalism and unique ethnic identity as well as acculturative experiences when disaggregated by country of origin. In 2009, The Initiative on Asian Americans and

Pacific Islanders was introduced as an executive order by former President Obama to improve health care access and overall quality of life among this population (The White

House, 2009). As several Asians are foreign born, this study will fill a gap in identifying how the environment impacts behavior among Asian immigrants as well as US born

Asians, highlighting the potential disparities that may exist in meeting physical activity guidelines. The following chapter will explore the methods to conducting this dissertation.

CHAPTER 3

METHODS

The study was designed to determine whether objective and perceived built environmental characteristics are associated with physical activity levels and to examine whether psychosocial ethnic constructs impact the relationship between built environment and physical activity among Asian Americans living in an urban environment.

Research Design

The study was a cross-sectional design and utilized multiple levels of data sources including primary and secondary data. Primary data collected included participant self- reported data collected using a Built Environment and Physical Activity Survey and secondary data measured characteristics of neighborhoods that were related to physical activity such as parks and recreational facilities, as well as objective built environment measures that corresponded to neighborhood walkability.

Research Sites

Participants were recruited from community-based organizations (CBOs) associated with Chinese, Korean, and Vietnamese communities. The selected CBOs were located within Philadelphia County and within the selected zip codes of participants’ residence.

Figure 3-1. Philadelphia County Zip Codes (ArcMap 10.4)

Participants

A total of 241 participants were recruited during the study period. Participants were eligible to participate in the study if they met the following criteria: 1) self-

identified as being Chinese, Korean, or Vietnamese, 2) currently lived in Philadelphia

County, and 3) were between the ages of 20-69 years old (Appendix A). Data was collected from participants who attended the CBOs located within Philadelphia. The identified CBOs are established collaborating partners of the Center for Asian Health,

Lewis Katz School of Medicine of Temple University and Temple University Student

Organizations. One participant did not meet the eligibility criteria of residing in

Philadelphia County and was excluded from the analysis. Thus, a total of 240 participants were included in the study.

Instrumentation

Primary Data Measures

Modified Neighborhood Environment Walkability Scale - Abbreviated (NEWS-A).

A modified version of the NEWS-A was administered to participants (Appendix C).

NEWS-A was used to assess the perception of neighborhood design features that are related to being physically active. There were 54 items with 12 subscales presented on a

Likert scale (see Table 3-1). Responses ranged on a scale from 1 to 5 (none to all) for residential density, 1 to 5 (1-5 min to 31+ min) for land use mix-diversity, and 1 to 4

(none to all; strongly disagree to strongly agree) for the remaining subscales that included land use mix-access, street connectivity, infrastructure for safety and walking, aesthetics, traffic hazards, crime, streets not having many cul-de-sacs, physical barriers to walking, parking difficult in local shopping areas, and hilly streets in neighborhood. The reliability of the original scale has ranged from 0.58 to 0.90. Higher values on each subscale indicate a more favorable value of the environmental characteristic (Saelens et al., 2003).

Table 3-1. Neighborhood Environment Walkability Scale

# of Subscale Answer responses items 1-Residential Density 6 None, a few, some, most, all 1-5 min, 6-10 min, 11-20 min, 21- 2-Land use mix- diversity 23 30min, 31+min, don’t know 3-Land use mix- access 3 Strongly disagree -- strongly agree 4-Street connectivity 2 Strongly disagree -- strongly agree 5-Infrastructure and safety for walking 6 Strongly disagree -- strongly agree 6-Aesthetics 4 Strongly disagree -- strongly agree 7-Traffic hazards 3 Strongly disagree -- strongly agree 8-Crime 3 Strongly disagree -- strongly agree 9-Lack of many cul-de-sacs 1 Strongly disagree -- strongly agree 10-Physical barriers to walking 1 Strongly disagree -- strongly agree 11-Parking difficult in local shopping areas 1 Strongly disagree -- strongly agree 12-Hilly streets in the neighborhood 1 Strongly disagree -- strongly agree

The NEWS subscales were calculated using the scoring protocol from Sallis and

Colleagues (2003). For the residential density subscale, the following equation was provided to determine the weighted density relative to households per unit area based on a single-family detached residence:

Residential density= Single-family detached + (12 * row houses/townhouses 1-3 stories)

+ (10 * apts/condos 1-3 stories) + (25 * apts/condos 4-6 stories) + (50 * apts/condos 7-

12 stories) + (75 * apts/condos 13+ stories) (Saelens, Sallis, Black, et al., 2003).

For the land-use mix diversity subscale, responses were reverse scored and ‘do not know’ was coded as a ‘1’ since it was most likely that the distance to that facility was greater than 31 minutes, also coded as ‘1.’ The remaining subscales were scored regularly. The

NEWS was scored by obtaining the mean score for each subscale. A higher score in the subscale was indicated as higher neighborhood walkability for the following subscales: 1-

Residential Density, 2-Land use mix- diversity, 3-Land use mix- access, 4-Street

connectivity, 5-Infrastructure and safety for walking, 6-Aesthetics, 9-Lack of many cul- de-sacs, 11-Parking difficult in local shopping areas. On the contrary, a higher score in the subscale was indicated as lower neighborhood walkability for the following subscales: 7-Traffic hazards, 8-Crime, 10-Physical barriers to walking, and 12-Hilly streets in the neighborhood.

In addition, a composite neighborhood walkability score was calculated by converting the subscale scores to z-scores and averaging them for a total score (Nichani et al., 2019). Before obtaining the z-scores for the lower walkability subscales, they were reverse coded to match the other subscales (i.e. higher score indicated higher neighborhood walkability).

Neighborhood Physical Activity Questionnaire (NPAQ). The NPAQ measured recreational and transport-related walking and cycling within a 10-15 minute walk of one’s home and outside of one’s neighborhood and an overall level of physical activity in a usual week (Giles-Corti et al., 2006) (Appendix D). The broad categories included walking/cycling for transport, walking/cycling for recreation, health, or fitness, and total walking and physical activity. Walking for transport was measured through recalling the frequency and duration of walking that occurs within and outside of the neighborhood.

Walking for recreation, health, or fitness gathered information about the total frequency and duration of walking within and outside the neighborhood. Other LTPA’s were also measured, which capture information about moderate and vigorous intensity activity.

Total walking, cycling, vigorous, and moderate and physical activity were also converted into metabolic equivalents (METs) using the Compendium of Physical Activity

(Ainsworth et al., 2011). Reliability measures using Kappa coefficients ranged from 0.67

to 0.95 for the NPAQ, and total METs physical activity had a reliability of 0.82 (Giles-

Corti et al., 2006).

Table 3-2. Activities Converted into METs.

Activity METs/minute Walking for transport 3.5 Cycling for transport 4.3 Walking for recreation/health/fitness 6.8 Cycling for recreation/health/fitness 8.0 Moderate activity 4.0 Vigorous activity 8.0

Physical activity outcomes included the number of times per week and frequency

(hours/minutes per week) of: 1) walking within one’s neighborhood for transportation and recreation/health/fitness, 2) walking outside of one’s neighborhood for transportation and recreation/health/fitness, 3) cycling within one’s neighborhood for transportation and recreation/health/fitness, 4) cycling outside of one’s neighborhood for transportation and recreation/health/fitness, 5) moderate activity, 6) vigorous activity, 7) Total minutes of physical activity and 8) Total physical activity expressed in METs. Active transportation was also calculated as a measured indicated by walking and/or cycling for the purposes of active transport. Furthermore, walking and cycling for recreation was also measured. For the objective measures, moderate to vigorous physical activity was combined to create moderate to vigorous physical activity (MVPA).

Total Physical Activity

Walking/Cycling Other physical activity

Within the Outside of the Moderate Activity neighborhood neighborhood

Active transportation Active transportation Vigorous activity

For For recreation/health/fitness recreation/health/fitness

Figure 3-2. Breakdown of Physical Activity Domains based on the NPAQ.

Neighborhood Self-Selection (Appendix E). Neighborhood-self-selection is defined as lifestyle preferences that impact both the location and travel choice among individuals (Frank et al., 2007). Neighborhood self-selection is a form of bias in conducting neighborhood studies. Three questions were asked regarding one’s decision to live in their current neighborhood based on affordability, ease of walking, and low transportation costs (Frank et al., 2007). The average score of the three responses were calculated.

Ethnic Identity. Ethnic identity is a measure of self-concept, characterized as how individuals strongly associate themselves to the behaviors, knowledge, and awareness of cultural beliefs and traditions of one’s ethnic group (Iwamoto & Liu, 2010). These lived experiences related to race and ethnicity have led to health disparities in chronic disease and lifestyle behavior, and more specifically have been shown to affect physical activity levels among Asians (Brown et al., 2016). The level of ethnic identity was a continuous variable in assessing the impact of the built environment on physical activity. The

Revised Multi-Group Ethnic Identity Measures (MEIM-R) was used to examine ethnic identity among the study sample (Phinney & Ong, 2007) (Appendix F). It is a 6-item questionnaire with answers ranging on a 5-point scale from 1 (strongly disagree) to 5

(strongly agree) (Phinney & Ong, 2007) . Components of the questionnaire include exploration of (items 1, 4, and 5) and commitment to (items 2, 3, and 6) one’s ethnic identity (S. D. Brown et al., 2014). The Cronbach’s alpha was reported to be 0.81 to 0.89

(Phinney & Ong, 2007; Yoon, 2011). The MEIM-R was scored using a total ethnic identity score and the average scores among the two subscales, exploration of ethnic identity and commitment to one’s ethnic identity.

Demographics. Information about participant’s current address (street address, apartment/suite number, city, zip code), number of years lived at current address, country of birth, yearly income, employment status, education level, gender, and English proficiency was collected (Appendix G).

Body Mass Index (BMI). Body mass index was measured using self-reported height in feet/inches and/or centimeters and weight in pounds and/or kilograms without shoes. BMI was divided into cup points recommended for Asian Americans by the World

Health Organization (WHO) (WHO Expert Consultation, 2004). The BMI cut offs were as followed: underweight was less than 18.5, 18.5-22.9 was normal, 23-24.9 was overweight, and greater than equal to 25 was obese (Jih et al., 2014; WHO Expert

Consultation, 2004). (Appendix G).

Health Conditions. History of health conditions was collected by asking participants whether they were ever told by a physician that they had any of the following conditions: Hypertension/high blood pressure, diabetes/high blood sugar, high cholesterol, heart attack/coronary heart disease, stroke, sleep apnea, obesity, asthma, arthritis (rheumatoid, gout, fibromyalgia), and any type of cancer. (Appendix G).

Secondary Data Measures

Built Environment Walkability. Neighborhood Walkability is a composite walkability metric that was assessed in this study using the Walk Score. Walk Score is a web-based assessment tool that is publicly available online at www.walkscore.com. The Walk Score algorithm calculates a score based on the network distance (shortest distance to a destination along a network of routes) to different types of facilities: educational (e.g., schools), retail (e.g., grocery), food (e.g., restaurants), recreational (e.g., gym, park) and entertainment (e.g., theater), that are weighted and summed based on a linear combination to a specific address or location. Furthermore, the score is adjusted based on street-network characteristics, including population density, block length and intersection density (Walk Score, 2019b). A score from 0 to 100 was produced, with 0 indicating the lowest walkability (car dependent) and 100 being the highest (most walkable) (Walk

Score, 2019b). This measure was used to describe the characteristics of neighborhoods surrounding the study sample and describe macroscale features of the built environment

that impact walkability. Walk scores were obtained for each participant’s address within their zip codes. In addition, transit and bike scores were also collected as provided on the website interface. Each participant’s address obtained from the built environment and physical activity survey was entered as locations into the website interface.

Table 3-3. Walk Score Categories (Hirsch et al., 2013) (1) 0–24=very car-dependent: almost all errands require a car (2) 25–49=car-dependent: a few amenities within walking distance (3) 50–69 =somewhat walkable: some amenities within walking distance (4) 70–89 =very walkable: most errands can be accomplished on foot (5) 90–100 =walker’s paradise: daily errands do not require a car

Figure 3-3. Sample Walk Score Result

Proximity to Physical Activity Resources and Parks. Locations of private and public facilities for measuring the accessibility to physical activity resources were identified using geospatial data, a commercial database called InfoUSA, and through google search

(ESRI, Redlands, CA, USA). InfoUSA is an online database that provides a comprehensive list of businesses, including the name, location, and industry classification code (Standard Industrial Classification (SIC) system and North American Industry

Classification System (NAICS)) (ESRI, 2018). Sources of data that contribute to the database include directory listings (Yellow Pages, business white pages), annual reports,

10-K and Securities and Exchange Commission (SEC) information, government data, business magazines, newsletters or newspapers, and the US Postal Service (ESRI, 2018).

Businesses were verified by telephone. The following NAICS codes were applied to search the InfoUSA database are presented in Table 3-4. Examples of corresponding index entries for each code included: ‘712190- parks, nature, provincial parks, etc.’ and

‘713940- fitness centers, exercise centers, athletic club facilities, recreational sport club facilities, etc.’ We excluded resources relating to schools and churches as they were found to be insignificant venues that were associated with physical activity levels; one study reported that only 2% or less of adults engaged in exercise in these settings (Moore et al., 2008).

Table 3-4. North American Industry Classification System Codes for Identifying Physical Activity Resources using InfoUSA

Sector Code Category Arts, Entertainment, and 712190 Nature Parks and Other Similar Recreation (71) Institutions 713910 Golf Courses and Country Clubs 713920 Skiing Facilities 713940 Fitness and Recreational Sports Centers 713950 Bowling Centers 713990 All Other Amusement and Recreation Industries

In order to ensure that all recreational facilities were captured, a google search was conducted to determine all of the facilities, which included the following terms: gyms, recreation center, fitness clubs, yoga studios, martial arts gyms, CrossFit, and dance studios. Additional facilities were identified following this method. The following categories were developed for physical activity facilities and parks: 1) Gym and recreation center; 2) Specialty exercise facilities, which included yoga and dance studios,

Muy Thai/boxing gyms, clubs (tennis, golf), CrossFit, Pilates, and other specialized activities; 3) Parks, represented by areas of green space that individuals utilized for recreation.

Spatial data for parks and trails were obtained from the Pennsylvania Spatial Data

Access (PASDA), a publicly available geospatial database established by Pennsylvania

State University (PASDA, 2015). A shapefile of Pennsylvania local parks was obtained, which indicated the boundaries of local parks and line data for trails within Pennsylvania

(PASDA, 2015). This data was prepared by the Pennsylvania Department of

Conservation and Natural Resources.

Procedures

Researcher Training

Researchers collecting data for the neighborhood variables were trained using a protocol to identify physical activity resources through online sources. Research staff helping with conducting the bilingual surveys were trained to answer questions regarding the survey items (e.g., if a participant asks to clarify the meaning of a survey question).

Since the proposed study had multiple research procedures, quality control measures were be taken to ensure proper data collection.

Translation of Instruments

All instruments (excluding secondary data sources) were translated into Chinese,

Korean, and Vietnamese languages and was administered depending on the ethnicity and

English fluency of the participant. Surveys were translated by bilingual community staff members at the Center for Asian Health. All translated surveys were back translated to

English to ensure validity.

Participant Recruitment

Participants were recruited through the aid of research team members fluent in

Chinese, Korean, and Vietnamese. Participants were recruited from CBO’s arranged in affiliation with the Center for Asian Health and from events set up specifically for data collection within communities between July to December 2019. Data was collected from a total of 17 CBO’s within Philadelphia County. CBO’s were located in various neighborhoods within Philadelphia County.

Informed Consent

Participants were asked to read the Temple University IRB approved consent form before participating (Appendix B). Participants’ home address was collected on the survey for the purposes of conducting the environmental audits as well as identifying within what zip code they live. No other identifying information was collected. Consent forms and surveys were stored in a locked file steel cabinet in the Center for Asian

Health, Lewis Katz School of Medicine. Only the research team was able to have access to the survey data.

Data Collection

Data was collected in person from participants between July to December 2019.

Participants completed the survey consisting of demographics, BMI, health conditions, neighborhood self-selection, NEWS-A, ethnic identity, NPAQ, and PPAQ via paper and pencil. Each questionnaire took approximately 25 minutes to complete. Participants received $10 for completing the survey.

Secondary data sources were utilized to measure neighborhood level factors, which was measured concurrently with collecting participant surveys. Walk Score and secondary data of recreational facilities and parks were collected for each zip code within

Philadelphia County. Proximity to physical activity resources and parks were linked to respondents’ residential addresses using GIS on a vector data model.

Data Entry and Cleaning

Survey data was entered in Microsoft Excel and under a password protected file only shared with the research team. Participant survey data Participant’s addresses were cleaned to ensure they were correct before geocoding and analyzing geospatial data.

While entering participant’s addresses into the website interface to obtain Walk Scores, addresses which did not output a score were checked using Google Maps and corrected accordingly. The participant’s reported street names were checked using Google Maps to confirm they existed, and corrected any misspelt or incorrect addresses onto the data entry spreadsheet where collected data was entered for each participant.

GIS Analysis

In order to calculate physical activity resources for each participant to incorporate into the final data analysis model, geospatial data was imported into ArcMap 10.7.1

(Environmental Systems Research Institute, Inc., 2020). Prior to geocoding, an address locator was created from a reference data table in the form of a shapefile containing

Philadelphia county street line data. Geospatial data was cleaned to match proper GIS formatting before importing into ArcMap. Participant’s addresses were geocoded into

ArcMap as point data, mapped to latitude and longitude coordinates, in order to conduct the spatial analyses. Any addresses which were not successfully geocoded were re- matched, or manually entered into ArcMap using the editor tool by matching them on streets based on the Google Map image. Second, the physical activity facilities obtained from InfoUSA and online were also geocoded as point data. Parks were presented as polygon data (PASDA, 2015). A centroid was drawn within the middle of each park as a point for calculating the proximity measures.

The program used to conduct the proximity measures was the ArcGIS network analyst extension. More specifically, the ‘find nearest tool’ was applied, which uses line distance or travel modes using time or distance as the measure for nearest features (ESRI,

2020). The walking distance, walking time, and driving time were measured separately for the nearest park, gym/recreation center, and specialty exercise facility to a person’s home address. The walking distance follows roads for pedestrian traffic and optimizes travel distance, measured in meters. The walking time also follows roads for pedestrian traffic and is measured by the walking speed of 5 kilometers per hour. The driving time incorporates the movement of automobiles on the roads while considering traffic rules.

Below is the description of each measure that was calculated. 95

Table 3-5. GIS Measures using the Find Nearest Tool

Travel Mode Near Feature Park Driving time Gym/recreational center Specialty exercise facility Park Walking time Gym/recreational center Specialty exercise facility Park Walking Distance Gym/recreational center Specialty exercise facility

The proximity measures within the attribute tables was imported into STATA 16 for analysis and merged with the participant survey data as well as other secondary measures.

Data Analysis

A priori power analysis was conducted using G*Power 3.1 to determine the sample size needed to detect a level of effect for the outcome variables in the study, with the NPAQ as the main outcome variable. Predictor and covariate variables included perceived neighborhood walkability, proximity to a park and physical activity facility, physical activity level, demographics, individual Walk Score, and ethnic identity which is illustrated in Table 3-6. The outcome variable was physical activity level and examined different domains.

Descriptive and univariate analysis included mean and interquartile range, standard deviation, variance, and histograms. Normality and assumption tests were conducted to ensure covariate and outcome variables were normally distributed as well as appropriate for the regression models. Bivariate analyses included the Kruskal-Wallis equality of populations rank test, and spearman rank correlation were used for continuous

variables. Chi-square tests and frequencies were used for categorical variables. All data cleaning, univariate and bivariate testing was conducted in SAS 9.4 and STATA 16.

The outcome data, physical activity level, contained an excessive zero count data of outcome variables which includes structural zeros (those who do not exhibit any physical activity), and random zeros due to sampling variability in which those who exhibit the behavior may produce a zero outcome (He et al., 2014). Furthermore, the variance exceeded the mean for each outcome variable which indicated over-dispersion of data. Figure 3-4 and Table 3-6 illustrate the non-normal distribution of outcome variables which demonstrate the excessive zeros.

Figure 3-4. Histogram of Distribution of Outcomes

Table 3-6. Frequency of Participants Reporting no Physical Activity

Activity % Active transportation 41.84 Recreational walking/cycling 42.44 Walking for transport 31.91 Cycling for transport 80.35 Walking for recreation/health/fitness 61.85 Cycling for recreation/health/fitness 84.55 Moderate activity 85.96 Vigorous activity 75.22 Overall physical activity 23.43

To take into account the non-normal distribution and excessive zeros, zero- inflated negative binomial (ZINB) regression models were used to test each research question and hypothesis. This model contains two regression equations, the logit part,

which includes structural zeros (physical activity=0) versus those who engage in physical

activity, and count part models which include positive counts only (physical activity > 0)

(Lambert, 1992).

−휇 (1 − 휋) + 휋푒 , 푓표푟 푦푖 = 0 −휇 푦푖 Pr(푌푖 = 푦푖) = { 푒 휇 휋 , 푓표푟 푦푖 = 1,2,… 푦푖!

Each model was controlled for with covariates related to demographics. Data was

analyzed using STATA 16 with the procedure, zinb (StataCorp LLC, 2020). In order to

determine whether the ZINB was an appropriate fit for conducting the analysis, the

dispersion parameter produced in the output was examined to determine whether there

was over-dispersion, in which the case of ZINB would be an appropriate fit over the plain

negative binomial model. Huber-white sandwich estimators were utilized to correct for

the clustered sample design, participants residing in varying zip codes, and estimates the

variance of the maximum likelihood estimator and robust standard errors within the

ZINB models (Huber, 1967; White, 1980). The ‘vce(cluster clustvar)’ option was used to

apply the estimator (StataCorp LLC, 2020).

Table 3-7. Study Variables

Data Source (Year) Characteristics Physical Activity and Built Environment perceived neighborhood walkability Survey (2019) physical activity level demographics ethnic identity

Online (2019) walk score Online and GIS (2019) physical activity facilities and parks

For RQ1, the independent variable included perceived neighborhood walkability, controlling for age, gender, income, and education. For RQ2, the independent variables included perceived neighborhood walkability and ethnic identity. For RQ3, Walk Score

(by individual address) was the independent variable, controlling for age, gender, income, neighborhood self-selection, and BMI. For RQ4, the independent variable was proximity to physical activity facilities and parks, while controlling for age, gender, income, Asian

BMI, perceived crime, and neighborhood self-selection.

Table 3-8. Variables by Research Question

RQ1 Model RQ2 Model RQ3 Model RQ4 Model Perceived Perceived Objective Proximity to neighborhood neighborhood Neighborhood Physical Activity walkability walkability Walkability (Walk Facilities and Parks Interaction term: Score) Ethnic Identity Covariates: age, Covariates: age, Covariates: age, Covariates: age, gender, income, gender, income, gender, income, gender, income, education education, English BMI, Asian BMI, Fluency neighborhood neighborhood self- self-selection selection, perceived crime

IRB Approval

IRB was submitted to Temple University Institutional Review Board and was approved in April 22nd, 2019. Informed consent was obtained from participants in

English, Mandarin, Korean or Vietnamese. The informed consent forms are presented in the appendices.

100

CHAPTER 4

RESULTS

The primary purpose of this study was to explore how physical activity among

Asian Americans is influenced by perceived and objective characteristics of the built environment. The results include descriptive participant characteristics, bivariate analysis of built environment predictor variables, as well as regression models testing the hypotheses for the four research questions articulated in Chapter 1. The analytic results were interpreted and are discussed with respect to each research question and collectively. Study strengths and limitations are detailed. Recommendations for future research in the area are made and the clinical implications of the findings are discussed.

Descriptive Statistics

Demographic Characteristics

A total of 240 participants comprised the study sample. Participants were recruited from 17 community based organizations in Philadelphia County in Southeastern

Pennsylvania. The mean age of participants was 47 years (SD 13.63) and more than half were female (58.33%). Almost half (44%) had a college degree or more, followed by having graduated high school (36.86%), not completed high school (11.02%), and had an elementary education or lower (7.63%). The largest percentage of participants had an annual income of less than $20,000 (36.32%), followed by $30,000-39,999 (30.34%),

$20,000-29,999 (20.51%), and $40,000 or greater (12.82%). The majority were currently employed (64.35%). Most had been living at their current residence for 10 years or more

101

(44.02%), followed by less than 5 years (29.91%), and between 5-9 years (26.07%).

These and other demographic characteristics are presented below in Table 4-1.

Table 4-1. Demographic Characteristics of Study Participants

Demographic characteristics MeanSD or Range IQR1 n (%) Age 47.18 (13.63) 20-68 23.0 Neighborhood self-selection 2.84 (0.54) 1-4 0.67 Gender Female 140 (58.33) Male 100 (41.67) Education level Elementary school 18 (7.63) Below high school 26 (11.02) High school graduate 87 (36.86) College and above 105 (44.49) Annual income <$20,000 85 (36.32) $20,000-29,999 48 (20.51) $30,000-39,999 71 (30.34) ≥$40,000 30 (12.82) Employment status Currently employed 148 (64.35) Not currently employed 82 (35.65) Years lived at current residence <5 years 70 (29.91) 5-9 years 61 (26.07) ≥10 years 103 (44.02) 1Interquartile range

Neighborhood self-selection was assessed by a 3-item scale to determine the reasons in which participants moved to their neighborhood. Neighborhood self-selection is a potential bias in research on the built environment, as individuals may move to highly walkable neighborhoods by choice and, as a result, be more physically active (Frank et al., 2007). Neighborhood self-selection was a covariate used to examine the relationship between the objective measures of the built environment. The neighborhood self-

102

selection mean score was 2.84 (SD 0.54), with a range of 1 to 4. Higher scores suggest that participants moved to their neighborhood for affordability, ease of walking, and low transportation costs.

Health Characteristics

Health conditions of study participants were collected by self-report. BMI was calculated using height and weight and categorized into Asian BMI categories, which have lower cut-offs as compared to the categories typically used with European-

Americans (Jih et al., 2014; WHO Expert Consultation, 2004). The Asian BMI categories are: underweight is less than 18.5 kg/m2, 18.5-22.9 kg/m2 is a normal body weight, 23-

24.9 kg/m2 is overweight, and greater than or equal to 25 kg/m2 is considered obese (Jih et al., 2014; WHO Expert Consultation, 2004). Half of the participants (50.00%) were within the normal Asian BMI range, with the other half being in the overweight

(30.17%), obese (15.09%), and underweight (4.74%) ranges. The top three most commonly reported health conditions were high cholesterol (27.08%), hypertension/high blood pressure (16.67%), and diabetes/high blood sugar (11.67%). Less than one tenth of the sample reported that they were told by their health care professional they had arthritis

(7.92%), obesity (4.17%), sleep apnea (1.67%), asthma (1.67%), cancer (1.67%), heart attack/coronary heart disease (0.83%), and stroke (0.42%).

103

Table 4-2. Health Characteristics of Study Participants

Health Characteristics n (%) Body Mass Index (Asian) Underweight (18.5) 11 (4.74) Normal (18.5-22.9) 116 (50.00) Overweight (23-24.9) 70 (30.17) Obese (25) 35 (15.09) Health condition Hypertension/high blood pressure Yes 40 (16.67) No 200 (83.33) Diabetes/high blood sugar Yes 28 (11.67) No 212 (88.33) High cholesterol Yes 65 (27.08) No 175 (72.92) Heart attack/Coronary heart disease Yes 2 (0.83) No 238 (99.17) Stroke Yes 1 (0.42) No 239 (99.58) Sleep apnea Yes 4 (1.67) No 236 (98.33) Obesity Yes 10 (4.17) No 230 (95.83) Asthma Yes 4 (1.67) No 236 (98.33) Arthritis - rheumatoid, gout, fibromyalgia Yes 19 (7.92) No 221 (92.08) Cancer Yes 4 (1.67) No 236 (98.33) Cancer type Brain cancer 2 (50%) Breast cancer 2 (50%)

104

Cultural Characteristics of Participants

The acculturative characteristics of participants are found in Table 4-3.

Participants were born in China (32.90%), Vietnam (32.03%), and Korea (29.44%), followed by Taiwan (2.16%), and Hong Kong (0.43%). Only 3% were born in the United

States. With respect to language proficiency, the majority self-reported that they did not speak English well (63.09%), followed by well (19.74%), not at all (12.45%), and fluently (4.72%).

Ethnic identity is a construct used to measure identity to one’s ethnic group and known to influence health behaviors, specifically documented in Asian Americans

(Brown et al., 2016). A greater score on the MEIM-R indicated greater levels of ethnic identity. The mean level of ethnic identity in the study sample was 3.61 (SD 0.84), which suggested that participants generally exhibited high levels of ethnic identity. The mean level of the subscale exploration of one’s ethnic identity (3.82  0.87) was slightly higher than the subscale of commitment to one’s ethnic identity (3.41  0.93).

Table 4-3. Acculturative Characteristics of Study Participants Cultural characteristics n (%) Ethnic Group Chinese 83 (34.58) Korean 74 (30.83) Vietnamese 83 (34.58) Country of birth China 76 (32.90) Hong Kong 1 (0.43) Korea 68 (29.44) Taiwan 5 (2.16) United States 7 (3.03) Vietnam 74 (32.03) Self-reported English Fluency Not at all 29 (12.45) Not well 147 (63.09) Well 46 (19.74) Fluent 11 (4.72)

105

Table 4-3 (continued) Ethnic Identity meanSD range Exploration of one’s ethnic identity1 3.41 (0.93) 1.00-5.00 Commitment to one’s ethnic identity2 3.82 (0.87) 1.67-5.00 Total Ethnic Identity Score 3.61 (0.84) 1.67-5.00 1Items 1, 4, 5; 2Items 2, 3, 6

Geographic Distribution of Participants

Participant addresses were obtained to evaluate GIS-assessed characteristics of the built environment (i.e. gyms/recreation centers, specialty exercise facilities, parks), as well as to calculate objective neighborhood walkability (i.e. Walk Score). The geographic distribution of participants within Philadelphia County are presented below. Most participants resided in zip codes 19120 (20.83%), 19149 (16.67%), and 19145 (9.58%).

Figure 4-1 displays the count of participants that resided from each zip code. The blue zip codes represent where the most participants resided, and the yellows indicates zip codes with zero participants.

Table 4-4. Geographic Distribution of Participants by Zip Code Zip code n(%) Zip code n (%) 19102 1 (0.42) 19134 2 (0.83) 19103 1 (0.42) 19135 4 (1.67) 19104 4 (1.67) 19136 7 (2.92) 19107 6 (2.50) 19138 1 (0.42) 19111 9 (3.75) 19140 5 (2.08) 19114 2 (0.83) 19141 5 (2.08) 19115 2 (0.83) 19142 3 (1.25) 19120 50 (20.83) 19145 23 (9.58) 19121 9 (3.75) 19147 3 (1.25) 19122 14 (5.83) 19148 18 (7.50) 19124 9 (3.75) 19149 40 (16.67) 19125 1 (0.42) 19151 1 (0.42) 19126 9 (3.75) 19152 9 (3.75) 19127 1 (0.42) 19154 1 (0.42)

106

Figure 4-1. Map of Geographic Distribution of Participant's by Zip Code

Built Environment –Neighborhood Walkability

Perceived neighborhood walkability was measured using the Neighborhood

Environment Walkability Scale (NEWS), which consists of 12 subscales and a total

107

composite neighborhood walkability score. A higher score on each subscale suggests higher neighborhood walkability, except for the subscales crime, traffic hazards, hilly streets in the neighborhood, and other physical barriers to walking, in which a higher score denoted lower walkability (Saelens, Sallis, Black, et al., 2003). The mean score for residential density was 272.18 (SD 97.02, range 173-735), suggesting that participants, on average, lived in areas of lower residential density. Land use mix diversity was 2.98

(SD 0.75), indicating participants lived about a walking distance of 11-20 minutes away from facilities such as shops, grocery stores, and recreational centers, on average. The remaining subscales have a range of 1 to 4. The subscales denoting higher neighborhood walkability were hilly streets in one’s neighborhood (reverse scored: 1.61  0.79), physical barriers to walking (reverse scored: 1.49  0.75), street connectivity (3.12 

0.46), land use mix access (3.07  0.73), infrastructure and safety for walking/cycling

(2.94  0.45), lack of cul-de-sacs (2.62  0.81), crime (reverse scored: 2.25  0.69), traffic hazards (reverse scored: 2.51  0.43), parking difficulty in local shopping areas

(2.49  0.88), and aesthetics (2.33  0.63).

Objective neighborhood walkability was measured using the Walk Score. The mean Walk Score for the study sample was 82.54 (SD 10.21) out of 100, suggesting that participants generally lived in neighborhoods that were highly walkable. With respect to the distribution of Walk Scores based on categorizes created, the majority of participants resided in neighborhoods that were “very walkable” (65.83%), followed by “walker’s paradise” (22.08%), “somewhat walkable” (8.33%) and “car-dependent” (3.75%). The

Walk Score was used as a continuous variable in the regression models in Research

Question 3.

108

Table 4-5. Descriptive Built Environment Characteristics

Built Environment Characteristic meanSD range Perceived Measures Neighborhood Walkability (composite) -0.0001 (0.36) -1.12-0.97 Residential density 272.18 (97.0) 173.00-735.00 Land use mix-diversity 2.97 (0.75) 1.13-5.00 Land use mix – access 3.07 (0.73) 1.00-4.00 Street connectivity 3.12 (0.46) 2.00-4.00 Infrastructure and safety for walking 2.94 (0.45) 1.83-4.00 Aesthetics 2.33 (0.63) 1.00-4.00 Traffic hazardsª 2.51 (0.43) 1.33-4.00 Crimeª 2.25 (0.69) 1.00-4.00 Parking difficult in local shopping areas 2.49 (0.88) 1.00-4.00 Lack of cul-de-sacs 2.62 (0.81) 1.00-4.00 Hilly streets in neighborhoodª 1.61 (0.79) 1.00-4.00 Physical barriers to walkingª 1.49 (0.75) 1.00-4.00 Objective measures n (%) Walk Score Car-dependent (0–49) 9 (3.75) Somewhat walkable (50–69) 20 (8.33) Very walkable (70–89) 158 (65.83) Walker’s paradise (90-100) 54 (22.08) meanSD range Total Walk Score 82.54 (10.21) 40-99 Transit Score 68.97 (12.07) 45-100 Bike Score 67.93 (10.46) 36-98 ªreverse scored: a higher score denotes lower neighborhood walkability

Built Environment- GIS Measures

Objective assessment of the built environment were measured using GIS. A total of 308 parks were identified in Philadelphia County from the Pennsylvania Local Parks database (PASDA, 2015). With regard to physical activity facilities, there were a total of

252 gym/recreation centers and 218 other specialty exercise facilities in the county as of

December 4th, 2019. Specialty exercise facilities consisted of yoga and dance studios,

Muy Thai/boxing gyms, clubs (tennis, golf), Crossfit, Pilates, and other specialized activities. The nearest gym/recreation center was an average of 1030.83 (SD 37.97) meters away. The mean driving time from participants’ homes to the nearest

109

gym/recreation center was 2.68 (SD 0.08) minutes and the mean walking time was 11.92

(SD 0.43) minutes. The distance to the nearest park was 525.21 (SD 21.26) meters from the participant’s home. The mean driving time to the nearest park was 1.47 minutes (SD

0.06), and walking time was 5.96 (SD 0.22) minutes. Lastly, specialty exercise facilities had slight less accessibility, illustrating a total distance of 1135.39 (SD 47.58) meters away from the participant’s home, mean driving time of 3.06 (SD 0.11) minutes, and walking time of 13.63 (SD 0.57) minutes.

Table 4-6. GIS Measures of Park and Physical Activity Facility and Park Proximity

Recreation Environment n Count of Parks (Total) 308 Count of gyms/recreation center 252 Count of other recreational facilities 218 Gym/recreation center accessibility measures meanSD Driving time 2.68 (0.08) Walking time 11.92 (0.43) Distance 1030.83 (37.97) Park accessibility measures Driving time 1.47 (0.06) Walking time 5.96 (0.22) Distance 525.21 (21.26) Specialty exercise facility accessibility measure Driving time 3.06 (0.11) Walking time 13.63 (0.57) Distance 1135.39 (47.58)

Self-Reported Physical Activity

Participants reported the duration and level of physical activity they engaged in over the past week using the Neighborhood Physical Activity Questionnaire (Giles-Corti et al., 2006). Activity type was reported in hours per week. The mean number of hours spent engaging in active transportation (walking and cycling to and from places) was 1.80

(SD 5.54) hours and recreational walking/cycling was 2.63 (SD 7.52) hours. Participants

110

reported a mean of 2.82 (SD 9.41) hours and 1.02 (SD 2.50) hours of walking within and outside of one’s neighborhood for transport and recreation/health/fitness, respectively.

The average number of hours spent cycling within and outside of one’s neighborhood for transport and recreation/health/fitness was 0.32 (SD 0.91) and 0.33 (SD 0.97) hours, respectively. On average, the study sample engaged in 0.35 hours (SD 1.34) of moderate and 0.77 (SD 1.75) hours of vigorous activity. The average hours spent engaging in all forms of physical activity was 5.50 (SD 10.55) hours, which included a combination of walking and cycling, as well as moderate and vigorous physical activity (MVPA).

Gender differences were explored. Men reported significantly more cycling within the neighborhood compared to women (0.48±1.24 v. 0.21±0.58 hours, p=0.03).

Otherwise, there were no statistically significant differences between men and women in the duration of active transportation, recreational walking/cycling, walking inside and outside of the neighborhood, cycling outside of the neighborhood, moderate activity, vigorous activity, and total physical activity.

In addition, ethnic differences were investigated. Vietnamese Americans reported engaging in the greatest hours of recreational walking/cycling (4.44±11.61), walking within the neighborhood (4.95±15.17), vigorous activity (1.11±1.80), and total physical activity (7.70±15.06). Chinese Americans reported the greatest hours of active transportation (2.42±3.27), cycling within (0.57±1.21) and outside (0.54±1.31) of the neighborhood, walking outside of the neighborhood (1.63±3.06), and moderate activity

(0.66±1.97). There were statistically significant differences in physical activity levels by ethnic group for all outcomes except vigorous activity. Results are presented in Table 4-

111

Table 4-7. Duration of Physical Activity among Participants and by Gender and Ethnic Group

Total Male Femal Chinese Korean Vietnames e e Variable MeanSD Mea Mean p* Mean Mean Mean p* n Active Transportation 1.80 (5.54) 2.30 1.43 0.38 2.42 0.83 2.05 <0.00 Recreational walking/cycling 2.62 (7.52) 2.72 2.57 0.21 2.41 0.84 4.44 <0.00 Walking within one’s neighborhood 2.82 (9.41) 3.36 2.44 0.30 2.21 1.08 4.95 <0.00 Cycling within one’s neighborhood 0.32 (0.91) 0.48 0.21 0.03 0.57 0.00 0.34 0.00 Walking outside of one’s 1.02 (2.50) 0.87 1.13 0.37 1.63 0.61 0.90 <0.00 neighborhood Cycling outside of one’s 0.33 (0.97) 0.40 0.27 0.53 0.54 0.00 0.40 0.02 neighborhood Moderate activity 0.35 (1.34) 0.19 0.45 0.92 0.66 0.21 0.14 0.04 Vigorous activity 0.77 (1.75) 0.89 0.69 0.52 0.48 0.73 1.11 0.20 Total Physical activity 5.50 6.03 5.11 0.50 5.96 2.51 7.70 <0.00 (10.55) *A Kruskal-Wallis equality of populations rank test was used

112

Physical activity level was also calculated in METs, defined as a measure of energy expenditure based on the resting metabolic rate (Ainsworth et al., 2011). Total physical activity among the study sample was 1624.95 METs (SD 2765.22). The highest number of METs resulted from walking within one’s neighborhood for transport and recreation/health/fitness (671.53  2269.31), vigorous intensity activity (371.30 

837.74), and walking outside of one’s neighborhood for transport and recreation/health/fitness (248.27  616.54).

Table 4-8. Physical Activity Level of Study Participants in Metabolic Equivalents

Variable Total MeanSD Walking within one’s neighborhood 661.52 (2269.3) Cycling within one’s neighborhood 144.38 (413.5) Walking outside of one’s neighborhood 248.27 (616.5) Cycling outside of one’s neighborhood 148.61 (436.0) Moderate activity 82.95 (321.0) Vigorous activity 371.30 (837.7) Total Physical activity 1624.95 (2765.2)

Bivariate Analysis

Separate bivariate analyses were conducted using Spearman Rank correlations and to assess the relationship between the perceived neighborhood walkability subscales and each of the built environment variables. As seen in Table 4-9, several NEWS subscales were significantly correlated with each other, with the highest correlations found among hilly streets and physical barriers (ρ=0.59, p<0.00), followed by street connectivity and land use mix access (ρ=0.55, p<0.00). In Table 4-10, the correlations between built environment variables including perceived and objective measures were analyzed. The Walk Score was correlated with several built environment measures, with the highest correlations among walking time and distance (ρ=1.00, p<0.00) to a specialty

113

exercise facility, (ρ=1.00, p<0.00), and park (ρ=1.00, p<0.00). Driving time and walking time to the nearest gym/recreation center (ρ=0.94, p<0.00), specialty exercise facility

(ρ=0.94, p<0.00), and park (ρ=0.93, p<0.00) were highly correlated. NEWS and the Walk

Score were moderately correlated (ρ=0.24, p<0.00). These relationships suggest that variables of the built environment were highly correlated and, thus, should be evaluated separately within regression models.

114

Table 4-9. Correlation Matrix of NEWS Subscales (Spearman Rank) A B C D E F G H I J K L A. Residential density 1.00 B. Land use mix-diversity 0.15* 1.00 C. Land use mix – access -0.07 0.41* 1.00 D. Street connectivity -0.03 0.19* 0.51* 1.00 E. Infrastructure/safety -0.15* 0.04 0.27* 0.44 1.00 F. Aesthetics 0.02 0.05 0.10 -0.06 0.26** 1.00 G. Traffic hazardsª 0.12 0.24* 0.15* 0.22* -0.03 0.09 1.00 H. Crimeª 0.02 -0.07 -0.06 0.14* -0.02 -0.29* 0.09 1.00 I. Parking difficultly 0.12 0.23 0.25* 0.03 -0.19 -0.01 0.20* -0.09 1.00 J. Lack of cul-de-sacs -0.16* -0.18* -0.04 0.09 0.03 -0.04 0.01 0.11 0.00 1.00 K. Hilly streetsª -0.03 -0.04 -0.08 0.13 0.08 -0.04 0.00 0.37* -0.19* 0.16* 1.00 L. Physical barriersª 0.11 0.02 -0.11 0.02 -0.05 0.15 0.10 0.17 -0.19* 0.24* 0.59* 1.00 *Significant at p<0.05; ªreverse scoring

Table 4-10. Correlation Matrix of Built Environment Variables (Spearman Rank) Variable 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 1. NEWS Total 1.00 2. Driving time - gym /rec -0.06 1.00 3. Walking time- gym/rec -0.11 0.94* 1.00 4. Distance to gym/rec -0.11 0.94* 1.00* 1.00 5. Driving time – park -0.05 -0.04 -0.09 -0.09 1.00 6. Walking time- park -0.08 0.00 0.02 0.02 0.93* 1.00 7. Distance to park -0.08 0.00 0.02 0.02 0.93* 1.00* 1.00 8. Driving time- specialty 0.02 0.42* 0.39* 0.40* -0.11 -0.11 -0.11 1.00 9. Walk time- specialty -0.04 0.44* 0.47* 0.47* -0.16* -0.10 -0.10 0.94* 1.00 10. Distance to specialty fac -0.03 0.44* 0.47* 0.47* -0.16* -0.10 -0.10 0.94* 1.00* 1.00 11. Walk Score 0.24* -0.42* -0.55* -0.55* 0.24* 0.08 0.08 -0.42* -0.56* -0.56* 1.00 *Significant at p<0.05; ªreverse scoring

115

Analysis of Research Question 1

Research Question 1 examined whether Asian residents with higher self-reported neighborhood walkability reported higher levels of active transportation, recreational walking/cycling, and higher levels of engagement in physical activity in general. The results are presented by zero-inflated negative binomial (ZINB) regression models, incidence rate ratios examining perceived neighborhood walkability subscales, as well as their relationship to active transportation, recreational walking/cycling, and overall physical activity. In addition, ZINB models and incidence rate ratios are presented for overall composite neighborhood walkability (average scores of perceived neighborhood walkability subscales) and various types of physical activity.

Neighborhood Walkability and Active Transportation

Active transportation was defined as engagement in walking and/or cycling for the purposes of transport from one place to another; this is in contrast to transport specifically for recreational purposes. Each perceived neighborhood walkability subscale was included in separate ZINB regression models while controlling for age, gender, and income. The beta coefficients of each perceived neighborhood walkability subscale were included in in the count and logit (binary) models with respect to active transportation, presented to three decimal places. The count model only included participants who reported engagement in activity (PA>0); the binary model is a logistic regression model comparing those who reported no physical activity (PA=0) versus those who did engage in physical activity (PA=1).

116

For the count model, among all NEWS subscales, physical barriers to walking

(freeways, railways, rivers, and other structures) (β=-0.321, p=0.010) and a lower number of cul-de-sacs (suggesting greater neighborhood walkability) (β=-0.482, p=0.003) were associated with fewer weekly hours of active transportation. Land use mix-access was marginally associated with weekly hours of engaging in active transportation (β=0.639,

0.061). Other variables represented by the subscales of the measure were not statistically significant predictors in the models. For the binary model, residential density (β=0.316, p<0.001), physical barriers to walking (β=23.368, p<0.001), and a lower number of cul- de-sacs (β=8.358, p<0.001) were associated with a greater odds of not engaging in active transportation. Thus, those who perceived their neighborhoods to have high residential density, more cul-de-sacs, and fewer physical barriers to walking were more likely to engage in active transportation. Physical barriers to walking, such as rivers, railways, and freeways, had the strongest magnitude of association. Overall, neighborhood walkability was not associated with active transportation.

Table 4-11. ZINB Regression Models for Perceived Neighborhood Walkability Subscales and Active Transportation

β SE p-value Predictor

Count Models§ Residential density 0.001 0.001 0.242 Land use mix-diversity 0.438 0.301 0.146 Land use mix – access 0.639 0.342 0.061 Street connectivity -0.672 0.450 0.881 Infrastructure and safety for walking -0.184 0.192 0.336 Aesthetics 0.065 0.370 0.860 Traffic hazardsª 0.086 0.253 0.732 Crimeª 0.204 0.179 0.256 Parking difficult in local shopping areas 0.301 0.173 0.083 Lack of cul-de-sacs -0.482 0.161 0.003 Hilly streets in neighborhoodª 0.182 0.284 0.521 Physical barriers to walkingª -0.321 0.124 0.010

117

Table 4-11 (continued) Predictor β SE p-value Overall neighborhood walkability 0.194 0.399 0.628 Binary Models§ Residential density 0.316 0.010 <0.001 Land use mix-diversity -0.349 1.762 0.843 Land use mix – access -1.474 2.065 0.475 Street connectivity -4.142 4.351 0.341 Infrastructure and safety for walking -1.731 1.282 0.177 Aesthetics -1.784 2.553 0.485 Traffic hazardsª 0.584 0.834 0.484 Crimeª -0.979 1.389 0.481 Parking difficult in local shopping areas 0.138 0.917 0.881 Lack of cul-de-sacs 8.358 1.736 <0.001 Hilly streets in neighborhoodª -0.252 3.677 0.945 Physical barriers to walkingª 23.368 0.976 <0.001 Overall neighborhood walkability -3.278 1.897 0.084 β: exponentiated beta coefficient; SE: standard error [robust] § Model is adjusted for age, gender, and income ªreverse scored: a higher score denotes lower neighborhood walkability

The incidence rate ratios (IRR) represent the log ratio of the beta coefficients (β) presented in the count part of the ZINB models and depict the magnitude of the outcome variable associated with a one-unit change in the independent predictor. A one unit increase in having greater physical barriers to walking was associated with a 27% decrease in engaging in active transportation (p<0.001). Moreover, a one unit increase in fewer cul-de-sacs was associated with a 39% decrease in engaging in active transportation (p<0.001). A forest plot and table represent the IRR’s and 95% confidence intervals for each NEWS subscale.

118

0 0.5 1 1.5 2 2.5 3 3.5 4

Residential density Land use mix-diversity Land use mix – access Street connectivity Infrastructure and safety for walking Aesthetics Traffic hazardsª Crimeª Parking difficult in local shopping areas Lack of cul-de-sacs Hilly streets in neighborhoodª Physical barriers to walkingª Overall neighborhood walkability Incidence Rate Ratio

Figure 4-2. IRRs of NEWS Subscales and Active Transportation Table 4-12. IRRs of NEWS Subscales and Active Transportation 95% CI 95% CI IRR Predictor (Lower) (Upper)

Count Models§ Residential density 1.002 1.004 0.999 Land use mix-diversity 1.549 2.796 0.859 Land use mix – access 1.895 3.703 0.970 Street connectivity 0.935 2.258 0.368 Infrastructure and safety for walking 0.831 1.211 0.571 Aesthetics 1.067 2.203 0.517 Traffic hazardsª 1.090 1.789 0.665 Crimeª 1.742 1.176 0.863 Parking difficult in local shopping areas 1.350 1.898 0.962 Lack of cul-de-sacs 0.617* 0.847 0.450 Hilly streets in neighborhoodª 1.200 2.096 0.687 Physical barriers to walkingª 0.726* 0.568 0.926 Overall neighborhood walkability 1.121 0.555 2.651 IRR= incidence rate ratio; 95% CI: confidence interval § Model is adjusted for age, gender, and income ªreverse scored: a higher score denotes lower neighborhood walkability

119

Neighborhood Walkability and Recreational Walking/Cycling

Walking and cycling for the purposes of recreation, health, or fitness also was of interest. Each perceived neighborhood walkability subscale was entered into separate regression models while controlling for age, gender, and income. The beta coefficients of perceived neighborhood walkability subscales were presented for the count and logit models. For the count model, land use mix-diversity (β =0.574, p<0.001) and access

(β=0.500, p=0.045) were positively associated with a greater amount of walking and/or cycling for recreation among those who were physically active. Higher perceived crime

(β=-0.444, p=0.008) and fewer cul-de-sacs (β=-0.378, p=0.001) were associated with less weekly hours of recreational walking and/or cycling. Overall neighborhood walkability was not associated with greater hours of recreational walking/cycling.

Within the binary model comparing those who did and did not engage in physical activity, greater perceived land use mix diversity (β=-14.531, p=0.011), land use mix access (β=-8.167, p=0.007), and parking difficulty in local areas (β=-2.25, p=0.011) decreased the odds of not engaging in recreational walking/cycling. In contrast, residential density increased the odds of engaging in recreational walking/cycling

(β=0.641, p<0.001). Higher overall neighborhood walkability was associated with lower odds of not engaging in recreational walking/cycling (β=-4.61, p=0.013).

Table 4-13. ZINB Regression Models for Perceived Neighborhood Walkability Subscales and Recreational Walking and Cycling

β SE p-value Predictor

Count Models§ Residential density 0.001 0.001 0.618 Land use mix-diversity 0.574 0.135 <0.001 Land use mix – access 0.500 0.250 0.045 Street connectivity -0.254 0.313 0.935

120

Table 4-13 (continued) Predictor β SE p-value Infrastructure and safety for walking 0.328 0.213 0.124 Aesthetics 0.380 0.243 0.118 Traffic hazardsª -0.316 0.319 0.322 Crimeª -0.444 0.165 0.008 Parking difficult in local shopping areas -0.064 0.124 0.607 Lack of cul-de-sacs -0.378 0.118 0.001 Hilly streets in neighborhoodª 0.030 0.125 0.812 Physical barriers to walkingª 0.130 0.148 0.380 Overall neighborhood walkability -0.030 0.348 0.931 Binary Models§ Residential density 0.641 0.198 <0.001 Land use mix-diversity -14.531 5.743 0.011 Land use mix – access -8.167 3.017 0.007 Street connectivity -0.804 2.089 0.700 Infrastructure and safety for walking -1.676 0.943 0.076 Aesthetics -4.512 3.779 0.233 Traffic hazardsª 0.607 1.093 0.579 Crimeª -0.175 1.039 0.866 Parking difficult in local shopping areas -2.254 0.884 0.011 Lack of cul-de-sacs 2.138 1.822 0.241 Hilly streets in neighborhoodª -3.112 6.230 0.617 Physical barriers to walkingª -0.729 0.705 0.301 Overall neighborhood walkability -4.605 1.863 0.013 β: exponentiated beta coefficient; SE: standard error [robust] § Model is adjusted for age, gender, and income ªreverse scored: a higher score denotes lower neighborhood walkability

The incidence rate ratios (IRR) of the beta coefficients (β) are presented for the count model within the ZINB model. A one unit increase in land use mix diversity and access was associated with a 78% (p<0.001) and 65% (p=0.045) increase in recreational walking/cycling, respectively. A one unit increase in perceived crime was associated with a 35.8% decrease in recreational walking/cycling (p=0.008). Furthermore, a one unit increase in the number of cul-de-sacs was associated with a 69% decrease in recreational walking/cycling (p=0.001). A forest plot and table which represent the IRR’s and 95% confidence intervals are presented for each NEWS subscale.

121

0 0.5 1 1.5 2 2.5 3 3.5 4

Figure 4-3. IRRs of NEWS and Recreational Walking and Cycling

Table 4-14. IRRs of NEWS Subscales and Recreational Walking and Cycling

95% CI 95% CI IRR Predictor (Lower) (Upper)

Count Models§ Residential density 1.000 0.988 0.999 Land use mix-diversity 1.775** 1.310 2.405 Land use mix – access 1.648* 1.010 2.688 Street connectivity 0.975 0.528 1.799 Infrastructure and safety for walking 1.388 0.914 2.109 Aesthetics 1.495 0.546 4.098 Traffic hazardsª 0.729 0.390 1.362 Crimeª 0.642* 0.462 0.891 Parking difficult in local shopping areas 0.938 0.735 1.197 Lack of cul-de-sacs 0.686* 0.544 0.865 Hilly streets in neighborhoodª 1.030 0.807 1.316 Physical barriers to walkingª 1.138 0.852 1.521 Overall neighborhood walkability 0.970 0.490 1.920 IRR= incidence rate ratio; 95% CI: confidence interval * p-value <0.05. ; ** p <0.001. § Model is adjusted for age, gender, and income ªreverse scored: a higher score denotes lower neighborhood walkability

122

Perceived Neighborhood Walkability Subscales and Physical Activity

Each perceived neighborhood walkability subscale was investigated in separate regression models while controlling for age, gender, income, and education. The beta coefficients of perceived neighborhood walkability subscales are presented for the count and logit models. For the count model, land use mix-diversity (β =0.253, p=0.011) and access (β=0.464, p=0.011) were positively associated with weekly hours of physical activity, while residential density (β=-0.002, p=0.036) and fewer cul-de-sacs (β=-0.520, p<0.001) was negatively associated with weekly hours of physical activity. The contributions of the constructs represented by the other subscales did not reach statistical significance.

For the binary model, land use mix diversity (β=-0.859, p<0.001) and access (β=-

1.362, p=0.001), aesthetics (β=-1.022, p=0.003), and parking difficulty in local areas (β=-

0.7188, p=0.001) decreased the odds of not engaging in any physical activity. Thus, the greater the land use mix, aesthetics, and parking difficulty, the more likely that the respondent was physically active. Having fewer cul-de-sacs in the neighborhood was associated with increased odds of not engaging in any physical activity (β=0.5287, p=0.011). Other NEWS subscales variables were not statistically significant in the ZINB models.

123

ZINB Regression Models for Perceived Neighborhood Walkability Subscales and Overall Physical Activity in METs.

β SE p-value Predictor

Count Models§ Residential density -0.002 0.009 0.036 Land use mix-diversity 0.253 0.099 0.011 Land use mix – access 0.464 0.183 0.011 Street connectivity 0.256 0.175 0.413 Infrastructure and safety for walking 0.126 0.192 0.509 Aesthetics 0.117 0.186 0.532 Traffic hazardsª -0.068 0.251 0.788 Crimeª -0.133 0.158 0.400 Parking difficult in local shopping areas 0.069 0.123 0.576 Lack of cul-de-sacs -0.529 0.106 <0.001 Hilly streets in neighborhoodª 0.016 0.162 0.923 Physical barriers to walkingª 0.101 0.112 0.365 Binary Models§ Residential density 0.001 0.003 0.841 Land use mix-diversity -0.859 0.250 <0.001 Land use mix – access -1.362 0.428 0.001 Street connectivity -0.037 0.378 0.929 Infrastructure and safety for walking -0.304 0.556 0.585 Aesthetics -1.022 0.348 0.003 Traffic hazardsª 0.254 0.365 0.486 Crimeª 0.316 0.464 0.495 Parking difficult in local shopping areas -0.719 0.218 0.001 Lack of cul-de-sacs 0.556 0.219 0.011 Hilly streets in neighborhoodª 0.898 0.266 0.735 Physical barriers to walkingª 0.096 0.339 0.778 β: exponentiated beta coefficient; SE: standard error [robust] § Model is adjusted for age, gender, income and education. ªreverse scored: a higher score denotes lower neighborhood walkability

The incidence rate ratios (IRR) produced by the beta confidents in the ZINB count model are presented below. A one unit change in land use mix diversity and access were associated with an 28% and 59% increase in physical activity, respectively

(p=0.014; p=0.011). A one unit change in residential density and smaller number of cul- de-sacs were associated with a 0.001% and 41% decrease in overall physical activity

124

levels, respectively (p<0.001). Presented below are a forest plot and table illustrating the

IRR’s and 95% confidence intervals for each NEWS subscale. 0 0.5 1 1.5 2 2.5

Incidence Rate Ratio

Figure 4-4. IRR of NEWS and Physical Activity Table 4-15. IRRs of NEWS Subscales and Physical Activity

95% CI 95% CI IRR Predictor (Lower) (Upper)

Count Models§ Residential density 0.998* 0.997 1.000 Land use mix-diversity 1.288* 1.060 1.565 Land use mix – access 1.590* 1.111 2.276 Street connectivity 1.292 0.917 1.820 Infrastructure and safety for walking 1.135 0.780 1.652 Aesthetics 1.124 0.780 1.619 Traffic hazardsª 0.935 0.572 1.528 Crimeª 0.876 0.876 1.193 Parking difficult in local shopping areas 1.071 0.842 1.362 Lack of cul-de-sacs 0.589** 0.478 0.726 Hilly streets in neighborhoodª 1.016 0.740 1.395 Physical barriers to walkingª 1.107 0.889 1.378 IRR= incidence rate ratio; 95% CI: confidence interval * p-value <0.05. ; ** p <0.001 § Model is adjusted for age, gender, income and education.

125

Perceived Neighborhood Walkability and Physical Activity

Regression models for total perceived neighborhood walkability are presented in relation to different physical activity domains comprising total physical activity, including walking/cycling in and out of neighborhood, moderate activity, and vigorous activity, when controlling for age, gender, and income (not reported in the tables). The beta coefficients of perceived neighborhood walkability subscales are presented for the count and logit models. In the sample, higher perceived neighborhood walkability was associated with lower levels of cycling outside of the neighborhood (β=-2.772, p=0.023) with the count model, while controlling for sociodemographics, gender, and age. With the binary model, greater neighborhood walkability decreased the odds of not engaging in cycling inside (β=-8.483, p=0.002) and outside of the neighborhood (β=-6.000, p=0.001), and total physical activity expressed in METs (β=-2.136, p=0.011), also controlling for sociodemographics. The results are presented below in Table 4-16.

126

Table 4-16. ZINB Regression Models for Overall Perceived Neighborhood Walkability and Physical Activity Subtypes

Total physical Walking in Walking outside Cycling in Cycling outside Vigorous Moderate activity neigh neigh neigh neigh activity activity (METs) β (SE) β (SE) β (SE) β (SE) β (SE) β (SE) β (SE) Predictor

Count Model§ Perceived Neighborhood 0.545 (0.366) 0.632 (1.075) -0.856 (0.869) -2.772 (1.223)* 0.248 (0.396) -1.587 (1.159) 0.045 (0.240) Walkability

Binary Model§ Perceived -2.136 Neighborhood -6.683 (4.970) 0.619 (5.622) -8.483 (2.715)* -6.000 (1.801)** -0.830 (0.433) -1.672 (1.648) (0.842)* Walkability Fit statistics χ2 (p) 11.420 70.980 2.690 8.780 73.570 29.810 9.200 (0.121) (<0.001**) (0.610) (0.032*) (<0.001**) (<0.001**) (0.419) AIC 642.811 594.384 296.885 231.803 462.029 247.656 2398.548 BIC 696.805 645.758 334.165 260.688 526.594 284.936 2465.600 β: exponentiated beta coefficient; SE: standard error [robust]; AIC: Akaike Information Criterion; BIC: Bayesian Information Criterion; χ2: chi- square; p: χ2 p-value. * P-value <0.05. ; ** p <0.001; § Model is adjusted for age, gender, income and education.

127

In summary, and with respect to the NEWS subscales, having more barriers to walking and less cul-de-sacs predicted fewer hours of engaging in active transportation.

Having fewer cul-de-sacs, more physical barriers, and higher residential density predicted lower odds of engaging in active transportation. Greater perceived land-use mix, lower perceived crime, and having more cul-de-sacs predicted greater hours of engaging in recreational walking and cycling. Lower perceived land use mix, having greater parking difficulty in shopping areas, and higher residential density were associated with lower odds of engaging in recreational walking/cycling. Greater overall neighborhood walkability was associated with a greater likelihood of engaging in recreational walking and cycling, as well as cycling inside and outside of the neighborhood.

Greater perceived land use mix, more cul-de-sacs, and lower residential density were associated with greater weekly METs of physical activity. Higher perceived land use mix, greater aesthetics, having more cul-de-sacs, and more parking difficulty also were associated with a greater likelihood of engaging in overall physical activity. Overall neighborhood walkability was associated with an increased likelihood of engaging in physical activity.

Analysis of Research Question 2

Research Question 2 investigated whether cultural factors influenced the relationship between perceived neighborhood walkability and physical activity. Ethnic identity was conceptualized as one’s level of self-identity as an Asian American, specifically Chinese, Korean, or Vietnamese (Phinney & Ong, 2007). As several studies have suggested, ethnic identity in highly immigrant countries such as the United States is linked to health-related outcomes and behaviors such as diet and physical activity (S. D.

128

Brown et al., 2016; Henley et al., 2016; Moise et al., 2018; Murayama et al., 2017). ZINB models were used to explore the interactive effects between ethnic identity and perceived neighborhood walkability as predictors of active transportation, recreational walking/cycling, and overall physical activity. Ethnic identity and perceived neighborhood walkability were inserted into the model as full factorials, which allowed for the investigation of both possible interactions and main effects.

Separate interaction plots were created to illustrate the effects by plotting the predicted outcomes from the regression equations for active transportation, recreational walking/cycling, moderate and vigorous activity, as well as physical activity in general.

Within these plots, the independent variables perceived neighborhood walkability and ethnic identity were standardized in order to interpret the beta coefficient as a change in the duration of active transportation, recreational walking/cycling and physical activity for every change of standard deviation (SD) of these independent predictors. Interaction plots display the predicted outcomes of activity duration and level by high (+1 SD) and low (-1 SD) perceived neighborhood walkability, as well as high (+1 SD) and low ethnic identity (-1 SD), while holding sociodemographic covariates constant at their mean values.

Ethnic Identity Interactions - Active Transportation

The first ZINB model examined the interaction of ethnic identity on the relationship between neighborhood walkability and active transportation. Within the count model, there was an interaction between neighborhood walkability and ethnic identity in predicting active transportation (β=-1.370, p=0.005). Both greater neighborhood walkability (β=4.767, p=0.009) and greater ethnic identity (β=0.608,

129

p=0.003) were associated with more hours of active transportation as main effects.

Within the binary model, an interaction between neighborhood walkability and ethnic identity was observed (β=-6.836, p=0.004), with neighborhood walkability being significantly associated with active transportation (β=19.164, p=0.0013). The AIC and

BIC indicated that the model was a good fit.

Table 4-17. Moderated ZINB Model of Ethnic Identity and Perceived Neighborhood Walkability in Predicting Active Transportation in Asian American Residents

β (SE) p-value Predictor

Count Model§ Neighborhood Walkability (NW) 4.767 (1.815) 0.009 Ethnic Identity (EI) 0.608 (0.201) 0.003 NWEI -1.370 (0.487) 0.005

Binary Model§ Neighborhood Walkability 19.164 (7.693) 0.013 Ethnic Identity -0.783 (0.725) 0.121 NWEI -6.836 (2.373) 0.004 Fit statistics AIC 633.714 BIC 710.345 β: exponentiated beta coefficient; SE: standard error [robust] AIC: Akaike Information Criterion; BIC: Bayesian Information Criterion § Model is adjusted for age, gender, income, education

Figure 4-6 represents the difference in levels of active transportation by perceived neighborhood walkability and ethnic identity. High perceived neighborhood walkability was associated with 3.39 fewer hours per week of engaging in active transportation among those with high ethnic identity. In contrast, high perceived neighborhood walkability was associated with 1.00 greater hours/week of active transportation for those with low ethnic identity.

130

Figure 4-5. Predicted Values of Neighborhood Walkability and Active Transportation by Ethnic Identity

Ethnic Identity Interactions- Recreational Walking/Cycling

The ZINB model examined the interactive effect of ethnic identity on the relationship between neighborhood walkability and recreational walking/cycling. Within the count model, there was no interaction between neighborhood walkability and ethnic identity in predicting recreational walking/cycling (β=-0.066, p=0.838). Neighborhood walkability (β=0.206, p=0.936) and ethnic identity (β=-0.029, p=0.904) were not associated with recreational walking/cycling. Within the binary model, ethnic identity did not interact with neighborhood walkability (β=-0.821, p=0.823). However, ethnic identity was associated with lower odds of not engaging in recreational walking (β=-3.356, p<0.001). The AIC and BIC indicated that the model was a good fit.

131

Table 4-18. Moderated ZINB Model of Ethnic Identity and Perceived Neighborhood Walkability in Predicting Recreational Walking/cycling in Asian American Residents

β (SE) p-value Predictor

Count Model§ Neighborhood Walkability 0.206 (2.545) 0.936 Ethnic Identity -0.029 (0.244) 0.904 NWEI -0.063 (0.601) 0.916

Binary Model§ Neighborhood Walkability -3.258 (10.996) 0.767 Ethnic Identity -3.356 (0.633) <0.001 NWEI -0.821 (3.673) 0.823 Fit statistics AIC 607.279 BIC 680.590 β: exponentiated beta coefficient; SE: standard error [robust] AIC: Akaike Information Criterion; BIC: Bayesian Information Criterion § Model is adjusted for age, gender, income, education

Levels of recreational walking/cycling by were examined through exploring the interaction between perceived neighborhood walkability and ethnic identity. This interaction was not statistically significant. High neighborhood walkability was associated with 0.13 fewer hours per week of recreational walking/cycling among those with high ethnic identity. Furthermore, high perceived neighborhood walkability resulted in 1.10 greater hours per week of recreational walking/cycling among those with low ethnic identity.

Ethnic Identity Interactions – Vigorous Intensity Physical Activity

The ZINB model examined the possible interaction of ethnic identity on the relationship between neighborhood walkability and vigorous intensity physical activity.

In the count model, there was no interaction between neighborhood walkability and ethnic identity in predicting vigorous intensity physical activity (β=-0.173, p=0.695).

132

Neighborhood walkability (β=0.992, p=0.595) and ethnic identity (β=0.028, p=0.767) were not significantly associated with vigorous physical activity as main effects. Within the binary model, there were no interactive effects between ethnic identity and neighborhood walkability with vigorous intensity physical activity (β=0.618, p=0.309).

The AIC and BIC indicated that the model was a good fit.

Table 4-19. Moderated ZINB Model of Ethnic Identity and Perceived Neighborhood Walkability in Predicting Duration of Vigorous Activity in Asian American Residents

β (SE) p-value Predictor

Count Model§ Neighborhood Walkability 0.992 (1.863) 0.595 Ethnic Identity 0.028 (0.094) 0.767 NWEI -0.173 (0.442) 0.695

Binary Model§ Neighborhood Walkability -2.800 (2.620) 0.285 Ethnic Identity -0.374 (0.202) 0.064 NWEI 0.618 (0.608) 0.309 Fit statistics AIC 2446.722 BIC 430.988 β: exponentiated beta coefficient; SE: standard error [robust]; AIC: Akaike Information Criterion; BIC: Bayesian Information Criterion § Model is adjusted for age, gender, income, education, and English fluency

The predicted differences in levels of vigorous intensity physical activity duration by examining the interaction between perceived neighborhood walkability and ethnic identity were measured. Among those with high ethnic identity, high perceived neighborhood walkability was related to 0.25 more hours per week of vigorous intensity physical activity. Among those with low ethnic identity, high perceived neighborhood walkability was related to 0.8 more hours per week of vigorous intensity physical activity. These results were not statistically significant.

133

Ethnic Identity Interactions – Moderate Intensity Physical Activity

The ZINB model examined the interaction of ethnic identity on the relationship between neighborhood walkability and moderate intensity physical activity. In the count model, there was a significant interaction between neighborhood walkability and ethnic identity in predicting moderate intensity physical activity (β=3.801, p=0.002).

Neighborhood walkability (β=-14.316, p=0.001) and ethnic identity (β=-1.821, p<0.001) were negatively associated with moderate physical activity as main effects. In the binary model, an interaction between perceived neighborhood walkability and ethnic identity was observed (β=4.282, p=0.008). Both perceived neighborhood walkability (β=-15.504, p=0.017) and ethnic identity (β=-2.981, p=0.032) decreased the odds of not engaging in moderate physical activity as main effects. The AIC and BIC indicated that the model was a good fit.

Table 4-20. Moderated ZINB Model of Ethnic Identity and Perceived Neighborhood Walkability in Predicting Duration of Moderate Physical Activity in Asian Americans

β (SE) p-value Predictor

Count Model§ Neighborhood Walkability -14.316 (4.310) 0.001 Ethnic Identity -1.821 (0.461) <0.001 NWEI 3.801 (1.254) 0.002

Binary Model§ Neighborhood Walkability -15.504 (6.48) 0.017 Ethnic Identity -2.981 (1.39) 0.032 NWEI 4.282 (1.62) 0.008 Fit statistics AIC 239.402 BIC 310.573 β: exponentiated beta coefficient; SE: standard error [robust] AIC: Akaike Information Criterion; BIC: Bayesian Information Criterion § Model is adjusted for income, education

134

The interaction plot in Figure 4-9 depicts the differences in levels of moderate intensity physical activity duration by examining the interaction between perceived neighborhood walkability and ethnic identity. High perceived neighborhood walkability was related to 4.8 more minutes per week of moderate intensity physical activity among those with high ethnic identity. Furthermore, high perceived neighborhood walkability was related to 20.4 fewer minutes per week of moderate intensity physical activity among those with low ethnic identity.

Figure 4-6. Predicted Values of Neighborhood Walkability and Moderate Intensity Physical Activity by Ethnic Identity

Ethnic Identity Interactions – Overall Physical Activity

The last ZINB regression model examined the interaction of ethnic identity on the relationship between neighborhood walkability and overall physical activity and displayed all covariates in the model. In the count model, there was no statistically

135

significant interaction between neighborhood walkability and ethnic identity in predicting overall physical activity (β=-0.356, p=0.396). Both greater neighborhood walkability

(β=1.508, p=0.366) and ethnic identity (β=0.017, p=0.913) were not significantly associated with overall physical activity expressed in METs. Within the binary model, there was no statistically significant interaction between ethnic identity and neighborhood walkability in predicting physical inactivity (β=0.018, p=0.981). Ethnic identity decreased the odds of not engaging in physical activity as a main effect in the model (β=-

1.360, p=0.003).

Table 4-21. Moderated ZINB Model of Ethnic Identity and Perceived Neighborhood Walkability in Predicting Overall Physical Activity (METs) in Asian Americans

Predictor β (SE) p-value Count Model Neighborhood Walkability 1.508 (1.666) 0.366 Ethnic Identity 0.017 (0.156) 0.913 Gender (ref. female) 0.080 (0.226) 0.722 Age -0.018 (0.010) 0.021 Education (ref. elem. school) Below high school 0.641 (0.419) 0.126 High school graduate 0.414 (0.549) 0.451 College and above 0.183 (0.454) 0.687 Income (ref. <$20,000) $20,000-29,999 0.107 (0.208) 0.607 $30,000-39,999 0.305 (0.251) 0.224 ≥$40,000 0.160 (0.472) 0.735 English Fluency (ref. Not at all) Not well -0.381 (0.335) 0.255 Well -0.372 (0.531) 0.457 Fluent -0.671 (0.527) 0.203 Neighborhood WalkabilityEthnic Identity -0.356 (0.419) 0.396 Binary Model β (SE) p-value Neighborhood Walkability -1.499 (2.610) 0.566 Ethnic Identity -1.360 (0.457) 0.003 Gender (ref. female) -0.141 (0.431) 0.743 Age 0.114 (0.031) <0.001 Education (ref. elem. school) Below high school -0.608 (0.831) 0.464 High school graduate 0.005 (0.718) 0.995 136

Table 4-21 (continued) Binary Model β (SE) p-value College and above -0.801 (1.191) 0.501 Income (ref. <$20,000) $20,000-29,999 0.277 (0.686) 0.686 $30,000-39,999 0.534 (0.513) 0.297 ≥$40,000 2.459 (0.695) <0.001 English Fluency (ref. not at all) Not well -0.154 (0.714) 0.829 Well 0.109 (0.998) 0.913 Fluent -18.693 (1.073) <0.001 Neighborhood WalkabilityEthnic Identity 0.018 (0.758) 0.981 Fit statistics AIC 2254.998 BIC 2330.815 β: exponentiated beta coefficient; SE: standard error [robust] AIC: Akaike Information Criterion; BIC: Bayesian Information Criterion § Model is adjusted for age, gender, income, education, and English fluency

The interaction plot (not shown) illustrates the differences in levels of overall physical activity in METs by examining the interaction between perceived neighborhood walkability and ethnic identity. Among those with high ethnic identity, high perceived neighborhood walkability was associated with a decrease of 67 METs of overall physical activity. Furthermore, among those with low ethnic identity, high neighborhood walkability was associated with an increase of 763 METs per week of overall physical activity.

In summary, the findings for this research question identified an interaction between ethnic identity and the built environment in accounting for self-reported weekly hours of active transportation. Those participants with high ethnic identity and high perceived neighborhood walkability reported 3.39 fewer hours of active transportation compared to individuals who reported low neighborhood walkability. Participants with low ethnic identity and high perceived neighborhood walkability accumulated 1 hour of active transportation more each week compared to those reported lower neighborhood 137

walkability. Furthermore, an interaction was found between ethnic identity and the built environment in accounting for weekly hours of moderate intensity physical activity.

Participants with high ethnic identity and high neighborhood walkability accumulated

0.08 greater hours per week of moderate intensity physical activity, whereas those with low ethnic identity and high perceived neighborhood walkability was related to 0.34 fewer hours per week. The relationship between the built environment and overall physical activity was not moderated by level of ethnic identity of participants.

Results of Research Question 3

Research Question 3 investigated whether Asian residents living in neighborhoods with higher Walk Scores reported greater levels of active transportation, recreational walking/cycling, and overall physical activity. The Walk Score is a measure of objective neighborhood walkability, which was obtained for participants’ residential addresses. The ZINB regression models included the Walk Score as the independent variable. Sociodemographics such as age, gender, income, BMI, and neighborhood self- selection were included as covariates in the model. Higher neighborhood self-selection scores indicated that participants chose to live in their neighborhood due to affordability, ease of walking, and low transportation costs (Frank et al., 2007). The results are presented by ZINB regression models with beta coefficients, IRRs, and odds ratios for each count and binary model, respectively.

Walk Score and Active Transportation

The first ZINB model examined whether a greater Walk Score predicted engaging in active transportation, while controlling for theoretically relevant covariates. Within the

138

count model, the Walk Score (β=0.015, IRR=1.015, p=0.396) was not associated with weekly hours of active transportation. Among the covariates, higher neighborhood self- selection scores were positively associated with weekly hours of active transportation

(β=0.625, IRR=1.869, p=0.022). Other covariates in the model did not reach statistical significance.

Within the binary model, participants with a higher Walk Score were more likely to not engage in active transportation (β=0.205, OR=1.228, p=0.024). Furthermore, every one unit increase in the Walk Score was associated with a 22.8% greater odds of not engaging in active transportation. Older participants (β=0.590, OR=1.804, p=0.001), those with an income of ≥$40,000 (β=8.890, OR=7262.3, p=0.009), and those with a higher BMI (β=0.010, OR=1.010, p<0.001) were more likely to not engage in active transportation. Furthermore, participants who scores high on neighborhood self-selection

(β=-3.017, OR=0.049, p=0.029) were less likely to report engaging in active transportation.

Table 4-22. ZINB Regression Models for the Walk Score and Active Transportation

β (SE) IRR p-value Predictor

Count Model Gender (ref. female) 0.240 (0.317) 1.272 0.448 Age -0.116 (0.142) 0.988 0.415 Income (ref <$20,000) $20,000-29,999 0.362 (0.468) 1.436 0.439 $30,000-39,999 -0.975 (0.392) 0.907 0.803 ≥$40,000 -0.393 (0.744) 0.675 0.598 BMI 0.003 (0.001) 1.003 0.054 Neighborhood Self-Selection 0.625 (0.273) 1.869 0.022 Walk Score 0.015 (0.018) 1.015 0.396 Binary Model β (SE) OR Gender (ref. female) 1.446 (0.091) 4.248 0.651 Age 0.590 (0.175) 1.804 0.001

139

Table 4-22 (continued) Predictor β (SE) OR Income (ref <$20,000) $20,000-29,999 0.969 (1.864) 2.636 0.603 $30,000-39,999 3.110 (2.479) 22.431 0.210 ≥$40,000 8.890 (3.419) 7262.350 0.009 BMI 0.010 (0.003) 1.010 <0.001 Neighborhood Self-Selection -3.017 (1.378) 0.049 0.029 Walk Score 0.205 (0.091) 1.228 0.024 Fit statistics χ2 (p) 14.370 (0.073) AIC 607.397 BIC 667.419 β: exponentiated beta coefficient; SE: standard error [robust] IRR: incidence rate ratio; OR: odds ratio

Walk Score and Recreational Walking/Cycling

The ZINB regression model investigated whether the Walk Score was associated with recreational walking and/or cycling both inside and outside of one’s neighborhood.

Within the count model, the Walk Score (β=-0.007, IRR=0.992, p=0.598) was not associated with weekly hours of recreational walking/cycling. Having an income between

$20,000-29,999 (β=0.391, IRR=1.479, p=0.025) was positively associated with weekly hours of recreational walking/cycling. All other covariates in the model were not statistically significant.

Within the binary model, the Walk Score (β=0.043, OR=1.045, p=0.488) was not associated with recreational walking/cycling. Older participants (β=0.330, OR=1.391, p=0.029), however, were at greater odds of not engaging in recreational walking/cycling.

All other covariates were not statistically significant. The Wald chi-square goodness of fit test indicated that the model was a good fit (p=0.011).

140

Table 4-23. ZINB regression models for the Walk Score and recreational walking/cycling β (SE) IRR p-value Predictor

Count Model Gender (ref. female) -0.051 (0.322) 0.950 0.873 Age -0.010 (0.013) 0.990 0.433 Income (ref. <$20,000) $20,000-29,999 0.391 (0.174) 1.479 0.025 $30,000-39,999 0.538 (0.364) 1.712 0.140 ≥$40,000 -0.571 (0.589) 0.565 0.332 BMI -0.000 (0.000) 1.000 0.224 Neighborhood Self-Selection -0.038 (0.286) 0.963 0.895

Walk Score -0.007 (0. 014) 0.992 0.598 Binary Model OR Gender (ref. female) -1.221 (1.421) 0.295 0.390 Age 0.330 (0.151) 1.391 0.029 Income (ref. <$20,000) $20,000-29,999 0.747 (1.301) 2.110 0.566 $30,000-39,999 2.588 (1.613) 13.298 0.109 ≥$40,000 1.983 (1.655) 7.264 0.231 BMI -0.006 (0.005) 0.994 0.244 Neighborhood Self-Selection -1.753 (1.314) 0.173 0.182 Walk Score 0.043 (0.063) 1.045 0.488 Fit statistics χ2 (p) 19.860 (0.011*) AIC 614.534 BIC 674.446 β: exponentiated beta coefficient; SE: standard error [robust] IRR: incidence rate ratio; OR: odds ratio

Walk Score and Total Physical Activity in METS

This research question also investigated whether the Walk Score was associated with overall physical activity expressed in METs. Components of overall physical activity included: active transportation, walking and cycling for recreation/health/fitness, and other moderate-to-vigorous intensity physical activity (MVPA). Within the count model, the Walk Score (β=-0.006, IRR=0.994, p=0.530) was not associated with greater

METs. None of the covariates in the model were statistically significant.

141

Within the binary model, the Walk Score did not predict engaging in overall physical activity (β=0.002, OR=1.002, p=0.909). Older age was associated with a 12.3% increased odds of being physically inactive (β=0.116, OR=1.23, p=0.029). Furthermore, greater BMI was associated with 0.012% decreased odds of being physically inactive

(β=-0.012, OR=0.988, p=0.011). Lastly, participants with a higher neighborhood self- selection score were at a 73% decreased odds of being physically inactive (β=-1.309,

OR=0.270, p=0.016).

Table 4-24. ZINB Regression Models for the Walk Score and METs of Physical Activity

β (SE) IRR p-value Predictor

Count Model§ Gender 0.267 (0.267) 1.306 0.316 Age -0.015 (0.009) 0.985 0.135 Income (ref. <$20,000) $20,000-29,999 0.114 (0.250) 1.121 0.650 $30,000-39,999 0.136 (0.288) 1.145 0.637 ≥$40,000 -0.302 (0.427) 0.739 0.479 BMI 0.000 (0.000) 1.000 0.867 Neighborhood Self-Selection 0.238 (0.248) 1.269 0.337 Walk Score -0.006 (0.009) 0.994 0.530 Binary Model§ OR Gender -0.078 (0.505) 0.925 0.878 Age 0.116 (0.026) 1.123 <0.001 Income (ref. <$20,000) $20,000-29,999 -0.179 (0.572) 0.836 0.755 $30,000-39,999 -0.465 (0.662) 0.628 0.483 ≥$40,000 0.870 (0.786) 2.387 0.269 BMI -0.012 (0.005) 0.988 0.011 Neighborhood Self-Selection -1.309 (0.545) 0.270 0.016 Walk Score 0.002 (0.017) 1.002 0.909 Fit statistics χ2 (p) 8.550 (0.382) AIC 2310.638 BIC 2370.660

142

In summary, this research question found that a greater Walk Score was associated with lower odds of engaging in active transportation. The Walk Score, however, was not significantly associated with greater reported hours of engaging in active transportation or overall physical activity. Furthermore, the Walk Score did not predict a greater likelihood of recreational walking/cycling, MVPA, or overall physical activity.

Participants who reported higher neighborhood self-selection scores reported greater hours of active transportation and higher odds of being physically active. This suggests that neighborhood self-selection is a bias, and if not controlled for, could overestimate the associations between the Walk Score and physical activity. Older participants were less likely to engage in active transportation, recreational walking/cycling, as well as overall physical activity. Those with an income of at least

$40,000 were less likely to engage in active transportation, and those with an income of

$20,000-29,999 engaged in less hours of recreational walking/cycling. Lastly, a higher

BMI was associated with lower odds of engaging in active transportation but higher odds of engaging in overall physical activity, a counterintuitive finding in respect to the literature and which is discussed in detail in the Discussion section below.

Results of Research Question 4

Research Question 4 examined whether the proximity to the closest recreational facility and park measured objectively by GIS was predictive of physical activity.

Proximity/accessibility to fitness and recreational facilities and parks was measured in two ways--by driving and walking proximity from the home of the respondent to the nearest facility/park. Physical activity facilities were divided into two components: 1)

143

gyms/recreational centers, and 2) specialty exercise facilities, which included yoga and dance studios, Muy Thai/boxing gyms, clubs (tennis, golf), Crossfit, Pilates, and other facilities that offered only specialized activities. Parks were represented by areas of green space that individuals utilized for recreation. The following outcomes were examined: walking in one’s neighborhood for active transport or recreation/fitness/health, active transportation indicated as walking or cycling for transport, MVPA, and overall physical activity. All outcome measures were expressed in METs. Several covariates were controlled for in the regression models, including age, gender, income, Asian BMI, neighborhood self-selection, and perceived crime.

Driving Proximity to Recreational Facilities/Parks

The ZINB models are presented for the association between proximity to recreational facilities/parks and walking, active transportation, and MVPA. The first

ZINB model represents whether the shortest driving time to a gym/recreational center, specialty exercise facility, and park were associated with total walking within one’s neighborhood. Within the count model, driving time to a gym/recreation center (β=0.141,

IRR= 1.152, p=0.180), specialty exercise facility (β=0.119, IRR=1.126, p=0.112) or park

(β= 0.188, IRR=1.207, p=0.129) were not associated with greater METS from walking within the neighborhood. Participants who had a BMI which met the criteria for obesity engaged in greater METs of walking within the neighborhood (β=0.846, IRR=2.331, p=0.014). No other covariates in the model reached statistical significance.

In the binary model, driving time to a gym/recreation center (β=‐0.087,

OR=0.917, p=0.639), specialty exercise facility (β=0.167, OR=1.182, p=0.112), and park

(β= 0.038, OR=1.038, p=0.852) were not associated with walking within the

144

neighborhood. Older participants (β=0.085, OR=1.089, p<0.001) were more to not walk in their neighborhood, while higher neighborhood self-selection (β=‐1.386, OR=0.250, p=0.011) decreased the odds of not walking. All other covariates were not statistically significant.

The second ZINB model presented below investigated whether driving time to the nearest gym/recreation center, specialized exercise facility, or park was associated with active transportation. Within the count model, longer driving time to a specialty exercise facility was associated with engaging in greater weekly METS of active transportation

(β=0.233, IRR=1.263, p=0.024). Longer driving time to the nearest park was marginally associated with greater weekly METS of active transportation (β=0.272, IRR=1.313, p=0.056).Obesity was associated with greater hours of active transportation by a factor of

1.21 (β=1.213, IRR=3.264, p=0.001). Other covariates in the model were not statistically significant.

In the binary model, driving time to the nearest gym/recreation center, specialty exercise facility, and park were not associated with engaging in active transportation.

Older participants (β=0.095, OR=1.100, p<0.001) and participants with an income of

≥$40,000 (β=1.336, OR=3.805, p=0.041) were more likely to not engage in active transportation. Participants with higher neighborhood self-selection was associated with a lower odds of engaging in active transportation (β=-1.593, OR=0.203, p=0.012).

The third model investigated relationships with MVPA. Within the count model, a longer driving time to a specialized exercise facility (β=0.163, IRR=1.177, p=0.030) and park (β=0.167, IRR=1.182, p=0.037) were associated with higher METS of self-reported

145

MVPA. Younger age (β=-0.038, IRR=0.963, p=0.001) and normal BMI (β=-0.533,

IRR=0.472, p=0.009) were associated with decreased weekly METS of MVPA.

Within the binary model, older age (β=0.025, OR=1.026, p=0.024) and obesity

(β=-1.793, OR=0.167, p=0.017) was associated with greater and lower odds of not engaging in MVPA, respectively. None of the independent variables related to proximity were statistically significant.

146

Table 4-25. ZINB Model for Driving Time to the Nearest Physical Activity Facilities and Parks and Association with Physical Activity

Model 1: Walking in the Model 2: Active Model 3: MVPA

neighborhood Transportation β (SE) IRR β (SE) IRR β (SE) IRR Predictor

Count Model§ Gender ‐0.208 (0.26) 0.812 ‐0.174 (0.24) 0.918 ‐0.183 (0.24) 0.833 Age ‐0.014 (0.01) 0.986 ‐0.010 (0.01) 0.871 -0.038 (0.01) 0.963* Income (ref. <$20,000) $20,000-29,999 0.613 (0.25) 1.846* 0.421 (0.33) 1.523 ‐0.316 (0.56) 0.729 $30,000-39,999 0.386 (0.23) 1.471 ‐0.056 (0.27) 0.946 0.507 (0.26) 1.659 ≥$40,000 ‐0.561 (0.44) 0.570 ‐0.389 (0.49) 0.678 0.668 (0.53) 1.950 Asian BMI (ref. under. <18.5) Normal (18.5-22.9) 0.302 (0.37) 1.352 0.501 (0.30) 1.651 ‐0.751 (0.28) 0.472* Overweight (23-24.9) 0.565 (0.34) 1.759 0.437 (0.28) 1.540 ‐0.533 (0.39) 0.587 Obese (25) 0.846 (0.34) 2.331* 1.213 (0.35) 3.364* 0.006 (0.22) 1.006 Perceived Crime 0.114 (0.27) 1.120 0.001 (0.21) 1.001 0.073 (0.22) 1.076 Neighborhood self-selection 0.413 (0.24) 1.512 0.014 (0.25) 1.014 ‐0.120 (0.30) 0.886 Drive, gym/rec. center 0.141 (0.11) 1.152 0.145 (0.11) 1.156 ‐0.036 (0.09) 0.965 Drive, specialty exercise 0.119 (0.07) 1.126 0.233 (0.10) 1.263* 0.163 (0.07) 1.177* Drive, nearest park 0.188 (0.12) 1.207 0.272 (0.14) 1.313 0.167 (0.08) 1.182*

Binary Model§ OR OR OR Gender ‐0.113 (0.45) 0.893 ‐0.282 (0.34) 0.754 0.180 (0.36) 1.197 Age 0.085 (0.02) 1.089** 0.095 (0.18) 1.100** 0.025 (0.01) 1.026* Income (ref. <$20,000) $20,000-29,999 ‐0.500 (0.53) 0.607 ‐0.404 (0.59) 0.667 0.603 (0.45) 1.828 $30,000-39,999 ‐0.016 (0.57) 0.984 0.137 (0.67) 1.147 0.260 (0.41) 1.296 ≥$40,000 0.656 (0.51) 1.926 1.336 (0.65) 3.805* 0.575 (0.86) 1.778 Asian BMI (ref. under. <18.5) Normal (18.5-22.9) ‐0.048 (0.62) 0.953 0.628 (0.70) 1.873 ‐1.297 (0.80) 0.273 Overweight (23-24.9) 0.277 (0.61) 1.319 0.458 (0.69) 1.580 -0.391 (0.94) 0.676 Obese (25) ‐0.436 (0.89) 0.646 0.147 (0.90) 1.158 -1.793 (0.75) 0.167*

147

Table 4-25 (continued) Binary Model§ β (SE) OR β (SE) OR β (SE) OR Perceived Crime ‐0.747 (0.56) 0.473 ‐0.389 (0.40) 0.678 ‐0.366 (0.32) 0.693 Neighborhood self-selection ‐1.386 (0.54) 0.250* ‐1.593 (0.63) 0.203* ‐0.372 (0.29) 0.690 Drive, gym/rec. center ‐0.087 (0.18) 0.917 ‐0.089 (0.18) 0.915 0.051 (0.21) 1.052 Drive, specialty exercise 0.167 (0.12) 1.182 0.246 (0.15) 1.279 0.005 (0.13) 1.005 Drive, nearest park 0.038 (0.21) 1.038 0.258 (0.24) 1.294 ‐0.178 (0.27) 0.837 Fit statistics alpha 0.718 0.687 0.548 AIC 1814.07 1594.88 1011.677 BIC 1882.93 1664.26 1081.05 β: exponentiated beta coefficient; SE: standard error [robust] IRR: incidence rate ratio; OR: odds ratio * p-value <0.05. ; ** p <0.001

148

The ZINB model presented examined the relationship between proximity, specifically driving time to the nearest gym/recreational center, specialty exercise facility, and overall physical activity. In the count model, longer driving time to the nearest park was associated with increased weekly METs of physical activity (β=0.271, IRR=1.311, p=0.013). A one unit increase in driving time to the nearest park resulted in a 31.1% increase in weekly METs of physical activity. Obesity was associated with greater weekly METs of physical activity (β=1.005, IRR=2.732, p=0.043). All other covariates and independent variables in the model were not statistically significant.

Within the binary model, older participants (β=0.116, OR=1.123, p<0.001) were less likely to engage in physical activity, while those with higher neighborhood self- selection scores (β=-1.525, OR=0.218, p=0.005) were less likely to not engage in physical activity. Independent variables of proximity to the nearest gym/recreation center, specialty exercise facility and park measured by driving time were not statistically significant.

Table 4-26. ZINB Model for Access to Nearest Drive Time to Physical Activity Facilities and Parks and Association with Overall Physical Activity

β (SE) IRR p-value Predictor

Count Model§ Gender 0.034 (0.27) 1.035 0.899 Age ‐0.017 (0.01) 0.983 0.100 Income (ref. <$20,000) $20,000-29,999 0.204 (0.25) 1.226 0.409 $30,000-39,999 0.121 (0.26) 1.129 0.637 ≥$40,000 ‐0.168 (0.59) 0.845 0.775 Asian BMI (ref. underweight 18.5) Normal (18.5-22.9) 0.426 (0.60) 1.531 0.476 Overweight (23-24.9) 0.255 (0.49) 1.290 0.604 Obese (25) 1.005 (0.50) 2.732 0.043 Perceived Crime 0.259 (0.23) 1.296 0.269 Neighborhood self-selection 0.319 (0.34) 1.376 0.354

149

Table 4-26 (continued) Predictor β (SE) IRR p-value Drive time, gym/recreational center ‐0.001 (0.11) 0.999 0.989 Drive time, specialty exercise facility 0.042 (0.10) 1.043 0.677 Drive time, park 0.271 (0.11) 1.311 0.013 Binary Model§ OR Gender 0.000 (0.47) 1.000 1.000 Age 0.116 (0.02) 1.123 <0.001 Income (ref. <$20,000) $20,000-29,999 ‐0.344 (0.61) 0.709 0.571 $30,000-39,999 -0.225 (0.65) 0.799 0.729 ≥$40,000 1.235 (0.67) 3.438 0.065 Asian BMI (ref. underweight 18.5) Normal (18.5-22.9) 0.110 (0.93) 1.116 0.906 Overweight (23-24.9) 0.277 (0.69) 1.319 0.687 Obese (25) ‐0.337 (0.91) 0.714 0.710 Perceived Crime ‐0.632 (0.48) 0.531 0.192 Neighborhood self-selection ‐1.525 (0.55) 0.218 0.005 Drive time, gym/recreational center ‐0.209 (0.19) 0.811 0.279 Drive time, specialty exercise facility 0.085 (0.13) 1.089 0.528 Drive time, park 0.219 (0.31) 1.245 0.483 Fit statistics alpha 0.875 AIC 2279.252 BIC 2348.624

Walking Proximity to Recreational Facilities/Parks

The following ZINB models investigated the relationship between walking proximity measured by the time it would take to walk to the nearest gym/recreational center, specialized exercise facility, and park with engagement in different types of physical activity. This analysis also investigated whether walking distance was associated with self-reported physical activity. However, walking distance was highly correlated with walking time and was not presented after it was illustrated that these the ZINB regression models produced similar findings.

The first outcome examined was walking in the neighborhood. Within the count model, greater walking time to the nearest park was positively associated with walking

150

within the neighborhood (β=0.045, IRR=1.046, p=0.040). More specifically, a one unit increase in walking time was associated with a 5% increase in walking. With respect to covariates, participants with obesity (β=0.826, IRR=2.285, p=0.013) and those with an income of $20,000-29,999 were more likely to walk within the neighborhood (β=0.600,

IRR=1.822, p<0.001). Within the binary model, greater walking time to the nearest specialty exercise facility was associated with greater odds of not walking in the neighborhood (β=0.057, OR=1.058, p=0.010). Older age (β=-0.016, OR=0.984, p<0.001) and a higher neighborhood self-selection score (β=-1.379, OR=0.252, p=0.010) were associated with lower odds of not walking in the neighborhood.

The subsequent ZINB regression model examined whether walking time to the nearest gym/recreation center, specialty exercise facility, and park were predictors of engaging in active transportation measured in METs. Within the count model, greater walk time to the nearest specialty exercise facility (β=0.053, IRR=1.054, p=0.011) and park (β=0.060, IRR=1.062, p=0.025) were associated with greater use of active transportation. A one unit increase in walk time to the nearest specialty facility or park was associated with a 5.4% and 6.2% increase in active transportation, respectively. With respect to covariates, participants with obesity engaged in greater levels of active transportation (β=1.179, IRR=3.251, p=0.001). All other covariates in the model did not reach statistical significance.

Within the binary model, greater walking time to the nearest specialty exercise facility (β=0.082, OR=1.085, p<0.001) and park (β=0.120, OR=1.128, p=0.020) increased the likelihood of not engaging in active transportation. Older age (β=0.097,

OR=1.101, p<0.001) and an income above $40,000 (β=1.437, (OR=4.209, p=0.037)

151

increased the odds of not engaging in active transportation. Participants with greater neighborhood self-selection scores were less likely to not engage in active transportation

(β=-1.609, OR=0.200, p=0.010).

The following ZINB model investigated whether the time it takes to walk to the nearest gym/recreation center, specialty exercise facility, and park were associated with

MVPA. Longer walking time to the nearest specialty exercise facility was associated with increased levels of MVPA (β=0.031, IRR=1.031, p=0.042), while gym/recreational center (β=-0.014, IRR=0.984, p=0.500) and parks (β=0.042, IRR=1.043, p=0.052) were not. An income between $30,000-39,999 (β=0.635, IRR=1.887, p=0.039) was associated with greater MVPA, while a normal BMI (β=-0.972, IRR=0.418, p=0.002) was associated with decreased levels.

Within the binary model, older age (β=0.027, OR=1.028, p=0.024) increased the odds of not engaging in MVPA, while participants with obesity (β=-1.769, OR=0.170, p=0.014) were at decreased odds of not engaging in MVPA. None of the independent variables related to proximity were statistically significant.

152

Table 4-27. ZINB Model of Walking Time to the Nearest Physical Activity Facilities and Parks and Association with Physical Activity

Model 1: Walking in the Model 2: Active Model 3: MVPA

neighborhood Transportation Predictor β (SE) IRR β (SE) IRR β (SE) IRR

Count Model§ Gender ‐0.138 (0.29) 0.871 ‐0.118 (0.26) 0.889 ‐0.186 (0.25) 0.831 Age ‐0.016 (0.01) 0.984 ‐0.011 (0.08) 0.989 ‐0.037 (0.01) 0.964* Income (ref. <$20,000) $20,000-29,999 0.600 (0.25) 1.822* 0.372 (0.30) 1.451 ‐0.240 (0.59) 0.787 $30,000-39,999 0.394 (0.25) 1.484 0.022 (0.30) 1.022 0.635 (0.31) 1.887* ≥$40,000 ‐0.574 (0.48) 0.563 ‐0.454 (0.52) 0.635 0.751 (0.53) 2.119 Asian BMI (ref. under 18.5) Normal (18.5-22.9) 0.335 (0.39) 1.398 0.517 (0.33) 1.677 ‐0.872 (0.28) 0.418* Overweight (23-24.9) 0.486 (0.36) 1.625 0.341 (0.33) 1.406 ‐0.778 (0.41) 0.459 Obese (25) 0.826 (0.33) 2.285* 1.179 (0.36) 3.251* -0.147 (0.41) 0.863 Perceived Crime 0.033 (0.25) 1.034 -0.064 (0.18) 0.938 0.002 (0.21) 1.002 Neighborhood self-selection 0.371 (0.25) 1.449 -0.045 (0.25) 0.956 ‐0.100 (0.31) 0.905 Walk time, gym/rec. center 0.022 (0.02) 1.023 0.020 (0.18) 1.020 ‐0.014 (0.02) 0.986 Walk time, specialty exercise 0.026 (0.02) 1.026 0.053 (0.02) 1.054* 0.031 (0.01) 1.031* Walk time, park 0.045 (0.02) 1.046* 0.060 (0.03) 1.062* 0.042 (0.02) 1.043 Binary Model§ OR OR OR Gender ‐0.094 (0.43) 0.910 ‐0.284 (0.34) 0.753 0.148 (0.38) 1.159 Age 0.084 (0.02) 1.087** 0.097 (0.02) 1.101** 0.027 (0.01) 1.028* Income (ref. <$20,000) $20,000-29,999 ‐0.536 (0.52) 0.585 ‐0.474 (0.56) 0.623 0.560 (0.48) 1.751 $30,000-39,999 ‐0.037 (0.58) 0.964 0.175 (0.67) 1.191 0.254 (0.41) 1.289 ≥$40,000 0.711 (0.51) 2.037 1.437 (0.69) 4.209* 0.621 (0.90) 1.861 Asian BMI (ref. under 18.5) Normal (18.5-22.9) ‐0.060 (0.62) 0.941 0.661 (0.70) 1.937 ‐1.262 (0.77) 0.283 Overweight (23-24.9) 0.215 (0.60) 1.240 0.399 (0.70) 1.490 ‐0.357 (0.89) 0.700 Obese (25) ‐0.486 (0.89) 0.615 0.207 (0.88) 1.230 ‐1.769 (0.72) 0.170* Perceived Crime ‐0.831 (0.57) 0.436 -0.503 (0.43) 0.605 -0.366 (0.32) 0.693 Neighborhood self-selection ‐1.379 (0.54) 0.252* -1.609 (0.63) 0.200* ‐0.417 (0.29) 0.659

153

Table 4-27 (continued) Predictor β (SE) OR β (SE) OR β (SE) OR Walk time, gym/rec center ‐0.006 (0.04) 0.994 ‐0.022 (0.03) 0.978 ‐0.007 (0.03) 0.993 Walk time, specialty exercise facility 0.057 (0.02) 1.058* 0.082 (0.02) 1.085** 0.012 (0.03) 1.012 Walk time, park 0.073 (0.05) 1.076 0.120 (0.05) 1.128* ‐0.054 (0.05) 0.948 Fit statistics alpha 0.723 0.704 0.546 AIC 1810.107 1588.839 1010.709 BIC 1878.965 1658.212 1080.081 β: exponentiated beta coefficient; SE: standard error [robust] IRR: incidence rate ratio; OR: odds ratio * p-value <0.05. ; ** p <0.001.

154

The ZINB model presented below examined whether walk time to the gym/recreational center, specialty exercise facility and park predicted physical activity levels. In the count model, greater walk time to the nearest park was associated with greater physical activity levels (β=0.059, IRR=1.061, p=0.001). A one unit increase in walk time was equivalent to a 6.1% increase in physical activity. Furthermore, those with obesity reported a higher rate of physical activity (β=1.007, IRR=2.737, p=0.043). All of the other covariates did not reach statistical significance.

Within the binary model, older age was associated with greater odds of not engaging in physical activity (β=0.113, OR=1.119, p<0.001). Higher neighborhood self- selection was associated with a lower odds of not engaging in physical activity (β=-1.454,

OR=0.234, p=0.007). The independent predictors related to proximity were not statistically significant.

Table 4-28. ZINB Model for Nearest Walk Time to Physical Activity Facilities and Parks and Association with Overall Physical Activity

p- β (SE) IRR Predictor value

Count Model§ Gender 0.083 (0.30) 1.086 0.782 Age ‐0.017 (0.01) 0.983 0.075 Income (ref. <$20,000) $20,000-29,999 0.206 (0.23) 1.229 0.378 $30,000-39,999 0.168 (0.29) 1.183 0.557 ≥$40,000 ‐0.144 (0.63) 0.866 0.818 Asian BMI (ref. underweight 18.5) Normal (18.5-22.9) 0.431 (0.62) 1.539 0.487 Overweight (23-24.9) 0.219 (0.51) 1.244 0.670 Obese (25) 1.007 (0.50) 2.737 0.043 Perceived Crime 0.336 (0.35) 1.236 0.350 Neighborhood self-selection 0.212 (0.23) 1.399 0.350 Walk time, gym/recreational center 0.000 (0.02) 1.000 0.996 Walk time, specialty exercise facility 0.007 (0.02) 1.007 0.745 Walk time, park 0.059 (0.02) 1.061 0.001

155

Table 4-28 (continued) Binary Model§ OR Gender 0.016 (0.43) 1.016 0.971 Age 0.113 (0.02) 1.119 <0.001 Income (ref. <$20,000) $20,000-29,999 ‐0.353 (0.61) 0.703 0.562 $30,000-39,999 ‐0.283 (0.65) 0.754 0.663 ≥$40,000 1.275 (0.67) 3.578 0.058 Asian BMI (ref. underweight 18.5) Normal (18.5-22.9) 0.078 (0.95) 1.081 0.935 Overweight (23-24.9) 0.231 (0.71) 1.260 0.745 Obese (25) ‐0.339 (0.91) 0.713 0.708 Perceived Crime ‐0.702 (0.48) 0.496 0.145 Neighborhood self-selection ‐1.454 (0.53) 0.234 0.007 Walk time, gym/recreational center ‐0.025 (0.04) 0.975 0.519 Walk time, specialty exercise facility 0.036 (0.03) 1.037 0.201 Walk time, park 0.080 (0.06) 1.084 0.186 Fit statistics alpha 0.876 AIC 2278.314 BIC 2347.686

In summary, the findings illustrated that longer driving time to the nearest park was associated with greater METs associated with MVPA and overall PA, while longer driving time to a specialty exercise facility was linked to greater METs associated with walking and active transportation. Furthermore, a greater walking time to the nearest park was associated with greater weekly METs of walking, active transportation, and overall

PA, while a greater walking time to the nearest specialty exercise facility was associated with greater METs of active transportation and MVPA. At the same time, a greater walking time to a nearest specialty exercise facility decreased the odds of walking and active transportation, while a greater walking time to the nearest park decreased the odds of active transportation. Greater driving or walking time to the nearest gym/recreation center were not associated with any type of physical activity.

156

Several covariates in the model were significantly associated with physical activity behavior. Older participants were less likely to engage in MVPA and overall physical activity. Participants with an income between $20,000-29,999 were more likely to engage in greater METs of walking in their neighborhood, and participants with an income of $30,000-39,999 were more likely to engage in weekly METs of MVPA.

Participants with an income above $40,000 were less likely to engage in active transportation. Participants with obesity were more likely to engage in greater weekly

METs of walking, active transportation and overall physical activity, as well as participate in overall more MVPA. This finding was counterintuitive, as BMI is inversely related to levels of physical activity. Lastly, participants with greater neighborhood self- selection scores were more likely to engage in any walking in the neighborhood, active transportation and overall physical activity.

Additional Analyses

Additional analyses were conducted to further characterize physical activity patterns not directly addressed in the four primary research questions. These analyses aimed to determine levels of congruence between perceived and objective built environment characteristics and how they impacted engagement in physical activity.

First, as an extension of Research Question 4, predictors of park use were measured in relation to predicting whether participants reported that they walked to a park for recreational purposes. A logistic regression model was used while controlling for gender, age, income, BMI, neighborhood self-selection, and crime. Perceived closeness to a park within one’s neighborhood was associated with greater odds of using the park

(OR=1.733, p<0.001). However, the actual objective reported walking distance measured

157

by GIS (OR=1.000, p=0.508) and the Walk Score was not related to park use (OR=1.013, p=0.404).

Table 4-29. Built Environment Predictors of Park Use

OR (95% CI) p-value Predictor

Walking distance to a park (GIS) 1.000 (0.999, 1.001) 0.508 Perceived closeness to a park 1.733 (1.286, 2.319) <0.001 Walk Score 1.013 (0.983, 1.044) 0.404 Gender 0.647 (0.334, 1.250) 0.195 Age 1.007 (0.970, 1.045) 0.682 Income (ref. <$20,000) $20,000-29,999 1.082 (0.347, 3.376) 0.892 $30,000-39,999 1.665 (0.652, 4.252) 0.287 ≥$40,000 0.346 (0.062, 1.924) 0.225 Asian BMI (ref. underweight 18.5) Normal (18.5-22.9) 2.254 (0.489, 10.40) 0.297 Overweight (23-24.9) 1.665 (0.652, 4.252) 0.860 Obese (25) 1.632 (0.382, 6.976) 0.509 Perceived Crime 1.497 (0.757, 2.961) 0.246 Neighborhood self-selection 1.723 (0.757, 3.923) 0.195

The next analysis examined whether objective and perceived neighborhood walkability predicted meeting the physical activity guidelines (U.S. Department of Health and Human Services, 2018). Higher perceived neighborhood walkability was associated with greater odds of meeting the physical activity guidelines (OR=5.27, p<0.001), while the walk score was not statistically significant (OR=0.999, p=0.404).

Table 4-30. Built Environment Predictors and Meeting the Physical Activity Guidelines

OR p-value Predictor

Perceived neighborhood walkability 5.257 (1.055, 26.175) <0.001 Walk Score 0.999 (0.960, 1.040) 0.404 Gender 0.913 (0.448, 1.862) 0.803 Age 0.949 (0.915, 0.985) 0.006

158

Table 4-30 (continued) Predictor OR p-value Income (ref. <$20,000) $20,000-29,999 1.434 (0.550, 3.739) 0.461 $30,000-39,999 1.477 (0.561, 3.886) 0.429 ≥$40,000 0.646 (0.097, 4.280) 0.650 Asian BMI (ref. underweight 18.5) Normal (18.5-22.9) 1.516 (0.482, 4.770) 0.488 Overweight (23-24.9) 1.260 (0.381, 4.163) 0.705 Obese (25) 1.851 (0.544, 6.298) 0.324 Neighborhood self-selection 2.808 (1.718, 4.589) <0.001

The analysis also aimed to evaluate the relationship of ethnic group with different perceived and objective built environment characteristics in meeting the aerobic physical activity guidelines of 150 weekly minutes of aerobic physical activity, operationalized as at least 450 METs. This model controlled for gender and age. Additional analyses investigated whether there were differences in the relationship between built environment characteristics and meeting the physical activity guidelines by Asian ethnic groups. With respect to the Walk Score, Vietnamese Americans who lived in neighborhoods with a higher Walk Score were more likely to meet the physical activity guidelines (OR=4.352, p<0.001). Chinese Americans who perceived their neighborhood to be highly walkable were more likely to meet the physical activity guidelines (OR=7.745, p=0.005);

Vietnamese Americans were less likely to meet the guidelines (OR=2.6e-26, p<0.001).

Vietnamese Americans who lived further away from a park (OR=0.589, p<0.001) were less likely to meet the physical activity guidelines. However, they were more likely to meet the guidelines if they further from a gym/recreation center (OR=1.250, p<0.001) and specialty exercise facility (OR=1.335, p<0.001). Koreans who lived further away from a park were more likely to meet the guidelines (OR=1.001, p=0.027) but were less

159

likely to meet the guidelines the further away they were to a gym/recreational center

(OR=0.998, p<0.001). Lastly, Chinese Americans who lived further from a specialty exercise facility were less likely to meet the physical activity guidelines (OR=0.999, p=0.002).

Table 4-31. Ethnic Differences in the Relationship between Built Environment Factors and Physical Activity

OR 95% CI Predictor§

Ethnic Group*Walk Score Chinese 0.979 0.931, 1.029 Korean 0.950 0.899, 1.003 Vietnamese 4.352** 3.634, 5.212 Ethnic Group*NEWS Chinese 7.745* 1.841, 32.578 Korean 1.307 0.133, 12.867 Vietnamese 2.6e-126** 1.1e-138, 6.4e-114 Ethnic Group*Distance to park Chinese 1.000 0.999, 1.001 Korean 1.001* 1.000, 1.002 Vietnamese 0.589** 0.558, 0.622 Ethnic Group*Distance to gym/rec center Chinese 1.000 0.999, 1.001 Korean 0.998** 0.997, 0.999 Vietnamese 1.250** 1.223, 1.277 Ethnic Group*Distance to specialty facility Chinese 0.999* 0.997, 0.999 Korean 1.001 0.999, 1.001 Vietnamese 1.335** 1.295, 1.375 § The model controlled for gender and age

160

CHAPTER 5

DISCUSSION

Summary

The preceding chapter described the statistical analyses of the relationship of perceived and objective built environmental characteristics on neighborhood physical activity levels among 240 Asian Americans from Philadelphia County, Pennsylvania.

This study was the first, to the author’s knowledge, that examined validated built environment measures, such as the NEWS and GIS proximity to physical activity facilities near this population of individuals. The research questions, more specifically, examined how factors of the built environment are associated with both the odds of engaging in physical activity, as well as the duration and frequency of participation. The research questions were answered by ZINB regression models with robust standard errors, accounting for clustering effects to examine associations between built environment measures and physical activity. The main findings for each research question are highlighted in the brief paragraphs below and subsequently discussed in greater detail. Furthermore, a summary of findings from the ZINB models for each research question can be found in Appendix I.

Research Question 1 examined how perceived neighborhood walkability was associated with engagement in physical activity. Overall, the results indicated that perceived neighborhood walkability was associated with greater odds of engaging in physical activity as well as recreational walking/cycling, more specifically, but not active transportation. Perceived neighborhood walkability also was not associated with greater

161

weekly hours participating in active transportation, recreational walking/cycling, or physical activity.

Research Question 2 examined whether ethnic identity moderated the relationship between perceived neighborhood walkability and physical activity. Ethnic identity moderated the relationship between perceived neighborhood walkability and active transportation as well as moderate-intensity physical activity. However, ethnic identity was not associated with the overall level of physical activity, vigorous-intensity physical activity, or recreational walking/cycling.

Research Question 3 investigated if objective neighborhood walkability, measured with the Walk Score, was associated with physical activity. There were no significant associations for the relationship between objective neighborhood walkability

(Walk Score) and physical activity or recreational walking/cycling. The Walk Score was associated with lower odds of participation in active transportation but was not associated with weekly hours of activity.

Finally, Research Question 4 explored how the proximity of physical activity resources, specifically gyms/recreation centers, specialty exercise facilities, and parks, were associated with physical activity. Participants who lived in closer proximity to a park had greater odds of engaging in active transportation but reported a lower frequency of engaging in walking, active transportation, and physical activity. Participants within closer proximity to a specialty exercise facility also were at greater odds of engaging in walking and active transport but had a lesser freequency spent engaging in active transportation and MVPA. Proximity to a gym/recreation center was not associated with any of the physical activity measures.

162

Research Question 1 – Perceived Neighborhood Walkability and Physical Activity

The first hypothesis examined whether perceived neighborhood walkability predicted active transportation behavior, which constitutes walking and cycling for commuting to and from destinations (Wanner et al., 2012). This hypothesis was not supported. Perceived neighborhood walkability has been linked to greater active transportation in several large studies (Cerin et al., 2007, 2017; Saelens, Sallis, Black, et al., 2003; Thielman et al., 2015; Van Dyck et al., 2012). However, those observations were not replicated here. The present study had a much smaller sample size compared to these sample sizes, which ranged from several hundred to thousands of participants. Also, this study used a convenience sampling approach rather than an area-based sampling of various high and low walkable neighborhoods.

Nevertheless, specific aspects of neighborhood infrastructure were predictors of active transportation. Specifically, more physical barriers that included freeways, railways, and rivers, were associated with both a lower likelihood to engage in active transportation and lower levels of active transportation as measured by time. As found in the literature, large infrastructural and natural barriers to walking can influence one’s decision to engage in active transport (Hinckson et al., 2017; Van Dyck et al., 2012). In this case, rivers and freeways which surround Philadelphia County could serve as major structural barriers that may hinder active transportation and make it necessary and more convenient for study participants to travel by car. For example, individuals who work further away from home may utilize these freeways more to travel, resulting in less walking and cycling within the neighborhood as a means of transport.

A particularly interesting observation was the relationship between engagement in active transportation and the perception of cul-de-sacs in the neighborhood. Intuitively, 163

one would anticipate that having greater cul-de-sacs would serve as a barrier to walking, as they would limit the ability to walk further from home. However, participants that perceived their neighborhood to have more cul-de-sacs were more likely to engage in and report more hours of active transportation. The presence of cul-de-sacs has encouraged safer settings for physical activity to occur (Mecredy et al., 2011). In this respect, participants may have felt more comfortable living in a neighborhood with greater cul-de- sacs, which, in turn, positively influences on transport-related walking and cycling. These study findings are comparable to a multi-country study across metropolitan areas of the

United States, Australia, and Belgium, which found that physical barriers to walking, and fewer rather than greater cul-de-sacs were associated with lower levels of physical activity (Van Dyck et al., 2012).

Land use mix and access were not found to be associated with active transportation. A lack of relationship may suggest that participants lived greater distances to their workplace and which required them to commute by other means, such as by car or public transportation. Closer proximity to one’s workplace and to food shops have been associated with more minutes of walking each day than other destinations (Cerin et al., 2007). Although the NEWS-A includes several items that assess the built environment, it may not have captured unique destinations used by Asian Americans

(Cerin et al., 2007; Owen et al., 2004). For example, Asian Americans report a preference for shopping at ethnic grocery stores (G. X. Ma et al., 2019; L. Wang & Lo, 2007). If these stores are concentrated in a community that is further away, active transportation levels may be lower. Further investigation of transport-related walking patterns would

164

help improve understanding of land-use mix destinations, which would help identify which destinations are important for promoting this behavior.

The second hypothesis of this research question examined whether perceived neighborhood walkability predicted increased levels of recreational walking and cycling.

Greater overall neighborhood walkability was associated with increased odds of engaging in recreational walking and cycling, but not a greater duration of engagement in those activities. There were differences in the NEWS subscales that predicted recreational walking and cycling in comparison to active transportation. Similar to active transportation, participants who perceived more cul-de-sacs in their neighborhoods and lower residential density reported more recreational walking/cycling. However, participants living close to businesses and facilities (land use mix diversity) and with access to services (land use mix access) engaged in greater recreational walking and cycling behaviors.

Existing literature on recreational walking/cycling has supported findings regarding land use mix but not for other factors that were statistically significant in this study (Cerin, Sit, et al., 2013; Ghani et al., 2019; Sugiyama et al., 2014). For example, in a study of 13,745 adults from five continents (North and South America, Europe, Asia, and Australia), land use mix positively impacted recreational walking behaviors

(Sugiyama et al., 2014). At the same time, greater residential density and fewer cul-de- sacs were predictive of greater recreational walking behavior (Sugiyama et al., 2014).

These observations may indicate that other mechanisms related to street connectivity were unique to the current study. The counterintuitive study findings on residential

165

density within this study were due to the poor regression model fit when examining how residential density impacts recreational walking and cycling.

The present study found that perceived crime also was associated with recreational walking and cycling; participants who perceived their neighborhood to have more crime reported less recreational walking and cycling. Interestingly, perceived crime was not associated with other forms of physical activity, such as active transportation or

MVPA. This finding may indicate that individuals’ perceived safety from neighborhood crime may not be impacted by active transportation if it’s their main mode of getting to and from work, shopping, or places of worship. On the other hand, engaging in recreational activities can be highly influenced by one’s perceived level of safety and crime in their neighborhood (Rees-Punia et al., 2018; Sugiyama et al., 2014). Individuals who view their neighborhood to have more crime typically engage in less recreational walking (Rees-Punia et al., 2018; Sugiyama et al., 2014). Thus, reducing crime and increasing perceptions of neighborhood safety may impact recreational physical activity

(Rees-Punia et al., 2018; Sugiyama et al., 2014).

The third hypothesis of Research Question 1 investigated whether perceived neighborhood walkability impacted overall physical activity. Perceived neighborhood walkability was associated with greater odds of engaging in physical activity. These findings are in line with existing studies which have also suggested that neighborhood walkability can positively impact physical activity (Ding et al., 2013; M. J. Duncan et al.,

2005; Hinckson et al., 2017; Hoehner et al., 2005; Kaczynski et al., 2012). In the present study, land use mix diversity and access had the strongest associations with physical activity, replicating previous work (Cerin et al., 2014; Ding et al., 2013; Hoehner et al.,

166

2005; Nichani et al., 2019; Van Cauwenberg et al., 2018). The International Physical

Activity and the Environment Network (IPEN) Adult study, for example, is the most extensive multi-country epidemiological study of perceived neighborhood walkability

(Cerin et al., 2014). Land use mix access had the strongest association with daily minutes of MVPA. (Cerin et al., 2014). The current study replicated this finding with the use of a self-reported measure of physical activity rather than an objective assessment of activity with an accelerometer, as used in the IPEN study. The present findings also are in line with other large epidemiological studies and systematic reviews (Cerin et al., 2014;

Hoehner et al., 2005; Sugiyama et al., 2009; Van Cauwenberg et al., 2018).

The present study found that participants who perceived their neighborhood to be more aesthetically pleasing were more likely to engage in physical activity. This suggests that greenery, building murals, and attractive buildings where participants resided may provide a more positive and joyful environment to engage in physical activity within the study population based on their neighborhood. Neighborhood aesthetics have been associated with greater engagement in physical activity and leisure-based activities in numerous studies (Cerin et al., 2014; Hoehner et al., 2005; Sugiyama et al., 2009; Van

Cauwenberg et al., 2018). Individuals who perceived their neighborhood to have more attractive natural and built environment features are more likely to engage in physical activity (Cerin et al., 2014; Hoehner et al., 2005; Sugiyama et al., 2009; Van Cauwenberg et al., 2018).

In the present study, the greater number of cul-de-sacs in a neighborhood and lower residential density were associated with higher levels of physical activity. Both observations are in contrast to the supporting literature. In the present study, participants

167

living in areas with single-family residences and row homes or low-rise apartments, as compared to high rise apartments, engaged in greater amounts of physical activity.

However, these contradictory study findings were attributed to the poor model fit for residential density when running the zero-inflated regression, as mentioned previously with regards to findings on recreational walking/cycling.

Furthermore, participants who lived in neighborhoods with greater cul-de-sacs were more physically active. Studies that have investigated street connectivity related measures with overall physical activity have found conflicting findings regarding the impact of perceived cul-de-sacs in the neighborhood (Humpel et al., 2002; Mecredy et al., 2011; Owen et al., 2004; Wells & Yang, 2008). For example, Van Dyck and colleagues (2012) found that having fewer cul-de-sacs predicted greater activity levels. In contrast, a study conducted in adolescents measuring objective street connectivity found that having greater cul-de-sacs predicted greater physical activity (Mecredy et al., 2011).

The present results are consistent with the findings of this study and suggest that participants may perceive that there are greater opportunities to engage in physical activity if they reside within or near cul-de-sacs. However, the type of cul-de-sacs within the city are distinct from typical cul-de-sacs seen in suburban areas that have single- family residences. Generally, urban areas have fewer cul-de-sacs and, if present, have homes with multiple family residences. However, given that the presence of cul-de-sacs was associated with all three physical activity behaviors in this study, these types of locations may serve as settings to engage in physical activity.

A number of studies have used the NEWS as a measure of the built environment

(M. A. Adams et al., 2009; Cerin, Conway, et al., 2013; Cerin et al., 2010, 2014; Ding et

168

al., 2013; Kerr et al., 2016; Sugiyama et al., 2014). However, the present study is believed to be the first to use the measure in a predominantly Asian American population.

A handful of studies have been conducted in urban environments in Asia, which may not necessarily be comparable due to differences in distinct culture, climate, environmental policies, and overall urban and city design compared to Western countries (Cerin,

Conway, et al., 2013; Van Dyck et al., 2012). For example, built environment studies in

China differ by having greater significant residential and population density, and land use mix businesses that are more common (e.g., noodle shops), which have been adapted in their studies (Cerin et al., 2010, 2011). In a similar study, Bungum and colleagues (2012) examined four items that measured perceived neighborhoods among Asian Pacific

Islanders living in Las Vegas: access to physical activity and grocery stores, aesthetics, presence of sidewalks, and traffic. Environmental accommodations that promote physical activity, such as the presence of sidewalks as well as having recreational facilities and grocery stores were found to be a strong predictor of physical activity (Bungum et al.,

2012). The current study incorporated more comprehensive measures of the built environment, which was captured by the NEWS.

Although active transportation and recreation constitutes both walking and cycling, participants in the present study reported extremely low levels of cycling.

Research in the general population has found stronger associations between how neighborhood infrastructure affects walking (Owen et al., 2004; Sugiyama et al., 2012); there are far fewer studies on cycling (McCormack & Shiell, 2011; Van Dyck et al.,

2012). Heesch and colleagues (2014) found that none of the built environment factors they examined were associated with leisure cycling versus transport cycling, suggesting

169

that the built environment impacts these behaviors differently. Although cycling behaviors have continued to rise in the US, more individuals engage in cycling for exercise rather than as a means of active transport (Kerr et al., 2016; L. Ma & Dill, 2015).

The findings also could be an artifact of the research method used to address this question. While the NEWS-A measures several aspects of the built environment, it is limited by an inability to capture the neighborhood-cycling infrastructure, as only one item was included regarding cycling safety. Therefore, the perceived quality of cycling infrastructures among participants was unclear. Cycling as a form of active commuting has many reported benefits in urban environments, including Philadelphia, but is highly influenced by safety and proper infrastructure within the neighborhood (Dhakal et al.,

2018; Hoffman et al., 2014). A greater presence of bike lanes and trails are generally situated and connected near Center City; they are less common in more northern parts of the county (Bicycle Coalition of Greater Philadelphia, 2020). Thorough investigations of the presence of bike trails and other infrastructure for cycling can promote safe active transport, specifically cycling within one's neighborhood (A. Bauman et al., 2009;

Saelens & Handy, 2008).

In conclusion, among the Asian-Americans studied in this investigation, higher perceived neighborhood walkability was associated with higher levels of physical activity. Land use mix diversity and access also were associated with physical activity, supporting their importance as relevant elements of the built environment. At the same time, having fewer cul-de-sacs in the neighborhood was associated with lower self- reported active transportation, recreational walking cycling, and physical activity.

Consistent with previous research as well as an intuitive thought, greater perceived crime

170

negatively was associated with less recreational walking/cycling. In contrast, neighborhood aesthetics were associated with higher levels of overall physical activity.

These findings suggest that perceived neighborhood walkability contributes to physical activity behavior and that these characteristics could potentially be emphasized in physical activity interventions among Asian American populations.

Research Question 2 – Ethnic Identity and Physical Activity

The second research question examined cultural influences, and, more specifically, how the concept of ethnic identity is associated with perceived neighborhood walkability and self-reported physical activity. In the present study, ethnic identity was operationalized as one’s attachment to their ethnic group; in this case, one’s sub-ethnic

Asian group (Phinney, 1992). Participants scored relatively high on the MEIM-R, the gold standard measure of ethnic identity, with a mean score of 3.61  0.84. Asian

Americans have reported high levels of ethnic identity, even surpassing other ethnic groups, including Black/African American, Hispanic, and Non-Hispanic Whites (S. D.

Brown et al., 2014). High ethnic identity indicates a potential influence on physical activity behaviors (Henley et al., 2016; Romeike et al., 2016; Woolford et al., 2013).

The results partially supported the hypothesis that perceived neighborhood walkability and physical activity levels varied by the level of ethnic identity. More specifically, there was a relationship between neighborhood walkability and ethnic identity with regard to active transportation. Active transportation was highest when perceived neighborhood walkability was low, and ethnic identity was high. Participants who reported high ethnic identity and judged their neighborhoods to be lower in walkability reported 3.39 more hours per week of activity as compared those who with

171

higher perceived neighborhood walkability. This suggests that high ethnic identity is linked to greater active transportation behavior in those who perceive their neighborhood to be less walkable. Individuals with high ethnic identity, who appear to have strong attachments to the values and beliefs of their ethnic group, appear to engage in greater active transportation regardless of the perceived walkability of their local neighborhoods.

The research question also explored the relationship between ethnic identity and neighborhood walkability in predicting engagement in recreational walking and cycling, and MVPA. There was an interaction with respect to moderate-intensity physical activity; engagement was highest when both perceived neighborhood walkability and ethnic identity were low. Those who had a low ethnic identity and low perceived neighborhood walkability only engaged in 4.8 minutes more moderate activity compared to those with high perceived neighborhood walkability and low ethnic identity. This was unexpected, given that the original hypothesis suggested both neighborhood walkability and ethnic identity would be associated with higher levels of moderate-intensity physical activity.

However, the minimal difference in levels of moderate activity is unlikely to reflect a clinically meaningful finding at the individual level.

When physical activity was examined overall, there were no statistically significant interactions between neighborhood walkability and ethnic identity level.

However, ethnic identity was associated with greater odds of engaging in overall physical activity as a main effect. Although no studies have examined the interactive effect of ethnic identity on the built environment, some have investigated the impact of ethnic identity on physical activity by racial and ethnic groups. Smalley and colleagues (2016), for example, found that African American women who had a stronger attachment to their

172

ethnic identity exhibited greater motivation to exercise, but this study did not directly measure physical activity behavior. In a study of pregnant women with gestational diabetes in the US, with about half of the sample identifying as Asian-American, a one- unit increase in ethnic identity score was associated with more considerable self-reported minutes of physical activity (S. D. Brown et al., 2016). Therefore, this study’s findings contribute to the literature that also illustrated that ethnic identity is linked to higher odds of engaging in physical activity.

In summary, ethnic identity interacted with the perceived built environment to impact active transportation and moderate-intensity physical activity. These inconsistent findings may be attributed to other moderating factors, such as acculturation and sociodemographic factors related to immigration that also influence physical activity. For example, lower-income immigrants may also express high ethnic identity but are more physically active through their occupation or lack of car ownership, as has been observed in ethnic minority populations (Martinez et al., 2008; O’Loughlin et al., 2007; Wieland et al., 2015; Yi et al., 2016). Furthermore, acculturation has been studied with respect to engaging in leisure-time physical activity (LTPA) in Asian-American immigrants

(Becerra et al., 2015; K. Li & Wen, 2013; Yi et al., 2016). Acculturation is the process in which a person adopts the practices and values of US society (B. Berry & Berry, 2005; J.

W. Berry, 1997). For example, among Chinese immigrants, those who are less acculturated and personally identified more with their own Chinese culture were less likely to engage in LTPA (Yi et al., 2016). Given that acculturation and ethnic identity are related, these measures should be examined in conjunction with sociodemographic and immigration factors as predictors of physical activity (Becerra et al., 2015; K. Li &

173

Wen, 2013; Yi et al., 2016). Additional research is needed to understand the mechanisms underlying how cultural factors can promote or hinder physical activity behavior.

Research Question 3 – The Walk Score and Physical Activity

Research Question 3 was posited to investigate whether objective neighborhood walkability, measured by the Walk Score, was associated with greater active transportation, recreational walking/cycling, and overall levels of physical activity. A greater Walk Score predicted a lower likelihood of engaging in active transportation. This finding was unexpected. The Walk Score directly influences active transportation by its relationship to mixed land use and access to various businesses and facilities in one’s neighborhood (Carr et al., 2010). Studies have repeatedly found that a higher Walk Score is associated with greater walking for transport (Hall & Ram, 2018; Towne et al., 2016).

A potential reason for this counterintuitive finding in the present study may be that while study participants lived in neighborhoods with a higher Walk Score, other features such as sidewalk quality, the presence of graffiti and litter, and perceived crime could negatively impact the decision to walk locally. Although these were not directly measured, these factors are shown to be a barrier to walking even in highly walkable neighborhoods (Hirsch et al., 2013; Vargo et al., 2012).

With regard to the second hypothesis of this research question, the Walk Score also did not predict recreational walking and cycling. Two studies of Asian Americans have found a relationship between the Walk Score and overall walking behavior (Hirsch et al., 2013; Kelley et al., 2016). Hirsch and colleagues (2013), for example, examined relationships between the Walk Score and walking/cycling among participants in the

Multi-Ethnic Study of Atherosclerosis (MESA), where 11% of the sample was comprised

174

of Chinese Americans (n=4552). A higher Walk Score was associated with greater walking for transport and recreation (Hirsch et al., 2013). Similarly, the MASALA study in South Asians (n=906) investigated minutes spent walking for transport. The Walk

Score was associated with an additional 13 minutes of walking for men; there was no association for women (Kelley et al., 2016). The current study did not replicate these findings. There were two major differences between the current investigation and those studies. First, the sample size in the present study was much smaller and from concentrated areas of Philadelphia rather than the entire metropolitan area. Second, the socioeconomic status of participants also differed; for example, the majority of

MASALA study participants were highly educated and had a yearly income of above

$100,000 (Kelley et al., 2016). The present study, in contrast, represented Asian

Americans residing solely in urban neighborhoods with lower reported education and income levels.

With respect to the third hypothesis, the Walk Score did not predict a greater likelihood or duration of engagement in overall physical activity. There are several potential explanations. The mean Walk Score among the study sample was 83 out of 100, suggesting that the neighborhoods were highly walkable. The majority of study participants (~88%) lived in very walkable neighborhoods in Philadelphia. This is in comparison to other areas of the city that are less walkable and car-dependent, such as northeast sections of Philadelphia County, where fewer participants resided. In other studies, there has typically been a wider range of Walk Scores across the area of interest

(Cole et al., 2015; Hall & Ram, 2018; Towne et al., 2016). Since study participants mainly resided in higher walkable neighborhoods, further assessments of the quality of

175

the built environment should be examined directly within these communities in future research.

In summary, objective measures of neighborhood walkability were not associated with self-reported physical activity among Asian-Americans living in Philadelphia.

Participants already resided in highly walkable neighborhoods, which may explain the non-significant findings. Since the Walk Score itself measures accessibility to different commercial destinations, these spaces may not be directly associated with engaging in physical activity reported by this study population.

Research Question 4 – Proximity to Physical Activity Facilities/Parks and Physical Activity

The purpose of Research Question 4 was to determine the impact of proximity to recreational facilities and parks on physical activity behaviors using GIS measures.

Closer proximity to a park was hypothesized to be associated with greater participation in various forms of physical activity. The study findings did not fully support the hypothesis. Closer proximity to a park (shorter walk time) predicted a higher likelihood of engaging in active transportation, as predicted. However, proximity to a park was associated with a lower frequency of active transportation.

These contradictory findings may indicate that living near a park promotes the overall decision to engage in active transportation. However, park proximity may not provide a direct correlation suggesting that participants utilize these parks as a means of engaging active transport. Parks are known to mainly promote active transportation indirectly through the presence of trails, which can allow for a shorter route to a destination than following sidewalks along city streets (National Recreation and Park

176

Association, 2020). However, in this case, other features of the neighborhood environment maybe have been more important predictors of engaging in greater hours of active transportation.

This study also found that Asian residents who lived further away from a park were more likely to engage in greater physical activity. Again, this finding could suggest that participants consequentially engage in more exercise in order to walk to a park, or their physical activity behaviors are less related to park use. As suggested, there also may be an existing threshold for the maximum distance of a park, which will promote physical activity (Kaczynski et al., 2014). Accessibility has been defined as having a park that is at least 1 kilometer (1000 meters) away from one’s home (Schipperijn et al., 2017). Given that 93.5% of participants lived within 1 km of a park, the study population had high access to a park, and this subsequently showed to have positive effects on physical activity. Nevertheless, the clinical significance of these findings was minimal, where a one-unit increase in walk time corresponded to a 6% increase in physical activity METs.

A handful of existing studies and systematic reviews have investigated whether the distance to a park (proximity) and the number of parks (density) are associated with engaging in more physical activity (Kaczynski et al., 2014; Kaczynski & Henderson,

2007; Schipperijn et al., 2017). There is stronger evidence that having more parks nearby is associated with physical activity (Bancroft et al., 2015; Kaczynski et al., 2014;

Schipperijn et al., 2017). For example, Kaczynski et al. (2014) found that the density of parks rather than the proximity to the nearest park predicted greater physical activity among participants residing in Kansas City, Missouri. As the current study aimed to examine proximity rather than density, it would be important, in future research, to

177

examine whether having several parks nearby would promote physical activity levels differently in Asian Americans.

The additional analysis further examined whether proximity to a park was associated with park use. A total of 35% of participants reported that they utilized a park to engage in physical activity. Perceived closeness to a park within one’s neighborhood was associated with greater odds of park use (OR=1.733, p<0.001), but walking distance to a park, measured by GIS, was not statistically significant (OR=1.000, p=0.508). This finding suggests that proximity to a park did not lead to increased park use but perceived measures of park proximity were more salient to determine whether participants utilized a park.

Park use was measured by asking participants whether they had walked or cycled to a park for recreation, health, or fitness purposes. This study did not capture all recreational activities that may occur in a park. The literature addresses the importance of understanding context-specific behaviors such as walking, leisure activities, and MVPA, all of which can occur in a park setting (Kaczynski et al., 2014). The finding that proximity to parks led to overall increases in physical activity levels may disregard the context-specific behavior and wrongfully assume that individuals are more physically active in a park. In a recent study, physical activity was tracked using GPS, which allowed for a measure of how much exercise was associated with park-specific physical activity (Stewart et al., 2018). In their findings, only 3% of physical activity was associated with a park, despite having five or more parks in their neighborhood (Stewart et al., 2018).

178

Closer proximity to a gym/recreational center and specialty exercise facility was anticipated to be associated with greater participation in various forms of physical activity. Living closer to a gym was not associated with higher levels of active transportation, MVPA, and overall activity. However, being closer to a specialty exercise facility, such as a yoga studio, boxing/Muy Thai gym, dance studio, etc., were associated with greater odds of walking and active transportation. On the contrary, living closer to a specialty exercise facility was associated with lower levels of active transportation and

MVPA in those who reported engaging in these activities.

These findings are counterintuitive and generally at odds with other studies in the literature. For example, a multi-ethnic study (MESA) by Diez Roux and Colleagues

(2007) found that the density of recreational resources (team/dual sports, conditioning activities, and other individual activities) was associated with engagement in more physical activity. Recently, Gidlow and colleagues (2019) studied ten cities globally to determine associations between the numbers of recreational facilities associated with non-walking Leisure Time Physical Activity (LTPA), which can be comparative to

MVPA. In their pooled analysis of studies, participants engaged in more non-walking

LTPA when they had more recreational facilities within a 0.5 and 1-km mile of their home (Gidlow et al., 2019). However, The extent to which facility proximity influenced the amount of non-walking LTPA varied by city, gender, and socioeconomic status

(Gidlow et al., 2019). For example, in Baltimore, participants with higher education were more likely to engage in physical activity, the further they lived from a park (Gidlow et al., 2019).

179

Nevertheless, the clinical significance of the current findings for proximity to a specialty exercise facility is minimal. There was only a 5.4-8.6% increase in the frequency of activity with every unit increase of distance to a facility. While perhaps unlikely to be meaningful at the individual level, such increases in activity could have population-level effects on engagement in physical activity. Thus, the equivocal findings on recreational facilities may also allude to whether participants are utilizing these facilities or not. These exercise facilities provide only standalone types of activities compared to gyms and recreational centers that provide a wide range of activities. Thus, specialty facilities may be less utilized by this study population. However, participants engaged in more walking or cycling to travel to these facilities, or as a substitute, engaged in other types of physical activity within their neighborhood (e.g., active transport, recreational walking, or cycling).

In conclusion, the hypotheses for Research Question 4 were largely not confirmed.

The nearby presence of a park or specialty exercise facility was associated with walking, active transportation, MVPA, and overall levels of physical activity. However, a closer proximity to these physical activity resources led to lower levels of these behaviors. Parks are recognized as promoting overall physical and mental well-being (Cohen et al., 2007;

Kaczynski & Henderson, 2007; Larson et al., 2016). The quality of parks and facilities can play a role in determining whether participants utilize these facilities (Cohen et al.,

2007; Kaczynski & Henderson, 2007; Larson et al., 2016). Integrating parks and recreational facilities into the community by maintaining their safety and cleanliness as well as building other nearby destinations that support active travel is crucial. In the

180

future, studies should quantify the use of parks, gym/recreational facilities, and specialty facilities while relating them to measures of proximity.

General Discussion

There were several key sociodemographic factors that may have impacted self- reported physical activity. Study participants reported having relatively low incomes in relation to their self-reported high levels of education. Given that the majority of participants were immigrants, they may have been educated in their countries of origin but faced challenges related to socioeconomic adjustment after immigrating to the US

(Trinh-Shevrin et al., 2009). When comparing income levels among ethnic groups, Asian

American immigrants earn less income than non-Hispanic Whites with the same educational status and face significant challenges to move up in the primary labor market

(Trinh-Shevrin et al., 2009). As a result, multiple family members have to contribute to the median family income (Trinh-Shevrin et al., 2009). Given these racial and ethnic inequalities, study participants could have been employed in the secondary labor market and face barriers to being physically active and accessing physical activity resources.

Study participants also reported less time engaging in moderate and vigorous leisure activities compared to walking for active transportation and recreational walking.

Engagement in lower intensity leisure activities reported by Asian Americans lends support to the notion that LTPA is viewed as a commodity that requires income and free time which may not be readily available secondary to work demands (Kandula &

Lauderdale, 2005).

The population of Asian residents has doubled in Philadelphia County since 1990

(The Pew Charitable Trusts, 2019). Participants primarily lived within certain areas of

181

Philadelphia County that are widely recognized as ethnically dense Asian neighborhoods.

These neighborhoods mainly constituted the southern and northern regions of

Philadelphia County, including zip codes 19120, 19145, 19148, and 19149. Between

2000 and 2010, these zip codes experienced a significant increase in the Asian population, with a 218.5% increase in 19145, a 277.4% increase 19148, and 596% increase in 19149 (The Pew Charitable Trusts, 2011, 2018). Zip code 19120 already contained the highest population of Asians living there in 2000, and this number increased by 20.9% in 2010 (The Pew Charitable Trusts, 2011). Although these neighborhoods have higher concentrations of Asian Americans, there are wide disparities in health indicators based on state level data (City of Philadelphia Department of Public

Health, n.d.; The Pew Charitable Trusts, 2019). Neighborhoods with poorer health and socioeconomic outcomes were Lawndale-Crescentville (19120), Oak Land-Fernrock

(19120), and Olney-Fentonville (19120), located in the northern part of the county

(Drexel University Urban Health Collaborative & City of Philadelphia Department of

Public Health, 2019). These neighborhoods were characterized as having poorer socioeconomic factors, such as lower life expectancy, lower education rates and higher commute times to travel to work in comparison to other Philadelphia neighborhoods

(Drexel University Urban Health Collaborative & City of Philadelphia Department of

Public Health, 2019). Those neighborhoods with better health outcomes were also in the northeast, Oxford circle and Mayfair-Holmesburg (19149) and in the south, including

South Broad-Girard Estates (19145) and Snyder-Whitman (19148) (Drexel University

Urban Health Collaborative & City of Philadelphia Department of Public Health, 2019).

These were characterized as having better socioeconomic outcomes, such as greater

182

education levels, median income, and overall greater health outcomes (City of

Philadelphia Department of Public Health, n.d.; Drexel University Urban Health

Collaborative & City of Philadelphia Department of Public Health, 2019). Therefore, there were widespread differences in neighborhood socioeconomic conditions where participants resided.

Relationships between sociodemographic characteristics and physical activity also were examined. Almost two-thirds of participants (64.85%) met the recommendation of

450 metabolic equivalents of aerobic activity per week (U.S. Department of Health and

Human Services, 2018). This is higher than the national average, where 50% of

Americans have reported meeting the aerobic physical activity guidelines (CDC, 2015).

Participants engaged in 5.5 hours of activity per week, but with a wide range of participation. Almost 15% of the sample (14.23%) reported more than 10 hours of activity per week, while 36.2% reported less than one hour per week. Walking within the neighborhood as a means of active transport or recreation/health/fitness was the most common activity. Participants engaged less in non-walking MVPA and cycling as forms of physical activity.

Although participants generally reported engaging in relatively high levels of physical activity, levels of activity varied by ethnic group. Vietnamese Americans engaged in more overall physical activity in comparison to Chinese and Korean Americans.

However, Chinese Americans reported higher levels of active transportation. A previous study comparing these three groups did not report overall levels of physical activity levels but found that social determinants, such as age, gender, and marital status, predicted differences in physical activity behavior (Becerra et al., 2015). Much of the

183

literature has shown that Asian Americans engage in lower levels of physical activity compared to other ethnic groups. However, the disaggregated nature of the data and lumping all Asians into one category can further mask the disparities in physical activity that are seen by ethnic group (Kandula & Lauderdale, 2005; K. Li & Wen, 2013;

Maxwell et al., 2012).

In the current study, Chinese Americans, who perceived their environment to be highly walkable, met the physical activity guidelines. Furthermore, objective measures of walkability, including the Walk Score, and proximity to parks, predicted greater engagement in physical activity among all groups. Vietnamese Americans who lived in neighborhoods with higher Walk Scores were more likely to meet the physical activity guidelines. Vietnamese Americans who lived closer to a park, but further away from a gym/recreational center and specialty exercise facility were more likely to meet the physical activity guidelines. Koreans who lived further to a park, but closer to a gym/recreation center, were more likely to meet the guidelines. Lastly, Chinese

Americans who lived closer to a specialty exercise facility were more likely to meet the physical activity guidelines. Thus, the interactive effects of ethnic group on the relationship between the built environment and physical activity suggests the significant heterogeneity that may exist based on several characteristics.

The present study controlled for significant covariates such as BMI, age, gender, and income (Hall & Ram, 2018). Older participants were consistently less likely to engage in physical activity, while participants with obesity based on Asian BMI cutoffs were more likely to be physically active. These counterintuitive findings are in contrast to the literature showing that individuals living with obesity are less physically active

184

(Dietz, 2011; Pietiläinen et al., 2008). Acculturation may explain the discrepancy

(Becerra et al., 2015; Belza et al., 2004; Yi et al., 2015, 2016). For example, having greater acculturation is linked to engaging in more physical activity, but has also been linked to obesity, which is often attributed to elements of a Western diet and food culture

(Gong et al., 2019; Novotny et al., 2009; Oster & Yung, 2010; Serafica et al., 2013).

Older Asians who have been living in the US for long periods of time frequently report being less acculturated than younger immigrants and US-born counterparts, and many report lower levels of physical activity as well as obesity (Oster & Yung, 2010; Yi et al.,

2016). These findings should be further explored in relation to the built environment and may suggest the need for targeting physical activity promotion among certain sociodemographic groups.

In summary, this study contributes to the existing literature regarding how the built environment impacts physical activity among Asian Americans. The use of validated measures such as the NEWS and objective GIS measures of proximity are methodological advances. Previous studies of Asian Americans used simplistic built environment measures which only included perceived incivilities, crime, aesthetics, diversity of destinations in the neighborhood to determine overall physical activity patterns (Bungum et al., 2012; Y. Li et al., 2015). This study identified additional aspects of the built environment and allowed for a comparison of various physical activity domains, specifically active transportation and recreational walking/cycling. The study identified that physical barriers to walking negatively affected active transportation, while greater perceived crime negatively affected recreational/walking cycling. Most related studies have primarily focused on predictors of overall physical activity and

185

walking. This study adds to that literature by having investigated how ethnic identity is related to physical activity behaviors.

Strengths and Limitations

This study examined complex relationships between perceived and objective built environment characteristics in a sample of Asian-Americans. As an often understudied population, this is a strength of the study. The sampling methods for this study were unique as they incorporated participants from community settings that were predominantly immigrants. The majority of other studies have constituted large scale epidemiological data or included participants that came from university or health care systems (Ding et al., 2013; Ferdinand et al., 2012; Kärmeniemi et al., 2018; Schipperijn et al., 2017; Smith et al., 2017). Furthermore, the study was inclusive to non-English speaking individuals, as the survey was provided in multiple Asian languages. This allowed sampling of a unique population that is traditionally excluded from studies, as a majority of built environment studies have only included English speaking participants

(Ferdinand et al., 2012; Sallis et al., 2011).

The present study also utilized valid measurement tools of the built environment, such as the NEWS and GIS measures. The study appears to be the first to investigate perceived neighborhood walkability using the NEWS and objective measures using GIS in Asian Americans. Previous studies of Asian Americans used simple built environment measures that did not capture all aspects of the neighborhood (Bungum et al., 2012; Y. Li et al., 2015). The built environment has complex associations and can be moderated or impacted by several different psychosocial as well as sociodemographic factors (Van

186

Cauwenberg et al., 2017). This study controlled for several critical socioeconomic and health indicators such as income, neighborhood self-selection, perceived crime, and BMI.

Another major strength of the study was that the built environment variables were collected concurrently with self-reported study participant measures. In several previous studies, collection of participant data preceded the collection of Walk Scores or GIS measures by several years (Chiu et al., 2015; Kelley et al., 2016; Twardzik et al., 2019).

This can introduce significant bias, given that major changes to the built environmental can occur over time.

This study is not without limitations. The study was conducted in an urban setting among certain Asian ethnic groups and cannot be generalizable to suburban or other urban populations, or to the Asian ethnic groups that were not studied. Another significant limitation was that physical activity was measured by self-report, which may have introduced recall and social desirability bias. Also, the NPAQ is an appropriate measure of context-specific physical activity behavior but does not assess physical activity that occurs at work or at home. Participants tend to over-report their physical activity behaviors when asked to recall their activity (S. A. Adams et al., 2005; Brenner

& DeLamater, 2014; Kapteyn et al., 2018; Matthews et al., 2019). In a study by Kapetyn and colleagues (2018), how one perceives their fitness levels are primarily driven by environment and culture and found major differences in objective versus reported physical activity. Thus, the use of accelerometers, coupled with geographic positioning systems (GPS) could improve the accuracy of measuring physical activity in various settings and environments (Kapteyn et al., 2018).

187

Understanding the complexities of physical activity remains a challenge. Physical activity can occur in several different settings and can be influenced by other elements of the environment (Hall & Ram, 2018; Tuckel & Milczarski, 2015; Twardzik et al., 2019).

The algorithm for the Walk Score only incorporates parks and no other recreational facilities such as gyms, bike trails, walking trails where physical activity may occur (Carr et al., 2010). Also, studies on the Walk Score have used a wide range of means to capture physical activity behavior, ranging from self-report to objective measures such as accelerometers (Hall & Ram, 2018; Tuckel & Milczarski, 2015; Twardzik et al., 2019).

Studies have found that the Walk Score and overall MVPA were inversely associated when measured by accelerometers; on the other hand, self-report measures have found less consistent associations (Hall & Ram, 2018; Tuckel & Milczarski, 2015; Twardzik et al., 2019). The current study used self-reported measures of physical activity, which could have produced different findings if accelerometers were used. Thus, the wide range of measures for physical activity presents a challenge to decipher whether the Walk Score impacts physical activity based on existing literature.

Furthermore, the study sample size was small in comparison to the total population of Asians residing in Philadelphia County, which is now estimated at around

100,000. Larger sample sizes with a more diverse Asian ethnic population are needed to more fully explore these and other research questions. Also, since the sample of participants was recruited from community-based settings, participants generally lived in ethnically dense Asian neighborhoods with a higher concentration of Asian residents

(Appendix H). In future studies, participants should be recruited from a wider range of neighborhoods.

188

One crucial factor that this study did not capture was vehicle ownership. The survey failed to account for this measure. This information would help indicate whether participants mainly drove as a means of transportation and how this impacted their physical activity level.

Although objective built environment characteristics were captured, field assessments of parks and recreational facilities were not conducted. Participants came from several zip codes, and conducting built environment audits required several resources (e.g., costs, training, etc.), which was not deemed feasible for this project

(Griew et al., 2013). More frequently, researchers have directly assessed the quality of parks and recreational facilities as well as microscale features of the environment

(Besenyi et al., 2018; Cain et al., 2014, 2017; Kaczynski et al., 2014; Millstein et al.,

2013). Parks and recreational facility audits can assess the presence and quality of resources, such as walking/biking trails, and other recreational facilities, which are essential for promoting physical activity (Kaczynski et al., 2008; Schipperijn et al.,

2013). For example, park characteristics such as which facilities they offer, their size, and their quality, were not studied here, but have been in other investigations (Evenson et al.,

2016; Kaczynski et al., 2014; Stewart et al., 2018; Suminski et al., 2012; Van

Cauwenberg et al., 2017). Microscale features of the built environment have also been measured through direct observation and help to understand whether sidewalk quality and presence of graffiti and litter within proximity to one’s home can impact their physical activity behaviors (Millstein et al., 2013).

Lastly, seasonal differences in physical activity levels were not taken into consideration. Since the period of data collection was held between July to December,

189

participants may have reported lower physical activity levels over time. Addressing these limitations in future studies will provide stronger methodology in advancing the field of the built environment in affecting physical activity behaviors, especially among ethnic minority groups.

Implications

The quality of the built environment has a substantial impact on health and well- being. This study addressed key built environmental characteristics related to physical activity in an effort to promote health among Asian Americans. In order to maximize health outcomes, environmental and policy strategies place emphasis on identifying the most modifiable built environment factors (Devarajan et al., 2020; Fraser & Lock, 2011;

Kerr et al., 2016). For example, perceived aesthetics, crime, and land use mix development were found to be important factors in this study. This is consistent with previous findings in other populations (Ding et al., 2013; M. J. Duncan et al., 2005;

Hoehner et al., 2005; Maddison et al., 2009; Peachey & Baller, 2015; Saelens, Sallis, &

Frank, 2003; Van Dyck et al., 2012).

Built environment changes related to these characteristics would provide more sustainable and cost-effective solutions than changing macro-level infrastructures such as residential density and street connectivity in older cities. Neighborhoods which have various commercial destinations near recreational facilities such as parks are more likely to promote greater physical activity. Furthermore, urban planners have incorporated the idea of designing activity-friendly communities that maximize physical activity and health benefits (Smith et al., 2017). These initiatives directly relate to active transportation and overall walking due to the profound benefits in improving

190

cardiovascular health, obesity prevention, as well as benefits to reducing traffic congestion and air pollution (V. Brown et al., 2017; Freeman et al., 2013; Wanner et al.,

2012).

The study findings suggest the need to promote active transportation and recreational walking within Asian American communities and provide adequate built environment structures to promote these behaviors. Health education interventions can also be supplemented to increase awareness of the environment for promoting physical activity.

Interventions that aim to improve health outcomes and prevent chronic disease through physical activity behavior should consider the neighborhood environment, such as whether there are parks and recreational facilities, as well as adequate sidewalks to engage in physical activity. For example, studies may utilize map layouts of one’s neighborhood to identify resources to engage in physical activity or routes for recreational walking. These characteristics should be addressed in order to ensure participants have opportunities to engage in outdoor physical activity and improve their adherence to programs designed to promote activity.

The objective built environment findings provide less strong evidence that can drive major policy-level changes in order to enhance physical activity levels among

Asian Americans. The lack of compelling results may suggest the need to further assess and identify other built environmental factors that may be affecting immigrant minority populations. These factors can be further assessed with the use of environmental health assessment tools, which inform decision making and can provide further strategies to improve built environment conditions among many sectors of public health, housing, and urban planning. Health Impact Assessments (HIA), environmental public health tracking,

191

and cumulative risk assessments are essential tools for designing both major and minor policies that improve community design and the built environment (Koehler et al., 2018).

Resulting policies have involved improvements in urban green and blue spaces (water, coastal, riverfronts), infrastructures that promote public transit, walking and cycling, and also combat the urban heat island effect in metropolitan cities (Koehler et al., 2018). The utilization of these tools to address the built environment in Asian American communities can ensure that the built environment is positively influencing health and physical activity-related outcomes.

The neighborhood environment, as well as recreational facilities and parks, provide support to engage in a variety of activities in order to meet the physical activity guidelines and, ideally, improve or maintain health (Evenson et al., 2013). Given that the study findings were inconclusive with regards to proximity to parks and recreational facilities, there may be other factors in the urban environment which were not studied directly in these communities, such as the quality of parks and recreational facilities within neighborhoods. Community asset mapping is a tool used in the field to identify recreational programs and resources that are available within the community (Burns et al., 2012). It involves interviews and surveys with community members and associations, as well as conducting a walk through the neighborhood to identify these resources, which can then be mapped out and utilized by the community (Burns et al., 2012). While not used in the present study, the tool has great potential to identify relevant resources such as recreation centers, gyms, facilities, parks, and other physical activity programs in the built environment. Resources are then displayed on a geospatial map and shared with constituents. In this regard, participants and community members are empowered to

192

improve their health behaviors but also use this to drive policies and changes in the built environment to enhance their neighborhood conditions.

Similarly, asset mapping has been shown to outline disparities within neighborhoods. An asset map of physical activity resources was created in Pasadena,

Texas, a community with poor socioeconomic conditions and high rates of obesity (Stasi et al., 2019). The research team found that the cluster of resources was located in an area of lower population density, and there was a lack of recreational facilities open to the public. Thus, asset mapping has direct implications for identifying how infrastructures can be improved and be utilized to locate ideal spots for facilities and parks or green spaces to be built in order to maximize engagement in physical activity.

In addition to communicating the main findings of the study to the scientific community via peer review publications, the study findings will be disseminated to the community organizations that recruited study participants in Philadelphia County.

Community leaders play a significant role in advocating for their communities, specifically with regards to health promotion. An executive summary of the findings, written for the lay public, will be created and shared with community leaders and members. The report will highlight the benefits of physical activity and active transportation and articulate strategies to mobilize built environment resources to engage in physical activity. The report also can be used as a tool for community organizations to utilize when advocating for improving the built environment infrastructures within their community. These initiatives may involve community members collaborating with other policy and environmental organizations or forming an academic-community partnership to further assess their neighborhood environment.

193

Future Directions

The social-ecological model has become an increasingly important framework for conducting research in the physical activity field (McLeroy et al., 1988). Saelens and colleagues (2012) emphasize the need to incorporate multiple levels of the ecological model into physical activity study designs. Although the present study assessed specific cultural characteristics, it did not capture any psychosocial factors related to exercise.

Exercise-related psychosocial mediators such as self-efficacy, motivation, and social support from family and children may significantly impact the built environment (J. A.

Carlson et al., 2012; Maddison et al., 2009; X. Wang et al., 2017). Future research on additional neighborhood-level influences, such as how living in ethnically dense neighborhoods or ethnic enclaves within urban cities affects physical activity behaviors are warranted.

In addition, further research linking the relationship between the built environment and physical activity to its influence on developing chronic diseases such as obesity, heart disease, and cancer is becoming a critically important avenue of research for ethnic minority populations. These studies should aim to identify how poorly built environments may perpetuate physical inactivity, which has negative implications on the development as well as the management of chronic diseases.

Direct observations of built environment structures in Asian American communities may further inform built environment conditions. Assessing the quality, availability, and conditions of parks and recreational facilities using field audits are needed within immigrant communities. Facilities that are less maintained or poorly kept may negatively impact their use. As examples, the Community Park Audit Tool (CPAT) and Physical Activity Resource Assessment (PARA) are heavily used in research to 194

assess the characteristics mentioned above of parks and facilities, respectively (Besenyi et al., 2018; R. E. Lee et al., 2005). Also, built environment researchers are increasingly using qualitative research to understand the built environment better, especially among ethnic minority groups. For example, photovoice has been used as a method to capture community perceptions of the built environment in order to develop community-driven policies for physical activity (Gullón et al., 2019). Given that these groups experience unique immigration history and cultural backgrounds, qualitative methods (e.g., photovoice, interviews, and focus groups) would help researchers gain insights into deciphering the complex relationships that these factors have with the built environment

(Salvo et al., 2018).

Further studies on evaluating the impact of built environment changes in metropolitan regions with Asian Americans and underrepresented populations are also needed. The recurring development of urban parks, fitness installations, and other features in Philadelphia County provides opportunities to assess the impact that these spaces have in serving ethnic minority groups. Physical activity researchers should study future parks, trails, and bike lanes that are being built in the city concerning who are using these spaces. For example, the Rail Park is a previous rail corridor that will become miles of greenway and a multipurpose trail located within Philadelphia’s Chinatown district (Friends of the Rail Park, 2020; Zhang et al., 2018). Although this project presents an innovative design in urban planning and architecture, it may not promote physical activity equitably across all populations. Penn State University researchers are currently assessing the perception of its use among those who live and work within

Chinatown (Friends of the Rail Park, 2020; Zhang et al., 2018). Future similar research is

195

needed in this area with longitudinal follow-up in various regions on how they are impacting physical activity and health outcomes.

The above factors may be taken into consideration to initiate future studies that will build on this current work. The inclusion of larger samples of Asian Americans from various urban and suburban neighborhoods will provide a more thorough analysis of the socioeconomic and built environment characteristics that affect physical activity behaviors of Asian Americans. Also, future studies should compare ethnically dense and non-ethnically dense suburban and urban neighborhoods in the Philadelphia metropolitan area to examine whether living within enclaves provides a protective effect on health behaviors. For example, ethnic enclaves continue to host immigrants; while it can be positive for new immigrants to be near their own ethnic group, there can be significant disparities in resources and facilities in these neighborhoods. In addition, working with community organizations and utilizing community asset mapping may be a future step to identifying how to leverage built environment facilities in neighborhoods to promote positive active transportation and physical activity behaviors. Furthermore, studies will address longitudinal comparisons between the built environment and its impact on chronic disease development.

The current study has potential to contribute to the literature on the built environment and how it influences physical activity in Asian Americans through two major papers. The first paper will incorporate the perceived built environment and its impact on active transportation, recreational walking/cycling, and overall physical activity. This paper will highlight the unique built environment characteristics that were important predictors of physical activity and has not been previously studied in Asian

196

Americans. The second paper will focus on how objective characteristics of the built environment, including the Walk Score, and proximity to facilities impact physical activity behaviors across the subgroups of Asian Americans. There was evident heterogeneity in findings regarding how the objective built environment influenced different physical activity behaviors of these ethnic groups, which would be a novel and important contribution to the literature in this general area.

Conclusion

Physical activity is a lifestyle behavior that provides substantial benefits to psychological and physical health and well-being. It also serves as a preventive strategy to reduce the risk of developing these chronic diseases. Asian Americans are at risk of developing chronic diseases and, in some cases, at much higher rates than other ethnic groups in the US (Adia et al., 2020). Thus, strategies to prevent chronic diseases among this population are much needed and have focused on physical activity promotion and maintenance. Physical activity promotion strategies that are solely addressed at the individual level are insufficient for improving physical activity (Green, Richard, &

Potvin, 1996; McLeroy et al., 1988; Sallis et al., 2008). The built environment continues to play a major role in maintaining and sustaining physical activity behaviors that individuals participate in within their neighborhoods.

This study demonstrates that perceived neighborhood walkability has a profound impact on physical activity in Asian Americans. More specifically, having wide and accessible destinations in one’s neighborhood, known as the land-use mix, promotes overall physical activity. Greater aesthetically pleasing environments and areas with cul- de-sacs can provide spaces to engage in various forms of physical activity. Ethnic identity

197

moderated the relationship between the perceived built environment and active transportation and moderate-intensity physical activity, but not overall physical activity.

Mixed evidence was found for the influence of Walk Score and proximity to parks and recreational facilities as predictors of physical activity. While this study provides unique findings among a predominantly Asian American immigrant and non-English speaking population, the study had some methodological limitations, such as smaller sample size and using self-report to measure physical activity. Nevertheless, the major findings can be used to inform environmental and policy strategies to maximize active transportation and recreational based physical activity in order to sustain the health of Asian-American immigrants living in urban areas.

198

REFERENCES CITED

Adams, M. A., Frank, L. D., Schipperijn, J., Smith, G., Chapman, J., Christiansen, L. B., Coffee, N., Salvo, D., du Toit, L., Dygrýn, J., Hino, A. A. F., Lai, P., Mavoa, S., Pinzón, J. D., Van de Weghe, N., Cerin, E., Davey, R., Macfarlane, D., Owen, N., & Sallis, J. F. (2014). International variation in neighborhood walkability, transit, and recreation environments using geographic information systems: The IPEN adult study. International Journal of Health Geographics, 13. https://doi.org/10.1186/1476-072X-13-43

Adams, M. A., Ryan, S., Kerr, J., Sallis, J. F., Patrick, K., Frank, L. D., & Norman, G. J. (2009). Validation of the Neighborhood Environment Walkability Scale (NEWS) Items Using Geographic Information Systems. Journal of Physical Activity & Health, 6, S113–S123.

Adams, S. A., Matthews, C. E., Ebbeling, C. B., Moore, C. G., Cunningham, J. E., Fulton, J., & Hebert, J. R. (2005). The Effect of Social Desirability and Social Approval on Self-Reports of Physical Activity. American Journal of Epidemiology, 161(4), 389–398. https://doi.org/10.1093/aje/kwi054

Adia, A. C., Nazareno, J., Operario, D., & Ponce, N. A. (2020). Health Conditions, Outcomes, and Service Access Among Filipino, Vietnamese, Chinese, Japanese, and Korean Adults in California, 2011–2017. American Journal of Public Health, e1–e7. https://doi.org/10.2105/AJPH.2019.305523

Ainsworth, B. E., Haskell, W. L., Herrmann, S. D., Meckes, N., Bassett, D. R., Tudor- Locke, C., Greer, J. L., Vezina, J., Whitt-Glover, M. C., & Leon, A. S. (2011). 2011 Compendium of Physical Activities: A second update of codes and MET values. Medicine and Science in Sports and Exercise, 43(8), 1575–1581. https://doi.org/10.1249/MSS.0b013e31821ece12

Akers, A., Barton, J., Cossey, R., Gainsford, P., Griffin, M., & Micklewright, D. (2012). Visual color perception in green exercise: Positive effects on mood and perceived exertion. Environmental Science & Technology, 46(16), 8661–8666. https://doi.org/10.1021/es301685g

An, R., Xiang, X., Yang, Y., & Yan, H. (2016). Mapping the Prevalence of Physical Inactivity in U.S. States, 1984-2015. PLOS ONE, 11(12), e0168175. https://doi.org/10.1371/journal.pone.0168175

August, K. J., & Sorkin, D. H. (2011). Racial/ethnic disparities in exercise and dietary behaviors of middle-aged and older adults. Journal of General Internal Medicine, 26(3), 245–250. https://doi.org/10.1007/s11606-010-1514-7

199

Bancroft, C., Joshi, S., Rundle, A., Hutson, M., Chong, C., Weiss, C. C., Genkinger, J., Neckerman, K., & Lovasi, G. (2015). Association of proximity and density of parks and objectively measured physical activity in the United States: A systematic review. Social Science & Medicine, 138, 22–30. https://doi.org/10.1016/j.socscimed.2015.05.034

Baran, P. K., Smith, W. R., Moore, R. C., Floyd, M. F., Bocarro, J. N., Cosco, N. G., & Danninger, T. M. (2014). Park Use Among Youth and Adults: Examination of Individual, Social, and Urban Form Factors. Environment and Behavior, 46(6), 768–800. https://doi.org/10.1177/0013916512470134

Barton, J., & Pretty, J. (2010). What is the Best Dose of Nature and Green Exercise for Improving Mental Health? A Multi-Study Analysis. Environmental Science & Technology, 44(10), 3947–3955. https://doi.org/10.1021/es903183r

Bauman, A., & Bull, F. C. (2007). Environmental Correlates of Physical Activity And Walking in Adults and Children: A Review of Reviews.

Bauman, A., Bull, F., Chey, T., Craig, C. L., Ainsworth, B. E., Sallis, J. F., Bowles, H. R., Hagstromer, M., Sjostrom, M., Pratt, M., & The IPS Group. (2009). The International Prevalence Study on Physical Activity: Results from 20 countries. International Journal of Behavioral Nutrition and Physical Activity, 6(1), 21. https://doi.org/10.1186/1479-5868-6-21

Bauman, A. E., Reis, R. S., Sallis, J. F., Wells, J. C., Loos, R. J., & Martin, B. W. (2012). Correlates of physical activity: Why are some people physically active and others not? The Lancet, 380(9838), 258–271. https://doi.org/10.1016/S0140- 6736(12)60735-1

Becerra, M. B., Herring, P., Marshak, H. H., & Banta, J. E. (2015). Social Determinants of Physical Activity Among Adult Asian-Americans: Results from a Population- Based Survey in California. Journal of Immigrant and Minority Health, 17(4), 1061–1069. https://doi.org/10.1007/s10903-014-0074-z

Belza, B., Walwick, J., Shiu-Thornton, S., Schwartz, S., Taylor, M., & LoGerfo, J. (2004). Older adult perspectives on physical activity and exercise: Voices from multiple cultures. Preventing Chronic Disease, 1(4), A09.

Berry, B., & Berry, J. W. (2005). Acculturation: Living successfully in two cultures. International Journal of Intercultural Relations, 29(6), 697–712.

Berry, J. W. (1997). Immigration, Acculturation, and Adaptation. Applied Psychology, 46(1), 5–34. https://doi.org/10.1111/j.1464-0597.1997.tb01087.x

Berry, J. W., Phinney, J. S., Sam, D. L., & Vedder, P. (2006). Immigrant Youth: Acculturation, Identity, and Adaptation. Applied Psychology, 55(3), 303–332. https://doi.org/10.1111/j.1464-0597.2006.00256.x

200

Besenyi, G. M., Schooley, B., Turner-McGrievy, G. M., Wilcox, S., Wilhelm Stanis, S. A., & Kaczynski, A. T. (2018). The Electronic Community Park Audit Tool (eCPAT): Exploring the Use of Mobile Technology for Youth Empowerment and Advocacy for Healthy Community Policy, Systems, and Environmental Change. Frontiers in Public Health, 6. https://doi.org/10.3389/fpubh.2018.00332

Bicycle Coalition of Greater Philadelphia. (2020). The Regional Bike Map. https://bicyclecoalition.org/resources/bike-maps/

Blackwell, D. L., Lucas, J. W., & Clarke, T. C. (2014). Summary health statistics for U.S. adults: National health interview survey, 2012. Vital and Health Statistics. Series 10, Data from the National Health Survey, 260, 1–161.

Blair, S. N., Goodyear, N. N., Gibbons, L. W., & Cooper, K. H. (1984). Physical Fitness and Incidence of Hypertension in Healthy Normotensive Men and Women. JAMA, 252(4), 487–490. https://doi.org/10.1001/jama.1984.03350040017014

Boarnet, M. G., Greenwald, M., & McMillan, T. E. (2008). Walking, Urban Design, and Health: Toward a Cost-Benefit Analysis Framework. Journal of Planning Education and Research, 27(3), 341–358. https://doi.org/10.1177/0739456X07311073

Boone, C. G., Buckley, G. L., Grove, J. M., & Sister, C. (2009). Parks and People: An Environmental Justice Inquiry in Baltimore, Maryland. Annals of the Association of American Geographers, 99(4), 767–787. https://doi.org/10.1080/00045600903102949

Boone, J. E., Gordon-Larsen, P., Stewart, J. D., & Popkin, B. M. (2008). Validation of a GIS facilities database: Quantification and implications of error. Annals of Epidemiology, 18(5), 371–377. https://doi.org/10.1016/j.annepidem.2007.11.008

Bornstein, D. B., & Davis, W. J. (2014). The Transportation Profession’s Role in Improving Public Health. Institute of Transportation Engineers. ITE Journal; Washington, 84(7), 18–24.

Brenner, P. S., & DeLamater, J. D. (2014). Social Desirability Bias in Self-reports of Physical Activity: Is an Exercise Identity the Culprit? Social Indicators Research, 117(2), 489–504. JSTOR.

Bronfenbrenner, U. (1979). The Ecology of Human Development. Harvard University Press.

Brown, J. C., Winters-Stone, K., Lee, A., & Schmitz, K. H. (2012). Cancer, Physical Activity, and Exercise. Comprehensive Physiology, 2(4), 2775–2809. https://doi.org/10.1002/cphy.c120005

201

Brown, S. C., Pantin, H., Lombard, J., Toro, M., Huang, S., Plater-Zyberk, E., Perrino, T., Perez-Gomez, G., Barrera-Allen, L., & Szapocznik, J. (2013). Walk Score®: Associations with Purposive Walking in Recent Cuban Immigrants. American Journal of Preventive Medicine, 45(2), 202–206. https://doi.org/10.1016/j.amepre.2013.03.021

Brown, S. D., Ehrlich, S. F., Kubo, A., Tsai, A.-L., Hedderson, M. M., Quesenberry, C. P., & Ferrara, A. (2016). Lifestyle behaviors and ethnic identity among diverse women at high risk for type 2 diabetes. Social Science & Medicine, 160, 87–93. https://doi.org/10.1016/j.socscimed.2016.05.024

Brown, S. D., Unger Hu, K. A., Mevi, A. A., Hedderson, M. M., Shan, J., Quesenberry, C. P., & Ferrara, A. (2014). The Multigroup Ethnic Identity Measure—Revised: Measurement invariance across racial and ethnic groups. Journal of Counseling Psychology, 61(1), 154–161. https://doi.org/10.1037/a0034749

Brown, V., Moodie, M., Mantilla Herrera, A. M., Veerman, J. L., & Carter, R. (2017). Active transport and obesity prevention—A transportation sector obesity impact scoping review and assessment for Melbourne, Australia. Preventive Medicine, 96, 49–66. https://doi.org/10.1016/j.ypmed.2016.12.020

Brownson, R. C., Chang, J. J., Eyler, A. A., Ainsworth, B. E., Kirtland, K. A., Saelens, B. E., & Sallis, J. F. (2004). Measuring the environment for friendliness toward physical activity: A comparison of the reliability of 3 questionnaires. American Journal of Public Health, 94(3), 473–483. https://doi.org/10.2105/ajph.94.3.473

Brownson, R. C., Hoehner, C. M., Day, K., Forsyth, A., & Sallis, J. F. (2009). Measuring the Built Environment for Physical Activity. American Journal of Preventive Medicine, 36(4 Suppl), S99-123.e12. https://doi.org/10.1016/j.amepre.2009.01.005

Bruton, C. M., & Floyd, M. F. (2014). Disparities in built and natural features of urban parks: Comparisons by neighborhood level race/ethnicity and income. Journal of Urban Health: Bulletin of the New York Academy of Medicine, 91(5), 894–907. https://doi.org/10.1007/s11524-014-9893-4

Bungum, T. J., Landers, M., Azzarelli, M., & Moonie, S. (2012). Perceived environmental physical activity correlates among Asian Pacific Islander Americans. Journal of Physical Activity & Health, 9(8), 1098–1104.

Burbidge, S., & Goulias, K. (2009). Active travel behavior. Transportation Letters, 1(2), 147–167. https://doi.org/10.3328/TL.2009.01.02.147-167

Burns, J., Paul, D., & Paz, S. (2012). Participatory Asset Mapping: A Community Lab Research Toolkit. http://www.communityscience.com/knowledge4equity/AssetMappingToolkit.pdf

202

Cain, K. L., Gavand, K. A., Conway, T. L., Geremia, C. M., Millstein, R. A., Frank, L. D., Saelens, B. E., Adams, M. A., Glanz, K., King, A. C., & Sallis, J. F. (2017). Developing and Validating an Abbreviated Version of the Microscale Audit for Pedestrian Streetscapes (MAPS-Abbreviated). Journal of Transport & Health, 5, 84–96. https://doi.org/10.1016/j.jth.2017.05.004

Cain, K. L., Millstein, R. A., Sallis, J. F., Conway, T. L., Gavand, K. A., Frank, L. D., Saelens, B. E., Geremia, C. M., Chapman, J., Adams, M. A., Glanz, K., & King, A. C. (2014). Contribution of streetscape audits to explanation of physical activity in four age groups based on the Microscale Audit of Pedestrian Streetscapes (MAPS). Social Science & Medicine (1982), 116, 82–92. https://doi.org/10.1016/j.socscimed.2014.06.042

Cao, X., Handy, S. L., & Mokhtarian, P. L. (2006). The Influences of the Built Environment and Residential Self-Selection on Pedestrian Behavior: Evidence from Austin, TX. Transportation, 33(1), 1–20. https://doi.org/10.1007/s11116- 005-7027-2

Carlson, J. A., Sallis, J. F., Conway, T. L., Saelens, B. E., Frank, L. D., Kerr, J., Cain, K. L., & King, A. C. (2012). Interactions between psychosocial and built environment factors in explaining older adults’ physical activity. Preventive Medicine, 54(1), 68–73. https://doi.org/10.1016/j.ypmed.2011.10.004

Carlson, S. A., Adams, E. K., Yang, Z., & Fulton, J. E. (2018). Percentage of Deaths Associated With Inadequate Physical Activity in the United States. Preventing Chronic Disease, 15, E38. https://doi.org/10.5888/pcd18.170354

Carlson, S. A., Fulton, J. E., Pratt, M., Yang, Z., & Adams, E. K. (2015). Inadequate physical activity and health care expenditures in the United States. Progress in Cardiovascular Diseases, 57(4), 315–323. https://doi.org/10.1016/j.pcad.2014.08.002

Carr, L. J., Dunsiger, S. I., & Marcus, B. H. (2010). Walk ScoreTM As a Global Estimate of Neighborhood Walkability. American Journal of Preventive Medicine, 39(5), 460–463. https://doi.org/10.1016/j.amepre.2010.07.007

Carr, L. J., Dunsiger, S. I., & Marcus, B. H. (2011). Validation of Walk Score for estimating access to walkable amenities. British Journal of Sports Medicine, 45(14), 1144–1148. https://doi.org/10.1136/bjsm.2009.069609

CDC. (2015). Physical Activity Basics. Physical Activity Basics. https://www.cdc.gov/physicalactivity/basics/index.htm

Cerin, E., Barnett, A., Sit, C. H. P., Cheung, M., Lee, L. J., Ho, S., & Chan, W. (2011). Measuring walking within and outside the neighborhood in Chinese elders: Reliability and validity. BMC Public Health, 11, 851. https://doi.org/10.1186/1471-2458-11-851

203

Cerin, E., Cain, K. L., Conway, T. L., Van Dyck, D., Hinckson, E., Schipperijn, J., De Bourdeaudhuij, I., Owen, N., Davey, R. C., Hino, A. A. F., Mitáš, J., Orzanco- Garralda, R., Salvo, D., Sarmiento, O. L., Christiansen, L. B., Macfarlane, D. J., Schofield, G., & Sallis, J. F. (2014). Neighborhood environments and objectively measured physical activity in 11 countries. Medicine and Science in Sports and Exercise, 46(12), 2253–2264. https://doi.org/10.1249/MSS.0000000000000367

Cerin, E., Conway, T. L., Cain, K. L., Kerr, J., De Bourdeaudhuij, I., Owen, N., Reis, R. S., Sarmiento, O. L., Hinckson, E. A., Salvo, D., Christiansen, L. B., MacFarlane, D. J., Davey, R., Mitáš, J., Aguinaga-Ontoso, I., & Sallis, J. F. (2013). Sharing good NEWS across the world: Developing comparable scores across 12 countries for the neighborhood environment walkability scale (NEWS). BMC Public Health, 13(1), 309. https://doi.org/10.1186/1471-2458-13-309

Cerin, E., Leslie, E., Toit, L. du, Owen, N., & Frank, L. D. (2007). Destinations that matter: Associations with walking for transport. Health & Place, 13(3), 713–724. https://doi.org/10.1016/j.healthplace.2006.11.002

Cerin, E., Nathan, A., van Cauwenberg, J., Barnett, D. W., Barnett, A., & Council on Environment and Physical Activity (CEPA) – Older Adults working group. (2017). The neighbourhood physical environment and active travel in older adults: A systematic review and meta-analysis. The International Journal of Behavioral Nutrition and Physical Activity, 14(1), 15. https://doi.org/10.1186/s12966-017-0471-5

Cerin, E., Saelens, B. E., Sallis, J. F., & Frank, L. D. (2006). Neighborhood Environment Walkability Scale: Validity and development of a short form. Medicine and Science in Sports and Exercise, 38(9), 1682–1691. https://doi.org/10.1249/01.mss.0000227639.83607.4d

Cerin, E., Sit, C. H., Cheung, M., Ho, S., Lee, L. J., & Chan, W. (2010). Reliable and valid NEWS for Chinese seniors: Measuring perceived neighborhood attributes related to walking. The International Journal of Behavioral Nutrition and Physical Activity, 7, 84. https://doi.org/10.1186/1479-5868-7-84

Cerin, E., Sit, C. H. P., Barnett, A., Cheung, M., & Chan, W. (2013). Walking for Recreation and Perceptions of the Neighborhood Environment in Older Chinese Urban Dwellers. Journal of Urban Health, 90(1), 56–66. https://doi.org/10.1007/s11524-012-9704-8

Cerin, E., Vandelanotte, C., Leslie, E., & Merom, D. (2008). Recreational facilities and leisure-time physical activity: An analysis of moderators and self-efficacy as a mediator. Health Psychology: Official Journal of the Division of Health Psychology, American Psychological Association, 27(2S), S126-135. https://doi.org/10.1037/0278-6133.27.2(Suppl.).S126

204

Charmaraman, L., & Grossman, J. M. (2010). Importance of race and ethnicity: An exploration of Asian, Black, Latino, and multiracial adolescent identity. Cultural Diversity & Ethnic Minority Psychology, 16(2), 144–151. https://doi.org/10.1037/a0018668

Chatman, D. G. (2009). Residential Choice, the Built Environment, and Nonwork Travel: Evidence Using New Data and Methods. Environment and Planning A: Economy and Space, 41(5), 1072–1089. https://doi.org/10.1068/a4114

Chiu, M., Shah, B. R., Maclagan, L. C., Rezai, M.-R., Austin, P. C., & Tu, J. V. (2015). Walk Score® and the prevalence of utilitarian walking and obesity among Ontario adults: A cross-sectional study. Health Reports, 26(7), 3–10.

Choi, J., Lee, M., Lee, J., Kang, D., & Choi, J.-Y. (2017). Correlates associated with participation in physical activity among adults: A systematic review of reviews and update. BMC Public Health, 17(1), 356. https://doi.org/10.1186/s12889-017- 4255-2

City of Philadelphia Department of Public Health. (n.d.). Health of the City 2019. file:///C:/Users/tug31272/Downloads/Health_of_City_2019-FINAL.pdf

Cohen, D. A., Golinelli, D., Williamson, S., Sehgal, A., Marsh, T., & McKenzie, T. L. (2009). Effects of park improvements on park use and physical activity: Policy and programming implications. American Journal of Preventive Medicine, 37(6), 475–480. https://doi.org/10.1016/j.amepre.2009.07.017

Cohen, D. A., McKenzie, T. L., Sehgal, A., Williamson, S., Golinelli, D., & Lurie, N. (2007). Contribution of public parks to physical activity. American Journal of Public Health, 97(3), 509–514. https://doi.org/10.2105/AJPH.2005.072447

Cole, R., Dunn, P., Hunter, I., Owen, N., & Sugiyama, T. (2015). Walk Score and Australian adults’ home-based walking for transport. Health & Place, 35, 60–65. https://doi.org/10.1016/j.healthplace.2015.06.011

Commodore-Mensah, Y., Selvin, E., Aboagye, J., Turkson-Ocran, R.-A., Li, X., Himmelfarb, C. D., Ahima, R. S., & Cooper, L. A. (2018). Hypertension, overweight/obesity, and diabetes among immigrants in the United States: An analysis of the 2010–2016 National Health Interview Survey. BMC Public Health, 18(1), 773. https://doi.org/10.1186/s12889-018-5683-3

Commodore-Mensah, Y., Ukonu, N., Obisesan, O., Aboagye, J. K., Agyemang, C., Reilly, C. M., Dunbar, S. B., & Okosun, I. S. (2016). Length of Residence in the United States is Associated With a Higher Prevalence of Cardiometabolic Risk Factors in Immigrants: A Contemporary Analysis of the National Health Interview Survey. Journal of the American Heart Association, 5(11). https://doi.org/10.1161/JAHA.116.004059

205

Conroy, S. M., Shariff-Marco, S., Koo, J., Yang, J., Keegan, T. H. M., Sangaramoorthy, M., Hertz, A., Nelson, D. O., Cockburn, M., Satariano, W. A., Yen, I. H., Ponce, N. A., John, E. M., & Gomez, S. L. (2017). Racial/Ethnic Differences in the Impact of Neighborhood Social and Built Environment on Breast Cancer Risk: The Neighborhoods and Breast Cancer Study. Cancer Epidemiology and Prevention Biomarkers, 26(4), 541–552. https://doi.org/10.1158/1055-9965.EPI- 16-0935

Dahmann, N., Wolch, J., Joassart-Marcelli, P., Reynolds, K., & Jerrett, M. (2010). The active city? Disparities in provision of urban public recreation resources. Health & Place, 16(3), 431–445. https://doi.org/10.1016/j.healthplace.2009.11.005 de Sa, E., & Ardern, C. I. (2014). Associations between the built environment, total, recreational, and transit-related physical activity. BMC Public Health, 14, 693. https://doi.org/10.1186/1471-2458-14-693

Devarajan, R., Prabhakaran, D., & Goenka, S. (2020). Built environment for physical activity—An urban barometer, surveillance, and monitoring. Obesity Reviews, 21(1), e12938. https://doi.org/10.1111/obr.12938

Dhakal, N., Cherry, C. R., Ling, Z., & Azad, M. (2018). Using CyclePhilly data to assess wrong-way riding of cyclists in Philadelphia. Journal of Safety Research, 67, 145–153. https://doi.org/10.1016/j.jsr.2018.10.004

Dietz, W. H. (2011). Reversing the tide of obesity. Lancet (London, England), 378(9793), 744–746. https://doi.org/10.1016/S0140-6736(11)61218-X

Diez Roux, A. V., Evenson, K. R., McGinn, A. P., Brown, D. G., Moore, L., Brines, S., & Jacobs, D. R. (2007). Availability of Recreational Resources and Physical Activity in Adults. American Journal of Public Health, 97(3), 493–499. https://doi.org/10.2105/AJPH.2006.087734

Ding, D., Adams, M. A., Sallis, J. F., Norman, G. J., Hovell, M. F., Chambers, C. D., Hofstetter, C. R., Bowles, H. R., Hagströmer, M., Craig, C. L., Gomez, L. F., Bourdeaudhuij, I. D., Macfarlane, D. J., Ainsworth, B. E., Bergman, P., Bull, F. C., Carr, H., Klasson-Heggebo, L., Inoue, S., … Bauman, A. E. (2013). Perceived neighborhood environment and physical activity in 11 countries: Do associations differ by country? International Journal of Behavioral Nutrition and Physical Activity, 10(1), 57. https://doi.org/10.1186/1479-5868-10-57

Douglas, J. A., Briones, M. D., Bauer, E. Z., Trujillo, M., Lopez, M., & Subica, A. M. (2018). Social and environmental determinants of physical activity in urban parks: Testing a neighborhood disorder model. Preventive Medicine, 109, 119–124. https://doi.org/10.1016/j.ypmed.2018.01.013

206

Drexel University Urban Health Collaborative, & City of Philadelphia Department of Public Health. (2019). Closer to Home: The Health of Philadelphia’s Neighborhoods. https://www.phila.gov/media/20190801133844/Neighborhood- Rankings_7_31_19.pdf

Duncan, D. T. (2013). What’s your Walk Score®?: Web-based neighborhood walkability assessment for health promotion and disease prevention. American Journal of Preventive Medicine, 45(2), 244–245. https://doi.org/10.1016/j.amepre.2013.04.008

Duncan, D. T., Aldstadt, J., Whalen, J., & Melly, S. J. (2013). Validation of Walk Scores and Transit Scores for estimating neighborhood walkability and transit availability: A small-area analysis. GeoJournal, 78(2), 407–416. https://doi.org/10.1007/s10708-011-9444-4

Duncan, D. T., Aldstadt, J., Whalen, J., Melly, S. J., & Gortmaker, S. L. (2011). Validation of Walk Score® for Estimating Neighborhood Walkability: An Analysis of Four US Metropolitan Areas. International Journal of Environmental Research and Public Health, 8(11), 4160–4179. https://doi.org/10.3390/ijerph8114160

Duncan, M. J., Spence, J. C., & Mummery, W. K. (2005). Perceived environment and physical activity: A meta-analysis of selected environmental characteristics. International Journal of Behavioral Nutrition and Physical Activity, 2(1), 11. https://doi.org/10.1186/1479-5868-2-11

Environmental Systems Research Institute, Inc. (2020). ArcMap Documentation. https://desktop.arcgis.com/en/arcmap/

Eriksson, U., Arvidsson, D., & Sundquist, K. (2012). Availability of exercise facilities and physical activity in 2,037 adults: Cross-sectional results from the Swedish neighborhood and physical activity (SNAP) study. BMC Public Health, 12, 607. https://doi.org/10.1186/1471-2458-12-607

ESRI. (2018). Methodology Statement: 2018 US Business Locations and Business Summary Data. http://downloads.esri.com/esri_content_doc/dbl/us/J9940_Methodology_Stmt_Bu siness_Locations_Summary_2018.pdf

ESRI. (2020). Find Nearest Tool. https://doc.arcgis.com/en/arcgis-online/analyze/find- nearest.htm

Estabrooks, P. A., Lee, R. E., & Gyurcsik, N. C. (2003). Resources for physical activity participation: Does availability and accessibility differ by neighborhood socioeconomic status? Annals of Behavioral Medicine, 25(2), 100–104. https://doi.org/10.1207/S15324796ABM2502_05

207

Evenson, K. R., Jones, S. A., Holliday, K. M., Cohen, D. A., & McKenzie, T. L. (2016). Park characteristics, use, and physical activity: A review of studies using SOPARC (System for Observing Play and Recreation in Communities). Preventive Medicine, 86, 153–166. https://doi.org/10.1016/j.ypmed.2016.02.029

Evenson, K. R., Wen, F., Hillier, A., & Cohen, D. A. (2013). Assessing the contribution of parks to physical activity using global positioning system and accelerometry. Medicine and Science in Sports and Exercise, 45(10), 1981–1987. https://doi.org/10.1249/MSS.0b013e318293330e

Evenson, K. R., Wen, F., Metzger, J. S., & Herring, A. H. (2015). Physical activity and sedentary behavior patterns using accelerometry from a national sample of United States adults. International Journal of Behavioral Nutrition and Physical Activity, 12(1), 20. https://doi.org/10.1186/s12966-015-0183-7

Ewing, R. (2005). Can the physical environment determine physical activity levels? Exercise and Sport Sciences Reviews, 33(2), 69–75.

Ewing, R., Meakins, G., Hamidi, S., & Nelson, A. C. (2014). Relationship between urban sprawl and physical activity, obesity, and morbidity – Update and refinement. Health & Place, 26, 118–126. https://doi.org/10.1016/j.healthplace.2013.12.008

Ewing, R., Schmid, T., Killingsworth, R., Zlot, A., & Raudenbush, S. (2003). Relationship between urban sprawl and physical activity, obesity, and morbidity. American Journal of Health Promotion: AJHP, 18(1), 47–57. https://doi.org/10.4278/0890-1171-18.1.47

Ferdinand, A. O., Sen, B., Rahurkar, S., Engler, S., & Menachemi, N. (2012). The relationship between built environments and physical activity: A systematic review. American Journal of Public Health, 102(10), e7–e13. https://doi.org/10.2105/AJPH.2012.300740

Flowers, E. P., Freeman, P., & Gladwell, V. F. (2017). The Development of Three Questionnaires to Assess Beliefs about Green Exercise. International Journal of Environmental Research and Public Health; Basel, 14(10), 1172. http://dx.doi.org.libproxy.temple.edu/10.3390/ijerph14101172

Floyd, M. F., Spengler, J. O., Maddock, J. E., Gobster, P. H., & Suau, L. J. (2008). Park- based physical activity in diverse communities of two U.S. cities. An observational study. American Journal of Preventive Medicine, 34(4), 299–305. https://doi.org/10.1016/j.amepre.2008.01.009

Focht, B. C. (2009). Brief Walks in Outdoor and Laboratory Environments: Effects on Affective Responses, Enjoyment, and Intentions to Walk for Exercise. Research Quarterly for Exercise and Sport, 80(3), 611–620.

208

Forsyth, A., Schmitz, K. H., Oakes, M., Zimmerman, J., & Koepp, J. (2006). Standards for Environmental Measurement Using GIS: Toward a Protocol for Protocols. Journal of Physical Activity & Health, 3(s1), S241–S257. https://doi.org/10.1123/jpah.3.s1.s241

Frank, L. D., Andresen, M. A., & Schmid, T. L. (2004). Obesity relationships with community design, physical activity, and time spent in cars. American Journal of Preventive Medicine, 27(2), 87–96. https://doi.org/10.1016/j.amepre.2004.04.011

Frank, L. D., Saelens, B. E., Powell, K. E., & Chapman, J. E. (2007). Stepping towards causation: Do built environments or neighborhood and travel preferences explain physical activity, driving, and obesity? Social Science & Medicine (1982), 65(9), 1898–1914. https://doi.org/10.1016/j.socscimed.2007.05.053

Frank, L. D., Sallis, J. F., Saelens, B. E., Leary, L., Cain, K., Conway, T. L., & Hess, P. M. (2010). The development of a walkability index: Application to the Neighborhood Quality of Life Study. British Journal of Sports Medicine, 44(13), 924–933. https://doi.org/10.1136/bjsm.2009.058701

Frank, L. D., Schmid, T. L., Sallis, J. F., Chapman, J., & Saelens, B. E. (2005). Linking objectively measured physical activity with objectively measured urban form: Findings from SMARTRAQ. American Journal of Preventive Medicine, 28(2 Suppl 2), 117–125. https://doi.org/10.1016/j.amepre.2004.11.001

Fraser, S. D. S., & Lock, K. (2011). Cycling for transport and public health: A systematic review of the effect of the environment on cycling. European Journal of Public Health, 21(6), 738–743. https://doi.org/10.1093/eurpub/ckq145

Freeman, L., Neckerman, K., Schwartz-Soicher, O., Quinn, J., Richards, C., Bader, M. D. M., Lovasi, G., Jack, D., Weiss, C., Konty, K., Arno, P., Viola, D., Kerker, B., & Rundle, A. G. (2013). Neighborhood Walkability and Active Travel (Walking and Cycling) in New York City. Journal of Urban Health : Bulletin of the New York Academy of Medicine, 90(4), 575–585. https://doi.org/10.1007/s11524-012-9758- 7

Friends of the Rail Park. (2020). The Rail Park. Friends of the Rail Park. https://www.therailpark.org/

Gebel, K., Bauman, A., & Owen, N. (2009). Correlates of non-concordance between perceived and objective measures of walkability. Annals of Behavioral Medicine: A Publication of the Society of Behavioral Medicine, 37(2), 228–238. https://doi.org/10.1007/s12160-009-9098-3

Ghani, F., Rachele, J. N., Loh, V. H., Washington, S., & Turrell, G. (2019). Do Differences in Social Environments Explain Gender Differences in Recreational Walking across Neighbourhoods? International Journal of Environmental Research and Public Health, 16(11), 1980. https://doi.org/10.3390/ijerph16111980

209

Gidlow, C., Cerin, E., Sugiyama, T., Adams, M. A., Mitas, J., Akram, M., Reis, R. S., Davey, R., Troelsen, J., Schofield, G., & Sallis, J. F. (2019). Objectively measured access to recreational destinations and leisure-time physical activity: Associations and demographic moderators in a six-country study. Health & Place, 59, 102196. https://doi.org/10.1016/j.healthplace.2019.102196

Giles-Corti, B., Broomhall, M. H., Knuiman, M., Collins, C., Douglas, K., Ng, K., Lange, A., & Donovan, R. J. (2005). Increasing walking: How important is distance to, attractiveness, and size of public open space? American Journal of Preventive Medicine, 28(2 Suppl 2), 169–176. https://doi.org/10.1016/j.amepre.2004.10.018

Giles-Corti, B., & Donovan, R. J. (2002a). The relative influence of individual, social and physical environment determinants of physical activity. Social Science & Medicine (1982), 54(12), 1793–1812.

Giles-Corti, B., & Donovan, R. J. (2002b). Socioeconomic status differences in recreational physical activity levels and real and perceived access to a supportive physical environment. Preventive Medicine, 35(6), 601–611.

Giles-Corti, B., Timperio, A., Cutt, H., Pikora, T. J., Bull, F. C. L., Knuiman, M., Bulsara, M., Van Niel, K., & Shilton, T. (2006). Development of a reliable measure of walking within and outside the local neighborhood: RESIDE’s Neighborhood Physical Activity Questionnaire. Preventive Medicine, 42(6), 455– 459. https://doi.org/10.1016/j.ypmed.2006.01.019

Gladwell, V. F., Brown, D. K., Wood, C., Sandercock, G. R., & Barton, J. L. (2013). The great outdoors: How a green exercise environment can benefit all. Extreme Physiology & Medicine, 2, 3. https://doi.org/10.1186/2046-7648-2-3

Golden, S. D., & Earp, J. A. L. (2012). Social ecological approaches to individuals and their contexts: Twenty years of health education & behavior health promotion interventions. Health Education & Behavior: The Official Publication of the Society for Public Health Education, 39(3), 364–372. https://doi.org/10.1177/1090198111418634

Gong, S., Wang, K., Li, Y., & Alamian, A. (2019). The influence of immigrant generation on obesity among Asian Americans in California from 2013 to 2014. PloS One, 14(2), e0212740. https://doi.org/10.1371/journal.pone.0212740

Gordon-Larsen, P., Nelson, M. C., Page, P., & Popkin, B. M. (2006). Inequality in the built environment underlies key health disparities in physical activity and obesity. Pediatrics, 117(2), 417–424. https://doi.org/10.1542/peds.2005-0058

Green, L. W., Richard, L., & Potvin, L. (1996). Ecological foundations of health promotion. American Journal of Health Promotion: AJHP, 10(4), 270–281. https://doi.org/10.4278/0890-1171-10.4.270

210

Griew, P., Hillsdon, M., Foster, C., Coombes, E., Jones, A., & Wilkinson, P. (2013). Developing and testing a street audit tool using Google Street View to measure environmental supportiveness for physical activity. International Journal of Behavioral Nutrition and Physical Activity, 10(1), 103. https://doi.org/10.1186/1479-5868-10-103

Gullón, P., Díez, J., Conde, P., Ramos, C., Márquez, V., Badland, H., Escobar, F., & Franco, M. (2019). Using Photovoice to Examine Physical Activity in the Urban Context and Generate Policy Recommendations: The Heart Healthy Hoods Study. International Journal of Environmental Research and Public Health, 16(5). https://doi.org/10.3390/ijerph16050749

Hajna, S., Ross, N. A., Brazeau, A.-S., Bélisle, P., Joseph, L., & Dasgupta, K. (2015). Associations between neighbourhood walkability and daily steps in adults: A systematic review and meta-analysis. BMC Public Health, 15, 768. https://doi.org/10.1186/s12889-015-2082-x

Hall, C. M., & Ram, Y. (2018). Walk score® and its potential contribution to the study of active transport and walkability: A critical and systematic review. Transportation Research Part D: Transport and Environment, 61, 310–324. https://doi.org/10.1016/j.trd.2017.12.018

Han, B., Cohen, D., & McKenzie, T. L. (2013). Quantifying the contribution of neighborhood parks to physical activity. Preventive Medicine, 57(5), 483–487. https://doi.org/10.1016/j.ypmed.2013.06.021

Han, E., Powell, L., Slater, S., & Quinn, C. (2012). Validation of secondary commercial data sources for physical activity facilities in urban and nonurban settings. Journal of Physical Activity & Health, 9(8), 1080–1088.

Handy, S. L., Boarnet, M. G., Ewing, R., & Killingsworth, R. E. (2002). How the built environment affects physical activity: Views from urban planning. American Journal of Preventive Medicine, 23(2, Supplement 1), 64–73. https://doi.org/10.1016/S0749-3797(02)00475-0

Hastings, K. G., Jose, P. O., Kapphahn, K. I., Frank, A. T. H., Goldstein, B. A., Thompson, C. A., Eggleston, K., Cullen, M. R., & Palaniappan, L. P. (2015). Leading Causes of Death among Asian American Subgroups (2003–2011). PLoS ONE, 10(4). https://doi.org/10.1371/journal.pone.0124341

He, H., Tang, W., Wang, W., & Crits-Christoph, P. (2014). Structural zeroes and zero- inflated models. Shanghai Archives of Psychiatry, 26(4), 236–242. https://doi.org/10.3969/j.issn.1002-0829.2014.04.008

211

Heinrich, K. M., Haddock, C. K., Jitnarin, N., Hughey, J., Berkel, L. A., & Poston, W. S. C. (2017). Perceptions of Important Characteristics of Physical Activity Facilities: Implications for Engagement in Walking, Moderate and Vigorous Physical Activity. Frontiers in Public Health, 5, 319. https://doi.org/10.3389/fpubh.2017.00319

Henley, A., Sealy, D.-A., Hopp, J., & Brown, G. S. (2016). Ethnic Identity and Physical Activity among African, African American, and Black Female College Students. - . International Journal of Health Sciences and Research (IJHSR), 6(2), 286–294.

Hinckson, E., Cerin, E., Mavoa, S., Smith, M., Badland, H., Stewart, T., Duncan, S., & Schofield, G. (2017). Associations of the perceived and objective neighborhood environment with physical activity and sedentary time in New Zealand adolescents. International Journal of Behavioral Nutrition and Physical Activity, 14(1), 145. https://doi.org/10.1186/s12966-017-0597-5

Hirsch, J. A., Moore, K. A., Evenson, K. R., Rodriguez, D. A., & Roux, A. V. D. (2013). Walk Score® and Transit Score® and Walking in the Multi-Ethnic Study of Atherosclerosis. American Journal of Preventive Medicine, 45(2), 158–166. https://doi.org/10.1016/j.amepre.2013.03.018

Hoehner, C. M., Brennan Ramirez, L. K., Elliott, M. B., Handy, S. L., & Brownson, R. C. (2005). Perceived and objective environmental measures and physical activity among urban adults. American Journal of Preventive Medicine, 28(2 Suppl 2), 105–116. https://doi.org/10.1016/j.amepre.2004.10.023

Hoehner, C. M., & Schootman, M. (2010). Concordance of Commercial Data Sources for Neighborhood-Effects Studies. Journal of Urban Health, 87(4), 713–725. https://doi.org/10.1007/s11524-010-9458-0

Hoffman, M. F., Hayes, S., & Napolitano, M. A. (2014). Urban Youths’ Experiences and Perceptions of a Community Cycling Initiative. Urban Studies, 51(2), 300–318. https://doi.org/10.1177/0042098013489741

Holland, A. T., & Palaniappan, L. P. (2012). Problems With the Collection and Interpretation of Asian-American Health Data: Omission, Aggregation, and Extrapolation. Annals of Epidemiology, 22(6), 397–405. https://doi.org/10.1016/j.annepidem.2012.04.001

Huber, P. J. (1967). The behavior of maximum likelihood estimates under nonstandard conditions (In Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability., Vol. 1). University of California Press.

Hughey, S. M., Walsemann, K. M., Child, S., Powers, A., Reed, J. A., & Kaczynski, A. T. (2016). Using an environmental justice approach to examine the relationships between park availability and quality indicators, neighborhood disadvantage, and racial/ethnic composition. Landscape and Urban Planning, 148, 159–169. https://doi.org/10.1016/j.landurbplan.2015.12.016

212

Humpel, N., Owen, N., & Leslie, E. (2002). Environmental factors associated with adults’ participation in physical activity: A review. American Journal of Preventive Medicine, 22(3), 188–199.

Huston, S. L., Evenson, K. R., Bors, P., & Gizlice, Z. (2003). Neighborhood environment, access to places for activity, and leisure-time physical activity in a diverse North Carolina population. American Journal of Health Promotion: AJHP, 18(1), 58–69. https://doi.org/10.4278/0890-1171-18.1.58

Islam, N. S., Khan, S., Kwon, S., Jang, D., Ro, M., & Trinh-Shevrin, C. (2010). Methodological Issues in the Collection, Analysis, and Reporting of Granular Data in Asian American Populations: Historical Challenges and Potential Solutions. Journal of Health Care for the Poor and Underserved, 21(4), 1354– 1381. https://doi.org/10.1353/hpu.2010.0939

Iwamoto, D. K., & Liu, W. M. (2010). The Impact of Racial Identity, Ethnic Identity, Asian Values and Race-Related Stress on Asian Americans and Asian International College Students’ Psychological Well-Being. Journal of Counseling Psychology, 57(1), 79–91. https://doi.org/10.1037/a0017393

Jack, E., & McCormack, G. R. (2014). The associations between objectively-determined and self-reported urban form characteristics and neighborhood-based walking in adults. The International Journal of Behavioral Nutrition and Physical Activity, 11, 71. https://doi.org/10.1186/1479-5868-11-71

James, P., Hart, J. E., Arcaya, M. C., Feskanich, D., Laden, F., & Subramanian, S. V. (2015). Neighborhood Self-Selection: The Role of Pre-Move Health Factors on the Built and Socioeconomic Environment. International Journal of Environmental Research and Public Health, 12(10), 12489–12504. https://doi.org/10.3390/ijerph121012489

Jáuregui, A., Salvo, D., Lamadrid-Figueroa, H., Hernández, B., Rivera, J. A., & Pratt, M. (2017). Perceived neighborhood environmental attributes associated with leisure- time and transport physical activity in Mexican adults. Preventive Medicine, 103S, S21–S26. https://doi.org/10.1016/j.ypmed.2016.11.014

Jih, J., Mukherjea, A., Vittinghoff, E., Nguyen, T. T., Tsoh, J. Y., Fukuoka, Y., Bender, M. S., Tseng, W., & Kanaya, A. M. (2014). Using appropriate body mass index cut points for overweight and obesity among Asian Americans. Preventive Medicine, 65, 1–6. https://doi.org/10.1016/j.ypmed.2014.04.010

John, C. St. (1987). Racial Differences in Neighborhood Evaluation Standards. Urban Affairs Quarterly, 22(3), 377–398. https://doi.org/10.1177/004208168702200302

Jones, S. A., Moore, L. V., Moore, K., Zagorski, M., Brines, S. J., Diez Roux, A. V., & Evenson, K. R. (2015). Disparities in physical activity resource availability in six US regions. Preventive Medicine, 78, 17–22. https://doi.org/10.1016/j.ypmed.2015.05.028

213

Joseph, R. P., & Maddock, J. E. (2016). Observational Park-based physical activity studies: A systematic review of the literature. Preventive Medicine, 89, 257–277. https://doi.org/10.1016/j.ypmed.2016.06.016

Kaczynski, A. T. (2012). Development of a Detailed Log Booklet for Social Ecological Physical Activity Research. Environmental Health Insights, 6, 1–11. https://doi.org/10.4137/EHI.S8086

Kaczynski, A. T., Besenyi, G. M., Stanis, S. A. W., Koohsari, M. J., Oestman, K. B., Bergstrom, R., Potwarka, L. R., & Reis, R. S. (2014). Are park proximity and park features related to park use and park-based physical activity among adults? Variations by multiple socio-demographic characteristics. International Journal of Behavioral Nutrition and Physical Activity, 11(1), 146. https://doi.org/10.1186/s12966-014-0146-4

Kaczynski, A. T., & Havitz, M. E. (2009). Examining the relationship between proximal park features and residents physical activity in neighborhood parks. https://www.scienceopen.com/document?vid=4fabbb6b-fd0f-471e-b2c4- 3bd2218a3da0

Kaczynski, A. T., & Henderson, K. A. (2007). Environmental Correlates of Physical Activity: A Review of Evidence about Parks and Recreation. Leisure Sciences, 29(4), 315–354. https://doi.org/10.1080/01490400701394865

Kaczynski, A. T., Potwarka, L. R., & Saelens, B. E. (2008). Association of park size, distance, and features with physical activity in neighborhood parks. American Journal of Public Health, 98(8), 1451–1456. https://doi.org/10.2105/AJPH.2007.129064

Kaczynski, A. T., Robertson-Wilson, J., & Decloe, M. (2012). Interaction of perceived neighborhood walkability and self-efficacy on physical activity. Journal of Physical Activity & Health, 9(2), 208–217.

Kalra, R., Patel, N., Arora, P., & Arora, G. (2019). Cardiovascular Health and Disease Among Asian-Americans (from the National Health and Nutrition Examination Survey). The American Journal of Cardiology, 124(2), 270–277. https://doi.org/10.1016/j.amjcard.2019.04.026

Kandula, N. R., & Lauderdale, D. S. (2005). Leisure Time, Non-leisure Time, and Occupational Physical Activity in Asian Americans. Annals of Epidemiology, 15(4), 257–265. https://doi.org/10.1016/j.annepidem.2004.06.006

Kao, D., Carvalho Gulati, A., & Lee, R. E. (2016). Physical Activity Among Asian American Adults in Houston, Texas: Data from the Health of Houston Survey 2010. Journal of Immigrant and Minority Health, 18(6), 1470–1481. https://doi.org/10.1007/s10903-015-0274-1

214

Kapteyn, A., Banks, J., Hamer, M., Smith, J. P., Steptoe, A., Soest, A. van, Koster, A., & Wah, S. H. (2018). What they say and what they do: Comparing physical activity across the USA, England and the Netherlands. J Epidemiol Community Health, 72(6), 471–476. https://doi.org/10.1136/jech-2017-209703

Kärmeniemi, M., Lankila, T., Ikäheimo, T., Koivumaa-Honkanen, H., & Korpelainen, R. (2018). The Built Environment as a Determinant of Physical Activity: A Systematic Review of Longitudinal Studies and Natural Experiments. Annals of Behavioral Medicine, 52(3), 239–251. https://doi.org/10.1093/abm/kax043

Kelley, E. A., Kandula, N. R., Kanaya, A. M., & Yen, I. H. (2016). Neighborhood Walkability and Walking for Transport Among South Asians in the MASALA Study. Journal of Physical Activity & Health, 13(5), 514–519. https://doi.org/10.1123/jpah.2015-0266

Kerr, J., Emond, J. A., Badland, H., Reis, R., Sarmiento, O., Carlson, J., Sallis, J. F., Cerin, E., Cain, K., Conway, T., Schofield, G., Macfarlane, D. J., Christiansen, L. B., Van Dyck, D., Davey, R., Aguinaga-Ontoso, I., Salvo, D., Sugiyama, T., Owen, N., … Natarajan, L. (2016). Perceived Neighborhood Environmental Attributes Associated with Walking and Cycling for Transport among Adult Residents of 17 Cities in 12 Countries: The IPEN Study. Environmental Health Perspectives, 124(3), 290–298. https://doi.org/10.1289/ehp.1409466

King, K. E., & Clarke, P. J. (2015). A Disadvantaged Advantage in Walkability: Findings From Socioeconomic and Geographical Analysis of National Built Environment Data in the United States. American Journal of Epidemiology, 181(1), 17–25. https://doi.org/10.1093/aje/kwu310

Koehler, K., Latshaw, M., Matte, T., Kass, D., Frumkin, H., Fox, M., Hobbs, B. F., Wills-Karp, M., & Burke, T. A. (2018). Building Healthy Community Environments: A Public Health Approach. Public Health Reports, 133(1_suppl), 35S-43S. https://doi.org/10.1177/0033354918798809

Koohsari, M. J., Badland, H., & Giles-Corti, B. (2013). (Re)Designing the built environment to support physical activity: Bringing public health back into urban design and planning. Cities, 35, 294–298. https://doi.org/10.1016/j.cities.2013.07.001

Koohsari, M. J., Badland, H., Sugiyama, T., Mavoa, S., Christian, H., & Giles-Corti, B. (2015). Mismatch between perceived and objectively measured land use mix and street connectivity: Associations with neighborhood walking. Journal of Urban Health: Bulletin of the New York Academy of Medicine, 92(2), 242–252. https://doi.org/10.1007/s11524-014-9928-x

Laidley, T. (2016). Measuring Sprawl: A New Index, Recent Trends, and Future Research. Urban Affairs Review, 52(1), 66–97. https://doi.org/10.1177/1078087414568812

215

Lambert, D. (1992). Zero-Inflated Poisson Regression, with an Application to Defects in Manufacturing. Technometrics, 34(1), 1–14. JSTOR. https://doi.org/10.2307/1269547

Larson, L. R., Jennings, V., & Cloutier, S. A. (2016). Public Parks and Wellbeing in Urban Areas of the United States. PLOS ONE, 11(4), e0153211. https://doi.org/10.1371/journal.pone.0153211

Lavie, C. J., Ozemek, C., Carbone, S., Katzmarzyk, P. T., & Blair, S. N. (2019). Sedentary Behavior, Exercise, and Cardiovascular Health. Circulation Research, 124(5), 799–815. https://doi.org/10.1161/CIRCRESAHA.118.312669

Lee, C. D., Folsom, A. R., & Blair, S. N. (2003). Physical activity and stroke risk: A meta-analysis. Stroke, 34(10), 2475–2481. https://doi.org/10.1161/01.STR.0000091843.02517.9D

Lee, I., Ewing, R., & Sesso, H. D. (2009). Relationship between the built environment and physical activity levels: The Harvard Alumni Health Study. American Journal of Preventive Medicine, 37(4), 293–298. https://doi.org/10.1016/j.amepre.2009.06.007

Lee, J. W., Brancati, F. L., & Yeh, H.-C. (2011). Trends in the Prevalence of Type 2 Diabetes in Asians Versus Whites: Results from the United States National Health Interview Survey, 1997–2008. Diabetes Care, 34(2), 353–357. https://doi.org/10.2337/dc10-0746

Lee, K. J., & Stodolska, M. (2017). Asian North Americans’ Leisure: A Critical Examination of the Theoretical Frameworks Used in Research and Suggestions for Future Study. Leisure Sciences, 39(6), 524–542. https://doi.org/10.1080/01490400.2016.1215944

Lee, R. E., Booth, K. M., Reese-Smith, J. Y., Regan, G., & Howard, H. H. (2005). The Physical Activity Resource Assessment (PARA) instrument: Evaluating features, amenities and incivilities of physical activity resources in urban neighborhoods. The International Journal of Behavioral Nutrition and Physical Activity, 2, 13. https://doi.org/10.1186/1479-5868-2-13

Lee, R. E., Mama, S. K., Adamus-Leach, H. J., & Soltero, E. G. (2015). Contribution of Neighborhood Income and Access to Quality Physical Activity Resources to Physical Activity in Ethnic Minority Women over Time. American Journal of Health Promotion, 29(4), 210–216. https://doi.org/10.4278/ajhp.130403-QUAN- 148

Leggett, S. S., Wesley, J. M., Myla, A., McCoy, P., & Betson, N. (2018). Availability and Quality of Physical Activity Resources in Neighborhood Parks for Pregnant Women and Women of Childbearing Age. Preventing Chronic Disease, 15. https://doi.org/10.5888/pcd15.170411

216

Leslie, E., Coffee, N., Frank, L., Owen, N., Bauman, A., & Hugo, G. (2007). Walkability of local communities: Using geographic information systems to objectively assess relevant environmental attributes. Health & Place, 13(1), 111–122. https://doi.org/10.1016/j.healthplace.2005.11.001

Li, F., Fisher, K., Brownson, R., & Bosworth, M. (2005). Multilevel modelling of built environment characteristics related to neighbourhood walking activity in older adults. Journal of Epidemiology and Community Health, 59(7), 558–564. https://doi.org/10.1136/jech.2004.028399

Li, K., & Wen, M. (2013). Racial and Ethnic Disparities in Leisure-time Physical Activity in California: Patterns and Mechanisms. Race and Social Problems, 5(3), 147–156. https://doi.org/10.1007/s12552-013-9087-9

Li, Y., Kao, D., & Dinh, T. Q. (2015). Correlates of neighborhood environment with walking among older Asian Americans. Journal of Aging and Health, 27(1), 17– 34. https://doi.org/10.1177/0898264314535636

Liem, R., Lim, B. A., & Liem, J. H. (2000). Acculturation and emotion among Asian Americans. Cultural Diversity & Ethnic Minority Psychology, 6(1), 13–31.

Liu, X., Chen, Y., Boucher, N. L., & Rothberg, A. E. (2017). Prevalence and change of central obesity among US Asian adults: NHANES 2011–2014. BMC Public Health, 17. https://doi.org/10.1186/s12889-017-4689-6

López, G., Ruiz, N. G., & Patten, E. (2017, September 8). Key facts about Asian Americans, a diverse and growing population. Pew Research Center. http://www.pewresearch.org/fact-tank/2017/09/08/key-facts-about-asian- americans/

Lopez, G., Ruiz, N., & Patten, E. (2017, September 8). Key facts about Asian Americans, a diverse and growing population. Pew Research Center. http://www.pewresearch.org/fact-tank/2017/09/08/key-facts-about-asian- americans/

Lovasi, G. S., Neckerman, K. M., Quinn, J. W., Weiss, C. C., & Rundle, A. (2009). Effect of Individual or Neighborhood Disadvantage on the Association Between Neighborhood Walkability and Body Mass Index. American Journal of Public Health, 99(2), 279–284. https://doi.org/10.2105/AJPH.2008.138230

Ma, G. X., Zhu, L., Shive, S. E., Zhang, G., Senter, Y. R., Topete, P., Seals, B., Zhai, S., Wang, M., & Tan, Y. (2019). The Evaluation of IDEAL-REACH Program to Improve Nutrition among Asian American Community Members in the Philadelphia Metropolitan Area. International Journal of Environmental Research and Public Health, 16(17). https://doi.org/10.3390/ijerph16173054

217

Ma, L., & Dill, J. (2015). Associations between the objective and perceived built environment and bicycling for transportation. Journal of Transport & Health, 2(2), 248–255. https://doi.org/10.1016/j.jth.2015.03.002

Maddison, R., Hoorn, S. V., Jiang, Y., Mhurchu, C. N., Exeter, D., Dorey, E., Bullen, C., Utter, J., Schaaf, D., & Turley, M. (2009). The environment and physical activity: The influence of psychosocial, perceived and built environmental factors. The International Journal of Behavioral Nutrition and Physical Activity, 6, 19. https://doi.org/10.1186/1479-5868-6-19

Marmot, M., Friel, S., Bell, R., Houweling, T. A. J., Taylor, S., & Commission on Social Determinants of Health. (2008). Closing the gap in a generation: Health equity through action on the social determinants of health. Lancet (London, England), 372(9650), 1661–1669. https://doi.org/10.1016/S0140-6736(08)61690-6

Martinez, S. M., Ayala, G. X., Arredondo, E. M., Finch, B., & Elder, J. (2008). Active transportation and acculturation among Latino children in San Diego County. Preventive Medicine, 47(3), 313–318. https://doi.org/10.1016/j.ypmed.2008.01.018

Matthews, C. E., Berrigan, D., Fischer, B., Gomersall, S. R., Hillreiner, A., Kim, Y., Leitzmann, M. F., Saint-Maurice, P., Olds, T. S., & Welk, G. J. (2019). Use of previous-day recalls of physical activity and sedentary behavior in epidemiologic studies: Results from four instruments. BMC Public Health, 19(2), 478. https://doi.org/10.1186/s12889-019-6763-8

Maxwell, A. E., Crespi, C. M., Alano, R. E., Sudan, M., & Bastani, R. (2012). Health Risk Behaviors Among Five Asian American Subgroups in California: Identifying Intervention Priorities. Journal of Immigrant and Minority Health, 14(5), 890–894. https://doi.org/10.1007/s10903-011-9552-8

McCormack, G. R., & Shiell, A. (2011). In search of causality: A systematic review of the relationship between the built environment and physical activity among adults. The International Journal of Behavioral Nutrition and Physical Activity, 8, 125. https://doi.org/10.1186/1479-5868-8-125

McDonald, J. T., & Kennedy, S. (2004). Insights into the “healthy immigrant effect”: Health status and health service use of immigrants to Canada. Social Science & Medicine (1982), 59(8), 1613–1627. https://doi.org/10.1016/j.socscimed.2004.02.004

McGinn, A. P., Evenson, K. R., Herring, A. H., Huston, S. L., & Rodriguez, D. A. (2007). Exploring associations between physical activity and perceived and objective measures of the built environment. Journal of Urban Health: Bulletin of the New York Academy of Medicine, 84(2), 162–184. https://doi.org/10.1007/s11524-006-9136-4

218

McKenzie, B. (2015). Who Drives to Work? Commuting by Automobile in the United States: 2013. 28.

McKenzie, T. L., Cohen, D. A., Sehgal, A., Williamson, S., & Golinelli, D. (2006). System for Observing Play and Recreation in Communities (SOPARC): Reliability and Feasibility Measures. Journal of Physical Activity and Health, 3(s1), S208–S222. https://doi.org/10.1123/jpah.3.s1.s208

McLeroy, K. R., Bibeau, D., Steckler, A., & Glanz, K. (1988). An ecological perspective on health promotion programs. Health Education Quarterly, 15(4), 351–377.

McMillan, T. E., Cubbin, C., Parmenter, B., Medina, A. V., & Lee, R. E. (2010). Neighborhood sampling: How many streets must an auditor walk? International Journal of Behavioral Nutrition and Physical Activity, 7(1), 20. https://doi.org/10.1186/1479-5868-7-20

Mecredy, G., Pickett, W., & Janssen, I. (2011). Street Connectivity is Negatively Associated with Physical Activity in Canadian Youth. International Journal of Environmental Research and Public Health, 8(8), 3333–3350. https://doi.org/10.3390/ijerph8083333

Millstein, R. A., Cain, K. L., Sallis, J. F., Conway, T. L., Geremia, C., Frank, L. D., Chapman, J., Van Dyck, D., Dipzinski, L. R., Kerr, J., Glanz, K., & Saelens, B. E. (2013). Development, scoring, and reliability of the Microscale Audit of Pedestrian Streetscapes (MAPS). BMC Public Health, 13(1), 403. https://doi.org/10.1186/1471-2458-13-403

Mohai, P., & Bryant, B. (1991). Race, Poverty & the Distribution of Environmental Hazards: Reviewing the Evidence. Race, Poverty & the Environment, 2(3/4), 3– 27. JSTOR.

Moise, R. K., Meca, A., Schwartz, S. J., Unger, J. B., Lorenzo-Blanco, E. I., Ángel Cano, M., Szapocznik, J., Piña-Watson, B., Des Rosiers, S. E., Baezconde-Garbanati, L., Soto, D. W., Pattarroyo, M., Villamar, J. A., & Lizzi, K. M. (2018). The use of cultural identity in predicting health lifestyle behaviors in Latinx immigrant adolescents. Cultural Diversity and Ethnic Minority Psychology. https://doi.org/10.1037/cdp0000229

Moody, J. S., Prochaska, J. J., Sallis, J. F., McKenzie, T. L., Brown, M., & Conway, T. L. (2004). Viability of parks and recreation centers as sites for youth physical activity promotion. Health Promotion Practice, 5(4), 438–443. https://doi.org/10.1177/1524839903258222

Moore, L. V., Diez Roux, A. V., Evenson, K. R., McGinn, A. P., & Brines, S. J. (2008). Availability of Recreational Resources in Minority and Low Socioeconomic Status Areas. American Journal of Preventive Medicine, 34(1), 16–22. https://doi.org/10.1016/j.amepre.2007.09.021

219

Moore-Harrison, T., & Lightfoot, J. T. (2010). Driven to Be Inactive?—The Genetics of Physical Activity. Progress in Molecular Biology and Translational Science, 94, 271–290. https://doi.org/10.1016/B978-0-12-375003-7.00010-8

Murayama, H., Spencer, M. S., Sinco, B. R., Palmisano, G., & Kieffer, E. C. (2017). Does Racial/Ethnic Identity Influence the Effectiveness of a Community Health Worker Intervention for African American and Latino Adults With Type 2 Diabetes? Health Education & Behavior, 44(3), 485–493. https://doi.org/10.1177/1090198116673821

Nagel, C. L., Carlson, N. E., Bosworth, M., & Michael, Y. L. (2008). The Relation between Neighborhood Built Environment and Walking Activity among Older Adults. American Journal of Epidemiology, 168(4), 461–468. https://doi.org/10.1093/aje/kwn158

National Recreation and Park Association. (2020). Active Transportation and Parks and Recreation. https://www.nrpa.org/contentassets/f768428a39aa4035ae55b2aaff372617/activetr ansportation_final.high.pdf

Nichani, V., Vena, J. E., Friedenreich, C. M., Christie, C., & McCormack, G. R. (2019). A population-based study of the associations between neighbourhood walkability and different types of physical activity in Canadian men and women. Preventive Medicine, 129, 105864. https://doi.org/10.1016/j.ypmed.2019.105864

Novotny, R., Williams, A. E., Vinoya, A. C., Oshiro, C. E. S., & Vogt, T. M. (2009). US acculturation, food intake, and obesity among Asian-Pacific hotel workers. Journal of the American Dietetic Association, 109(10), 1712–1718. https://doi.org/10.1016/j.jada.2009.07.013

O’Loughlin, J., Maximova, K., Tan, Y., & Gray-Donald, K. (2007). Lifestyle risk factors for chronic disease across family origin among adults in multiethnic, low-income, urban neighborhoods. Ethnicity & Disease, 17(4), 657–663.

Oster, A., & Yung, J. (2010). Dietary Acculturation, Obesity, and Diabetes Among Chinese Immigrants in New York City. Diabetes Care, 33(8), e109–e109. https://doi.org/10.2337/dc09-2291

Osypuk, T., Alonso, A., & L. (2015). Understanding the Healthy Immigrant Effect and Cardiovascular Disease. Circulation, 132(16), 1522–1524. https://doi.org/10.1161/CIRCULATIONAHA.115.018707

Osypuk, T. L., Roux, A. V. D., Hadley, C., & Kandula, N. (2009). Are Immigrant Enclaves Healthy Places to Live? The Multi-ethnic Study of Atherosclerosis. Social Science & Medicine (1982), 69(1), 110–120. https://doi.org/10.1016/j.socscimed.2009.04.010

220

Owen, N., Humpel, N., Leslie, E., Bauman, A., & Sallis, J. F. (2004). Understanding environmental influences on walking; Review and research agenda. American Journal of Preventive Medicine, 27(1), 67–76. https://doi.org/10.1016/j.amepre.2004.03.006

Oyserman, D., Fryberg, S. A., & Yoder, N. (2007). Identity-based motivation and health. Journal of Personality and Social Psychology, 93(6), 1011–1027. https://doi.org/10.1037/0022-3514.93.6.1011

Palaniappan, L. P., Araneta, M. R. G., Assimes, T. L., Barrett-Connor, E. L., Carnethon, M. R., Criqui, M. H., Fung, G. L., Venkat Narayan, K. M., Patel, H., Taylor- Piliae, R. E., Wilson, P. W. F., & Wong, N. D. (2010). Call to Action: Cardiovascular Disease in Asian Americans A Science Advisory From the American Heart Association. Circulation, 122(12), 1242–1252. https://doi.org/10.1161/CIR.0b013e3181f22af4

Paquet, C., Daniel, M., Kestens, Y., Léger, K., & Gauvin, L. (2008). Field validation of listings of food stores and commercial physical activity establishments from secondary data. The International Journal of Behavioral Nutrition and Physical Activity, 5, 58. https://doi.org/10.1186/1479-5868-5-58

Parker, N. H., O’Connor, D. P., Kao, D. T., & Lee, R. E. (2016). Do Neighborhood Physical Activity Resources and Land Use Influence Physical Activity among African American Public Housing Residents? Journal of Health Care for the Poor and Underserved, 27(3), 1330–1344. https://doi.org/10.1353/hpu.2016.0135

PASDA. (2015). Pennsylvania Local Parks. https://www.pasda.psu.edu/uci/DataSummary.aspx?dataset=307

Patel, S., Kwon, S., Arista, P., Tepporn, E., Chung, M., Ko Chin, K., Rideout, C., Islam, N., & Trinh-Shevrin, C. (2015). Using evidence-based policy, systems, and environmental strategies to increase access to healthy food and opportunities for physical activity among Asian Americans, Native Hawaiians, and Pacific Islanders. American Journal of Public Health, 105 Suppl 3, S455-458. https://doi.org/10.2105/AJPH.2015.302637

Patterson, F., Zhang, G., Davey, A., Tan, Y., & Ma, G. X. (2016). American Heart Association’s Ideal Cardiovascular Health Metrics in Under-Represented Asian Americans. Journal of Community Health, 41(6), 1282–1289. https://doi.org/10.1007/s10900-016-0217-3

Peachey, A. A., & Baller, S. L. (2015). Perceived Built Environment Characteristics of On-Campus and Off-Campus Neighborhoods Associated With Physical Activity of College Students. Journal of American College Health: J of ACH, 63(5), 337– 342. https://doi.org/10.1080/07448481.2015.1015027

221

Phinney, J. S. (1992). The Multigroup Ethnic Identity Measure: A New Scale for Use with Diverse Groups. Journal of Adolescent Research, 7(2), 156–176. https://doi.org/10.1177/074355489272003

Phinney, J. S., & Ong, A. D. (2007). Conceptualization and measurement of ethnic identity: Current status and future directions. Journal of Counseling Psychology, 54(3), 271–281. https://doi.org/10.1037/0022-0167.54.3.271

Pietiläinen, K. H., Kaprio, J., Borg, P., Plasqui, G., Yki-Järvinen, H., Kujala, U. M., Rose, R. J., Westerterp, K. R., & Rissanen, A. (2008). Physical inactivity and obesity: A vicious circle. Obesity (Silver Spring, Md.), 16(2), 409–414. https://doi.org/10.1038/oby.2007.72

Pratt, M., Macera, C. A., Sallis, J. F., O’Donnell, M., & Frank, L. D. (2004). Economic interventions to promote physical activity: Application of the SLOTH model. American Journal of Preventive Medicine, 27(3, Supplement), 136–145. https://doi.org/10.1016/j.amepre.2004.06.015

Rees-Punia, E., Hathaway, E. D., & Gay, J. L. (2018). Crime, perceived safety, and physical activity: A meta-analysis. Preventive Medicine, 111, 307–313. https://doi.org/10.1016/j.ypmed.2017.11.017

Rhodes, R. E., Courneya, K. S., Blanchard, C. M., & Plotnikoff, R. C. (2007). Prediction of leisure-time walking: An integration of social cognitive, perceived environmental, and personality factors. International Journal of Behavioral Nutrition and Physical Activity, 4(1), 51. https://doi.org/10.1186/1479-5868-4-51

Rhodes, R. E., Saelens, B. E., & Sauvage-Mar, C. (2018). Understanding Physical Activity through Interactions Between the Built Environment and Social Cognition: A Systematic Review. Sports Medicine, 48(8), 1893–1912. https://doi.org/10.1007/s40279-018-0934-0

Romeike, K., Lechner, L., Vries, H. de, & Oenema, A. (2016). Development of a computer-tailored nutrition and physical activity intervention for lower-educated women of Dutch, Turkish and Moroccan origin using content matching and ethnic identity tailoring. BMC Public Health, 16(1), 1–16. https://doi.org/10.1186/s12889-016-3596-6

Saelens, B. E., & Handy, S. L. (2008). Built environment correlates of walking: A review. Medicine and Science in Sports and Exercise, 40(7 Suppl), S550-566. https://doi.org/10.1249/MSS.0b013e31817c67a4

Saelens, B. E., Sallis, J. F., Black, J. B., & Chen, D. (2003). Neighborhood-Based Differences in Physical Activity: An Environment Scale Evaluation. American Journal of Public Health, 93(9), 1552–1558.

222

Saelens, B. E., Sallis, J. F., & Frank, L. D. (2003). Environmental correlates of walking and cycling: Findings from the transportation, urban design, and planning literatures. Annals of Behavioral Medicine: A Publication of the Society of Behavioral Medicine, 25(2), 80–91. https://doi.org/10.1207/S15324796ABM2502_03

Sallis, J. F., Floyd, M. F., Rodríguez, D. A., & Saelens, B. E. (2012). The Role of Built Environments in Physical Activity, Obesity, and CVD. Circulation, 125(5), 729– 737. https://doi.org/10.1161/CIRCULATIONAHA.110.969022

Sallis, J. F., Hovell, M. F., Hofstetter, C. R., Elder, J. P., Hackley, M., Caspersen, C. J., & Powell, K. E. (1990). Distance between homes and exercise facilities related to frequency of exercise among San Diego residents. Public Health Reports (Washington, D.C.: 1974), 105(2), 179–185.

Sallis, J. F., Owen, N., & Fisher, E. B. (2008). Ecological models of health behavior. In Health behavior and health education: Theory, research, and practice, 4th ed (pp. 465–485). Jossey-Bass.

Sallis, J. F., Saelens, B. E., Frank, L. D., Conway, T. L., Slymen, D. J., Cain, K. L., Chapman, J. E., & Kerr, J. (2009). Neighborhood Built Environment and Income: Examining Multiple Health Outcomes. Social Science & Medicine (1982), 68(7), 1285–1293. https://doi.org/10.1016/j.socscimed.2009.01.017

Sallis, J. F., Slymen, D. J., Conway, T. L., Frank, L. D., Saelens, B. E., Cain, K., & Chapman, J. E. (2011). Income disparities in perceived neighborhood built and social environment attributes. Health & Place, 17(6), 1274–1283. https://doi.org/10.1016/j.healthplace.2011.02.006

Salvo, G., Lashewicz, B. M., Doyle-Baker, P. K., & McCormack, G. R. (2018). Neighbourhood Built Environment Influences on Physical Activity among Adults: A Systematized Review of Qualitative Evidence. International Journal of Environmental Research and Public Health, 15(5). https://doi.org/10.3390/ijerph15050897

Schipperijn, J., Bentsen, P., Troelsen, J., Toftager, M., & Stigsdotter, U. K. (2013). Associations between physical activity and characteristics of urban green space. Urban Forestry & Urban Greening, 12(1), 109–116. https://doi.org/10.1016/j.ufug.2012.12.002

Schipperijn, J., Cerin, E., Adams, M. A., Reis, R., Smith, G., Cain, K., Christiansen, L. B., Dyck, D. van, Gidlow, C., Frank, L. D., Mitáš, J., Pratt, M., Salvo, D., Schofield, G., & Sallis, J. F. (2017). Access to parks and physical activity: An eight country comparison. Urban Forestry & Urban Greening, 27, 253–263. https://doi.org/10.1016/j.ufug.2017.08.010

223

Serafica, R. C., Lane, S. H., & Ceria-Ulep, C. D. (2013). Dietary Acculturation and Predictors of Anthropometric Indicators Among Filipino Americans. SAGE Open, 3(3), 2158244013495543. https://doi.org/10.1177/2158244013495543

Smalley, K. B., Warren, J. C., McClendon, S., Peacock, W., & Caro, M. (2016). Ethnic Identity Attachment and Motivation for Weight Loss and Exercise Among Rural, Overweight, African-American Women. Clinical Medicine Insights. Women’s Health, 9(Suppl 1), 95–101. https://doi.org/10.4137/CMWH.S34691

Smith, M., Hosking, J., Woodward, A., Witten, K., MacMillan, A., Field, A., Baas, P., & Mackie, H. (2017). Systematic literature review of built environment effects on physical activity and active transport – an update and new findings on health equity. The International Journal of Behavioral Nutrition and Physical Activity, 14. https://doi.org/10.1186/s12966-017-0613-9

Sohail, Q. Z., Chu, A., Rezai, M. R., Donovan, L. R., Ko, D. T., & Tu, J. V. (2015). The Risk of Ischemic Heart Disease and Stroke Among Immigrant Populations: A Systematic Review. The Canadian Journal of Cardiology, 31(9), 1160–1168. https://doi.org/10.1016/j.cjca.2015.04.027

Srinivasan, S., & Guillermo, T. (2000). Toward improved health: Disaggregating Asian American and Native Hawaiian/Pacific Islander data. American Journal of Public Health, 90(11), 1731–1734. https://doi.org/10.2105/ajph.90.11.1731

Staimez, L. R., Weber, M. B., Narayan, K. M. V., & Oza-Frank, R. (2013). A systematic review of overweight, obesity, and type 2 diabetes among Asian American subgroups. Current Diabetes Reviews, 9(4), 312–331.

Stasi, S., Spengler, J., Maddock, J., McKyer, L., & Clark, H. (2019). Increasing Access to Physical Activity Within Low Income and Diverse Communities: A Systematic Review. American Journal of Health Promotion: AJHP, 890117119832257. https://doi.org/10.1177/0890117119832257

StataCorp LLC. (2020). STATA Manual- ZINB. https://www.stata.com/manuals13/rzinb.pdf

Stewart, O. T., Moudon, A. V., Littman, A. J., Seto, E., & Saelens, B. E. (2018). Why neighborhood park proximity is not associated with total physical activity. Health & Place, 52, 163–169. https://doi.org/10.1016/j.healthplace.2018.05.011

Stodolska, M., Shinew, K. J., Acevedo, J. C., & Izenstark, D. (2011). Perceptions of Urban Parks as Havens and Contested Terrains by Mexican-Americans in Chicago Neighborhoods. Leisure Sciences, 33(2), 103–126. https://doi.org/10.1080/01490400.2011.550220

Stokols, D. (1992). Establishing and maintaining healthy environments. Toward a social ecology of health promotion. The American Psychologist, 47(1), 6–22.

224

Stubbe, J. H., Boomsma, D. I., Vink, J. M., Cornes, B. K., Martin, N. G., Skytthe, A., Kyvik, K. O., Rose, R. J., Kujala, U. M., Kaprio, J., Harris, J. R., Pedersen, N. L., Hunkin, J., Spector, T. D., & Geus, E. J. C. de. (2006). Genetic Influences on Exercise Participation in 37.051 Twin Pairs from Seven Countries. PLOS ONE, 1(1), e22. https://doi.org/10.1371/journal.pone.0000022

Sugiyama, T., Cerin, E., Owen, N., Oyeyemi, A. L., Conway, T. L., Van Dyck, D., Schipperijn, J., Macfarlane, D. J., Salvo, D., Reis, R. S., Mitáš, J., Sarmiento, O. L., Davey, R., Schofield, G., Orzanco-Garralda, R., & Sallis, J. F. (2014). ׳Perceived neighbourhood environmental attributes associated with adults recreational walking: IPEN Adult study in 12 countries. Health & Place, 28, 22– 30. https://doi.org/10.1016/j.healthplace.2014.03.003

Sugiyama, T., Leslie, E., Giles-Corti, B., & Owen, N. (2009). Physical activity for recreation or exercise on neighbourhood streets: Associations with perceived environmental attributes. Health & Place, 15(4), 1058–1063. https://doi.org/10.1016/j.healthplace.2009.05.001

Sugiyama, T., Neuhaus, M., Cole, R., Giles-Corti, B., & Owen, N. (2012). Destination and route attributes associated with adults’ walking: A review. Medicine and Science in Sports and Exercise, 44(7), 1275–1286. https://doi.org/10.1249/MSS.0b013e318247d286

Suminski, R. R., Connolly, E. K., May, L. E., Wasserman, J., Olvera, N., & Lee, R. (2012). Park quality in racial/ethnic minority neighborhoods. Environmental Justice, 5(6), 271–278. https://doi.org/10.1089/env.2012.0013

Sylvia, L. G., Bernstein, E. E., Hubbard, J. L., Keating, L., & Anderson, E. J. (2014). A Practical Guide to Measuring Physical Activity. Journal of the Academy of Nutrition and Dietetics, 114(2), 199–208. https://doi.org/10.1016/j.jand.2013.09.018

Taylor, C. B., Sallis, J. F., & Needle, R. (1985). The relation of physical activity and exercise to mental health. Public Health Reports, 100(2), 195–202.

Taylor, W. C., Poston, W. S. C., Jones, L., & Kraft, M. K. (2006). Environmental Justice: Obesity, Physical Activity, and Healthy Eating. Journal of Physical Activity & Health, 3(s1), S30–S54. https://doi.org/10.1123/jpah.3.s1.s30

The Pew Charitable Trusts. (2011). A City Transformed: The Racial and Ethnic Changes in Philadelphia Over the Last 20 Years.

The Pew Charitable Trusts. (2018). Philadelphia’s Immigrants: Who they are and how they are changing the city. https://www.pewtrusts.org/- /media/assets/2018/06/pri_philadelphias_immigrants.pdf

The Pew Charitable Trusts. (2019). Philadelphia 2019: The State of the City. https://www.pewtrusts.org/-/media/assets/2019/10/sotc_2019.pdf

225

The White House. (2009). Initiative on Asian Americans and Pacific Islanders. The White House. https://obamawhitehouse.archives.gov/node/14486

Thielman, J., Rosella, L., Copes, R., Lebenbaum, M., & Manson, H. (2015). Neighborhood walkability: Differential associations with self-reported transport walking and leisure-time physical activity in Canadian towns and cities of all sizes. Preventive Medicine, 77, 174–180. https://doi.org/10.1016/j.ypmed.2015.05.011

Thornton, C. M., Conway, T. L., Cain, K. L., Gavand, K. A., Saelens, B. E., Frank, L. D., Geremia, C. M., Glanz, K., King, A. C., & Sallis, J. F. (2016). Disparities in Pedestrian Streetscape Environments by Income and Race/Ethnicity. SSM - Population Health, 2, 206–216. https://doi.org/10.1016/j.ssmph.2016.03.004

Thornton, L. E., Pearce, J. R., & Kavanagh, A. M. (2011). Using Geographic Information Systems (GIS) to assess the role of the built environment in influencing obesity: A glossary. The International Journal of Behavioral Nutrition and Physical Activity, 8, 71. https://doi.org/10.1186/1479-5868-8-71

Towne, S. D., Won, J., Lee, S., Ory, M. G., Forjuoh, S. N., Wang, S., & Lee, C. (2016). Using Walk ScoreTM and Neighborhood Perceptions to Assess Walking Among Middle-Aged and Older Adults. Journal of Community Health, 41(5), 977–988. https://doi.org/10.1007/s10900-016-0180-z

Trinh-Shevrin, C., Islam, N. S., Rey, M. J., & Trinh-Shevrin, C. (2009). Asian American Communities and Health: Context, Research, Policy, and Action. John Wiley & Sons, Incorporated. http://ebookcentral.proquest.com/lib/templeuniv- ebooks/detail.action?docID=427569

Tuckel, P., & Milczarski, W. (2015). Walk Score(TM), Perceived Neighborhood Walkability, and walking in the US. American Journal of Health Behavior, 39(2), 242–256. https://doi.org/10.5993/AJHB.39.2.11

Twardzik, E., Judd, S., Bennett, A., Hooker, S., Howard, V., Hutto, B., Clarke, P., & Colabianchi, N. (2019). Walk Score and objectively measured physical activity within a national cohort. J Epidemiol Community Health, 73(6), 549–556. https://doi.org/10.1136/jech-2017-210245

United States Census Bureau. (2015). Annual Estimates of the Resident Population: April 1, 2010 to July 1, 2014—2014 Population Estimates.

U.S. Department of Health and Human Services. (2014). Social Determinants of Health. https://www.healthypeople.gov/2020/topics-objectives/topic/social-determinants- of-health

226

U.S. Department of Health and Human Services. (2018). Physical Activity Guidelines for Americans, 2nd Edition. file:///Users/aishabhimla/Desktop/PhD%20Dissertation%20Files%202018- 2019/Articles/Physical_Activity_Guidelines_2nd_edition.pdf

U.S. Department of Health and Human Services, Office of Disease Prevention and Health Promotion. (2014). Healthy People 2020: Physical Activity. https://www.healthypeople.gov/2020/topics-objectives/topic/physical-activity

Van Cauwenberg, J., Cerin, E., Timperio, A., Salmon, J., Deforche, B., & Veitch, J. (2017). Is the Association between Park Proximity and Recreational Physical Activity among Mid-Older Aged Adults Moderated by Park Quality and Neighborhood Conditions? International Journal of Environmental Research and Public Health, 14(2). https://doi.org/10.3390/ijerph14020192

Van Cauwenberg, J., Nathan, A., Barnett, A., Barnett, D. W., Cerin, E., & Council on Environment and Physical Activity (CEPA)-Older Adults Working Group. (2018). Relationships Between Neighbourhood Physical Environmental Attributes and Older Adults’ Leisure-Time Physical Activity: A Systematic Review and Meta-Analysis. Sports Medicine (Auckland, N.Z.), 48(7), 1635–1660. https://doi.org/10.1007/s40279-018-0917-1

Van Dyck, D., Cerin, E., Conway, T. L., De Bourdeaudhuij, I., Owen, N., Kerr, J., Cardon, G., Frank, L. D., Saelens, B. E., & Sallis, J. F. (2012). Associations between perceived neighborhood environmental attributes and adults’ sedentary behavior: Findings from the USA, Australia and Belgium. Social Science & Medicine, 74(9), 1375–1384. https://doi.org/10.1016/j.socscimed.2012.01.018

Vargo, J., Stone, B., & Glanz, K. (2012). Google walkability: A new tool for local planning and public health research? Journal of Physical Activity & Health, 9(5), 689–697. https://doi.org/10.1123/jpah.9.5.689

Vaughan, K. B., Kaczynski, A. T., Wilhelm Stanis, S. A., Besenyi, G. M., Bergstrom, R., & Heinrich, K. M. (2013). Exploring the Distribution of Park Availability, Features, and Quality Across Kansas City, Missouri by Income and Race/Ethnicity: An Environmental Justice Investigation. Annals of Behavioral Medicine, 45(1), 28–38. https://doi.org/10.1007/s12160-012-9425-y

Walk Score. (2019a). Walk Score and API. https://www.walkscore.com/professional/api.php

Walk Score. (2019b). Walk Score Methodology. Walk Score. https://www.walkscore.com/methodology.shtml

Wang, L., & Lo, L. (2007). Immigrant Grocery-Shopping Behavior: Ethnic Identity versus Accessibility: Environment and Planning A. https://doi.org/10.1068/a3833

227

Wang, X., Conway, T. L., Cain, K. L., Frank, L. D., Saelens, B. E., Geremia, C., Kerr, J., Glanz, K., Carlson, J. A., & Sallis, J. F. (2017). Interactions of psychosocial factors with built environments in explaining adolescents’ active transportation. Preventive Medicine, 100, 76–83. https://doi.org/10.1016/j.ypmed.2017.04.008

Wanner, M., Götschi, T., Martin-Diener, E., Kahlmeier, S., & Martin, B. W. (2012). Active Transport, Physical Activity, and Body Weight in Adults: A Systematic Review. American Journal of Preventive Medicine, 42(5), 493–502. https://doi.org/10.1016/j.amepre.2012.01.030

Warburton, D. E. R., & Bredin, S. S. D. (2017). Health benefits of physical activity: A systematic review of current systematic reviews. Current Opinion in Cardiology, 32(5), 541–556. https://doi.org/10.1097/HCO.0000000000000437

Weiss, C. C., Purciel, M., Bader, M., Quinn, J. W., Lovasi, G., Neckerman, K. M., & Rundle, A. G. (2011). Reconsidering Access: Park Facilities and Neighborhood Disamenities in New York City. Journal of Urban Health : Bulletin of the New York Academy of Medicine, 88(2), 297–310. https://doi.org/10.1007/s11524-011- 9551-z

Wells, N. M., & Yang, Y. (2008). Neighborhood Design and Walking: A Quasi- Experimental Longitudinal Study. American Journal of Preventive Medicine, 34(4), 313–319. https://doi.org/10.1016/j.amepre.2008.01.019

Wen, M., Kandula, N. R., & Lauderdale, D. S. (2007). Walking for Transportation or Leisure: What Difference Does the Neighborhood Make? Journal of General Internal Medicine, 22(12), 1674–1680. https://doi.org/10.1007/s11606-007-0400- 4

Wen, M., Zhang, X., Harris, C. D., Holt, J. B., & Croft, J. B. (2013). Spatial Disparities in the Distribution of Parks and Green Spaces in the USA. Annals of Behavioral Medicine : A Publication of the Society of Behavioral Medicine, 45(Suppl 1), 18– 27. https://doi.org/10.1007/s12160-012-9426-x

White, H. (1980). A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test for Heteroskedasticity. Econometrica, 48(4), 817–838. JSTOR. https://doi.org/10.2307/1912934

WHO | Social determinants of health. (2018, February 13). WHO. http://www.who.int/social_determinants/en/

WHO Expert Consultation. (2004). Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet (London, England), 363(9403), 157–163. https://doi.org/10.1016/S0140-6736(03)15268-3

228

Wieland, M. L., Tiedje, K., Meiers, S. J., Mohamed, A. A., Formea, C. M., Ridgeway, J. L., Asiedu, G. B., Boyum, G., Weis, J. A., Nigon, J. A., Patten, C., & Sia, I. G. (2015). Perspectives on physical activity among immigrants and refugees to a small urban community in Minnesota. Journal of Immigrant and Minority Health / Center for Minority Public Health, 17(1), 263–275. https://doi.org/10.1007/s10903-013-9917-2

Williams, D. R., Mohammed, S. A., Leavell, J., & Collins, C. (2010). Race, socioeconomic status, and health: Complexities, ongoing challenges, and research opportunities. Annals of the New York Academy of Sciences, 1186(1), 69–101. https://doi.org/10.1111/j.1749-6632.2009.05339.x

Wolch, J. R., Byrne, J., & Newell, J. P. (2014). Urban green space, public health, and environmental justice: The challenge of making cities ‘just green enough.’ Landscape and Urban Planning, 125, 234–244. https://doi.org/10.1016/j.landurbplan.2014.01.017

Wolch, J., Wilson, J. P., & Fehrenbach, J. (2005). Parks and Park Funding in Los Angeles: An Equity-Mapping Analysis. Urban Geography, 26(1), 4–35. https://doi.org/10.2747/0272-3638.26.1.4

Wold, B., & Mittelmark, M. B. (2018). Health-promotion research over three decades: The social-ecological model and challenges in implementation of interventions. Scandinavian Journal of Public Health, 46(20_suppl), 20–26. https://doi.org/10.1177/1403494817743893

Wong, R. J., Chou, C., Sinha, S. R., Kamal, A., & Ahmed, A. (2014). Ethnic disparities in the association of body mass index with the risk of hypertension and diabetes. Journal of Community Health, 39(3), 437–445. https://doi.org/10.1007/s10900- 013-9792-8

Woolford, S. J., Woolford, C. J., Sami, A., & Williams, D. R. (2013). 158. Barriers to Physical Activity Among Black Adolescent Girls and the Role of Ethnic Identity in a United States and Caribbean Sample. Journal of Adolescent Health, 52(2, Supplement 1), S96–S97. https://doi.org/10.1016/j.jadohealth.2012.10.228

Wooller, J. J., Barton, J., Gladwell, V. F., & Micklewright, D. (2016). Occlusion of sight, sound and smell during Green Exercise influences mood, perceived exertion and heart rate. International Journal of Environmental Health Research, 26(3), 267– 280. https://doi.org/10.1080/09603123.2015.1109068

Yancey, A. K., Siegel, J. M., & McDaniel, K. L. (2002). Role Models, Ethnic Identity, and Health-Risk Behaviors in Urban Adolescents. Archives of Pediatrics & Adolescent Medicine, 156(1), 55–61. https://doi.org/10.1001/archpedi.156.1.55

229

Yang, W., Spears, K., Zhang, F., Lee, W., & Himler, H. L. (2012). Evaluation of personal and built environment attributes to physical activity: A multilevel analysis on multiple population-based data sources. Journal of Obesity, 2012, 548910. https://doi.org/10.1155/2012/548910

Yang, Y. (2016). A dynamic framework on travel mode choice focusing on utilitarian walking based on the integration of current knowledge. Journal of Transport & Health, 3(3), 336–345. https://doi.org/10.1016/j.jth.2016.03.002

Yang, Y., & Diez-Roux, A. V. (2017). Adults’ Daily Walking for Travel and Leisure: Interaction Between Attitude Toward Walking and the Neighborhood Environment. American Journal of Health Promotion, 31(5), 435–443. https://doi.org/10.1177/0890117116669278

Yi, S. S., Beasley, J. M., Kwon, S. C., Huang, K.-Y., Trinh-Shevrin, C., & Wylie-Rosett, J. (2016). Acculturation and activity behaviors in Chinese American immigrants in New York City. Preventive Medicine Reports, 4, 404–409. https://doi.org/10.1016/j.pmedr.2016.08.007

Yi, S. S., Roberts, C., Lightstone, A. S., Shih, M., & Trinh-Shevrin, C. (2015). Disparities in meeting physical activity guidelines for Asian Americans in two metropolitan areas in the United States. Annals of Epidemiology, 25(9), 656- 660.e2. https://doi.org/10.1016/j.annepidem.2015.05.002

Yoon, E. (2011). Measuring ethnic identity in the Ethnic Identity Scale and the Multigroup Ethnic Identity Measure-Revised. Cultural Diversity & Ethnic Minority Psychology, 17(2), 144–155. https://doi.org/10.1037/a0023361

Yu, C.-Y. (2014). Environmental supports for walking/biking and traffic safety: Income and ethnicity disparities. Preventive Medicine, 67, 12–16. https://doi.org/10.1016/j.ypmed.2014.06.028

Zhang, W., Oh, S., & vom Hofe, R. (2018). Transforming Abandoned Rail Corridors into Multi-purpose Trails: Applying a Spatial Hedonic Approach to Estimating the Economic Benefits of the Little Miami Scenic Trail in Cincinnati, Ohio, USA. Urban Rail Transit, 4(4), 274–287. https://doi.org/10.1007/s40864-018-0094-4

Zuniga-Teran, A. A., Orr, B. J., Gimblett, R. H., Chalfoun, N. V., Marsh, S. E., Guertin, D. P., & Going, S. B. (2017). Designing healthy communities: Testing the walkability model. Frontiers of Architectural Research, 6(1), 63–73. https://doi.org/10.1016/j.foar.2016.11.005

230

APPENDIX A

ELIGIBILITY SCREENING FORM

Please answer the following:

FOR OFFICE USE ONLY

Date:____(MM)____(DD)_____(YY) Site ID:______

Participant ID: ______Staff Name: ______

1. Are you between 20 to 65 years of age? ___Yes ___No

2. Are you Korean, Vietnamese, or Chinese? ___Yes ___No

3. What is your date of birth? _____(YYYY)___(MM)___(DD)

4. Do you permanently live in one of the following zip codes (see below) of

Philadelphia County? ___Yes ___No

Zip codes included: 19102, 19103, 19104, 19106, 19107, 19109, 19111, 19112, 19114, 19115, 19116, 19118, 19119, 19120, 19121, 19122, 19123, 19124, 19125, 19126, 19127, 19128, 19129, 19130, 19131, 19132, 19133, 19134, 19135, 19136, 19137, 19138, 19139, 19140, 19141, 19142, 19143, 19144, 19145, 19146, 19147, 19148, 19149, 19150, 19151, 19152, 19153, 19154. [STAFF ONLY} Eligibility Criteria:

If the participant answered ‘No’ to Question 1, 2, or 3 – they are not eligible

231

APPENDIX B

IRB CONSENT FORMS

232

233

234

235

236

237

238

APPENDIX C

NEIGHBORHOOD ENVIRONMENT WALKABILITY SCALE [ENGLISH]

SAELENS ET AL., 2003

We would like to find out more information about the way that you perceive or think about your neighborhood. Please answer the following questions about your neighborhood and yourself.

A. Types of residences in your neighborhood – please circle the answer that best applies to you and your neighborhood.

1. How common are detached single-family residences in your immediate neighborhood? 1 2 3 4 5 None A few Some Most All

2. How common are townhouses or row houses of 1-3 stories in your immediate neighborhood? 1 2 3 4 5 None A few Some Most All

3. How common are apartments or condos 1-3 stories in your immediate neighborhood? 1 2 3 4 5 None A few Some Most All

4. How common are apartments or condos 4-6 stories in your immediate neighborhood? 1 2 3 4 5 None A few Some Most All

5. How common are apartments or condos 7-12 stories in your immediate neighborhood? 1 2 3 4 5 None A few Some Most All

6. How common are apartments or condos more than 13 stories in your immediate neighborhood? 1 2 3 4 5 None A few Some Most All

239

B. Stores, facilities, and other things in your neighborhood About how long would it take to get from your home to the nearest businesses or facilities listed below if you walked to them? Please put only one check mark ( ) for each business or facility. 1-5 6-10 11-20 20-30 30+ Don’t min min min min min know Example: gas station 1. ___ 2. ___ 3. __ 4. ___ 5. ___ 8. ___ 1. Convenience/small grocery store 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 2. Supermarket 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 3. Hardware store 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 4. Fruit/vegetable market 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 5. Laundry/dry cleaners 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 6. Clothing store 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 7. Post office 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 8. Library 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 9. Elementary school 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 10. Other schools 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 11. Book store 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 12. Fast food restaurant 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 13. Coffee place 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 14. Bank/credit union 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 15. Non-fast food restaurant 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 16. Video store 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 17. Pharmacy/drug store 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 18. Salon/barber shop 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 19. Your job or school 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ [check here ___ if not applicable] 20. Bus or train stop 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 21. Park 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 22. Recreation Center 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___ 23. Gym or fitness facility 1. ___ 2. ___ 3. ___ 4. ___ 5. ___ 8. ___

C. Access to services – Please circle the answer that best applies to you and your neighborhood. Both local and within walking distance mean within a 10-15 minute walk from your home.

1. Stores are within easy walking distance of my home 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

2. There are many places to go within easy walking distance of my home. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

3. It is easy to walk to a transit stop (bus, train) from my home.

240

1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

D. Streets in my neighborhood – Please circle the answer that best applies to you and your neighborhood.

1. The distance between intersections in my neighborhood is usually short (100 yards or less; the length of a football field or less) 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

2. There are many alternative routes for getting from place to place in my neighborhood (I don’t have to go the same way every time). 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

E. Places for walking and cycling – Please circle the answer that best applies to you and your neighborhood.

1. There are sidewalks on most of the streets in my neighborhood. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

2. Sidewalks are separated from the road/traffic in my neighborhood by parked cars. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

3. There is a grass/dirt strip that separate the streets from the sidewalks in my neighborhood. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

4. My neighborhood streets are well lit at night. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

5. Walkers and bikers on the streets in my neighborhood can be easily seen by people in their homes. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

6. There are crosswalks and pedestrian signals to help walkers cross busy streets in my neighborhood. 1 2 3 4

241

Strongly disagree Somewhat disagree Somewhat agree Strongly agree

F. Neighborhood surroundings/aesthetics – Please circle the answer that best applies to you and your neighborhood.

1. There are trees along the streets in my neighborhood. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

2. There are many interesting things to look at while walking in my neighborhood. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

3. There are many attractive natural sights in my neighborhood (such as landscaping, views). 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

4. There are attractive buildings/homes in my neighborhood. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

G. Traffic Hazards – please circle the answer that best applies to you and your neighborhood.

1. There is so much traffic along nearby streets that it makes it difficult or unpleasant to walk in my neighborhood. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

2. The speed of traffic on most nearby streets is usually slow (30 mph or less). 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

3. Most drivers exceed the posted speed limits while driving in my neighborhood. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

H. Crime - please circle the answer that best applies to you and your neighborhood.

1. There is a high crime rate in my neighborhood. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

242

2. The crime rate in my neighborhood makes it unsafe to go on walks during the day. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

3. The crime rate in my neighborhood makes it unsafe to go on walks at night. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

I. Other - please circle the answer that best applies to you and your neighborhood.

1. Parking is difficult in local shopping areas. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

2. The streets in my neighborhood do not have many cul-de-sacs (dead-end streets). 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

3. The streets in my neighborhood are hilly, making my neighborhood difficult to walk in. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

4. There are major barriers to walking in my local area that make it hard to get from place to place (for example, freeways, railway lines, rivers). 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

243

APPENDIX D

NEIGHBORHOOD PHYSICAL ACTIVITY QUESTIONNAIRE [ENGLISH]

GILES-CORTI ET AL., 2006

Walking: In the next section, we ask you about walking in and around your neighborhood, then in section B, we ask about walking outside your neighborhood. In both sections we ask you about two types of walking: walking for transport (e.g., to the shop), then walking for recreation, health and fitness. If the walking that you do for transport is also for recreation, health or fitness, please record it only once.

For example: Linda lives 20 minutes away from work. She chooses to walk there rather than drive mainly because she wants to improve her fitness. If Linda records that she walks for transport (3 times per week for a total of 120 minutes), she would not repeat that information under walking for recreation, health, or fitness.

Section A: This section is about walking IN AND AROUND your neighborhood or local area - we mean everywhere within a 10-15 minute walk of your home. Section B is about walking OUTSIDE your neighborhood.

1. In a USUAL WEEK, do you walk in or around your neighborhood or new local area to get to or from somewhere (such as walking to a shop or to public transport) or for recreation, health or fitness (including walking your dog)? ___No (If checked, go to section B) ___Yes

WALKING FOR TRANSPORT IN AND AROUND YOUR NEIGHBOURHOOD

2. In a USUAL WEEK, how many times do you walk as a means of transport, such as going to and from work, walking to the shop or walking to public transport in your neighborhood or local area?

Write in number of times_____ (If 0, go to question 5)

3. Please estimate the total time you spend walking as a means of transport in your neighborhood or local area in a USUAL WEEK. (e.g., 5 times by 10 minutes = 50 minutes). Hours_____ Minutes_____

4. Tick all the places where you walk to as a means of transport in or around your neighborhood or local area in a USUAL WEEK.

244

Places you might walk to as a means of Tick ALL transport IN your neighborhood or local the places area in a USUAL week you walk to in a USUAL week E.g., To or from shops __ To or from work [or study] ___ To or from shops ___ To or from school ___ To or from café or restaurant ___ To or from friend’s house ___ Somewhere else (1) : Please write where ______Somewhere else (2) : Please write where ______

WALKING FOR RECREATION, HEALTH OR FITNESS IN AND AROUND YOUR NEIGHBOURHOOD.

5. In a USUAL WEEK, how many times do you walk for recreation, health or fitness (including walking your dog) in or around your neighborhood or local area?

Write in number of times_____ (If 0, go to question 8)

6. Please estimate the total time you spend walking for recreation, health or fitness in or around your neighborhood or local area in a USUAL WEEK. (E.g., 5 times by 20 minutes = 100 minutes). Hours_____ Minutes_____

7. Please tick all the places where you walk for recreation, health or fitness in or around your neighborhood or local area in a USUAL WEEK.

Places you might walk to as a means of Tick ALL recreation/health/fitness IN your the places neighborhood or local area in a USUAL you walk to week in a USUAL week E.g., Park, oval, or bushlands __ Beach ___ Park, oval, or bushlands ___ Around the neighborhood using the ___ streets/footpaths (no specific destination) To or from café or restaurant ___ To or from a shop ___ Somewhere else (1) : Please write where ___ Somewhere else (2) : Please write where ___

245

Section B: This section is about walking OUTSIDE your neighborhood or local area - we mean everywhere further than a 15-minute walk from your home. (For example, somewhere you walk to in the next suburb, or somewhere you drive to).

WALKING FOR TRANSPORT OUTSIDE YOUR NEIGHBOURHOOD

8. In a USUAL WEEK, do you walk outside your neighborhood or local area to get to or from somewhere (such as walking to a shop or to public transport) or for recreation, health or fitness (including walking your dog)? ___No (If checked, go to Section C) ___Yes

9. In a USUAL WEEK, how many times do you walk as a means of transport, such as going to and from work, walking to the shop or walking to public transport outside your neighborhood or local area?

Write in number of times_____ (If 0, go to question 12)

10. Please estimate the total time you spend walking as a means of transport outside your neighborhood or local area in a USUAL WEEK. (E.g., 5 times by 10 minutes = 50 minutes). Hours_____ Minutes_____

11. Tick all the places where you walk to as a means of transport outside your neighborhood or local area in a USUAL WEEK.

Places you might walk to as a means of Tick ALL transport OUTSIDE of your neighborhood the places or local area in a USUAL week you walk to in a USUAL week E.g., To or from shops __ To or from work [or study] ___ To or from shops ___ To or from school ___ To or from café or restaurant ___ To or from friend’s house ___ Somewhere else (1): Please write where ___ Somewhere else (2): Please write where ___

WALKING FOR RECREATION, HEALTH OR FITNESS OUTSIDE YOUR NEIGHBOURHOOD.

12. In a USUAL WEEK, how many times do you walk for recreation, health or fitness (including walking your dog) outside your neighborhood or local area?

246

Write in number of times_____ (If 0, go to Section C)

13. Please estimate the total time you spend walking for recreation, health or fitness outside your neighborhood or local area in a USUAL WEEK. (E.g., 1 time for 30 minutes = 30 minutes). Hours_____ Minutes_____

14. Tick the places where you walk for recreation, health or fitness outside your neighborhood or local area in a USUAL WEEK

Places you might walk to as a means of Tick ALL the recreation/health/fitness OUTSIDE of your places you neighborhood or local area in a USUAL week walk to in a USUAL week E.g., Park, oval, or bushlands __ Beach ___ Park, oval, or bushlands ___ Around the neighborhood using the ___ streets/footpaths (no specific destination) To or from café or restaurant ___ To or from a shop ___ Somewhere else (1): Please write where ___ Somewhere else (2): Please write where ___

Section C: This section is about cycling (biking) IN AND AROUND your neighborhood or local area - we mean everywhere within a 10-15 minute walk of your home. Section D is about cycling OUTSIDE your new neighborhood.

15. In a USUAL WEEK, do you bike in or around your new neighborhood or new local area to get to or from somewhere (such as biking to a shop or to public transport) or for recreation, health or fitness? ___No (If checked, go to section D) ___Yes

CYCLING FOR TRANSPORT IN AND AROUND YOUR NEW NEIGHBOURHOOD

16. In a USUAL WEEK, how many times do you bike as a means of transport, such as going to and from work, biking to the shop or biking to public transport in your neighborhood or local area?

Write in number of times_____ (If 0, go to question 19)

17. Please estimate the total time you spend biking as a means of transport in your neighborhood or local area in a USUAL WEEK. (e.g., 5 times by 10 minutes = 50 minutes). Hours_____ Minutes_____

247

18. Tick all the places where you bike to as a means of transport in or around your neighborhood or local area in a USUAL WEEK.

Places you might bike to as a means of Tick ALL transport IN your neighborhood or local the places area in a USUAL week you bike to in a USUAL week E.g., To or from shops __ To or from work [or study] ___ To or from shops ___ To or from school ___ To or from café or restaurant ___ To or from friend’s house ___ Somewhere else (1) : Please write where ______Somewhere else (2) : Please write where ______

CYCLING FOR RECREATION, HEALTH OR FITNESS IN AND AROUND YOUR NEW NEIGHBOURHOOD.

19. In a USUAL WEEK, how many times do you bike for recreation, health or fitness in or around your neighborhood or local area?

Write in number of times_____ (If 0, go to question 22)

20. Please estimate the total time you spend biking for recreation, health or fitness in or around your neighborhood or local area in a USUAL WEEK. (E.g., 5 times by 20 minutes = 100 minutes).

Hours_____ Minutes_____

21. Please tick all the places where you bike for recreation, health or fitness in or around your neighborhood or local area in a USUAL WEEK.

Places you might bike to as a means of Tick ALL the recreation/health/fitness IN your places you neighborhood or local area in a USUAL week bike to in a USUAL week E.g., Park, oval, or bushlands __ Beach ___ Park, oval, or bushlands ___ Around the neighborhood using the ___ streets/footpaths (no specific destination)

248

To or from café or restaurant ___ To or from a shop ___ Somewhere else (1) : Please write where ___ Somewhere else (2): Please write where ___

Section D: This section is about cycling OUTSIDE your neighborhood or local area - we mean everywhere further than a 15-minute walk from your home. (For example, somewhere you walk to in the next suburb, or somewhere you drive to).

CYCLING FOR TRANSPORT OUTSIDE YOUR NEW NEIGHBOURHOOD

22. In a USUAL WEEK, do you bike outside your neighborhood or local area to get to or from somewhere (such as biking to a shop or to public transport) or for recreation, health or fitness? ___No (If checked, go to Section E) ___Yes

23. In a USUAL WEEK, how many times do you bike as a means of transport, such as going to and from work, biking to the shop or biking to public transport outside your neighborhood or local area?

Write in number of times_____ (If 0, go to question 26)

24. Please estimate the total time you spend biking as a means of transport outside your neighborhood or local area in a USUAL WEEK. (E.g., 5 times by 10 minutes = 50 minutes). Hours_____ Minutes_____

25. Tick all the places where you bike to as a means of transport outside your neighborhood or local area in a USUAL WEEK.

Places you might bike to as a means of Tick ALL transport OUTSIDE your neighborhood or the places local area in a USUAL week you bike to in a USUAL week E.g., To or from shops __ To or from work [or study] ___ To or from shops ___ To or from school ___ To or from café or restaurant ___ To or from friend’s house ___ Somewhere else (1) : Please write where ___ Somewhere else (2) : Please write where ___

CYCLING FOR RECREATION, HEALTH OR FITNESS OUTSIDE YOUR NEIGHBOURHOOD.

249

26. In a USUAL WEEK, how many times do you bike for recreation, health or fitness outside your neighborhood or local area?

Write in number of times_____ (If 0, go to Section E)

27. Please estimate the total time you spend biking for recreation, health or fitness outside your neighborhood or local area in a USUAL WEEK. (E.g., 1 time for 30 minutes = 30 minutes). Hours_____ Minutes_____

28. Tick the places where you bike for recreation, health or fitness outside your neighborhood or local area in a USUAL WEEK

Places you might bike to as a means of Tick ALL transport OUTSIDE your neighborhood or the places local area in a USUAL week you bike to in a USUAL week E.g., Park, oval, or bushlands __ Beach ___ Park, oval, or bushlands ___ Around the neighborhood using the ___ streets/footpaths (no specific destination) To or from café or restaurant ___ To or from a shop ___ Somewhere else (1): Please write where ___ Somewhere else (2): Please write where ___

Section E: The next set of questions is about other leisure time physical activities that you do IN A USUAL WEEK, besides what you have already mentioned. Do not include walking or cycling.

29. In a USUAL WEEK, do you do any other vigorous or moderate intensity leisure time physical activities? Do not include any walking or cycling. ___No (If checked, go to Section F) ___Yes

30. In a USUAL WEEK, do you do any vigorous intensity leisure time physical activities like jogging, aerobics or competitive tennis? Do not include walking or cycling or moderate intensity physical activities. Vigorous intensity physical activities make you breathe harder or puff and pant. ___No (If checked, go to question 33) ___Yes

31. In a USUAL WEEK, how many times do you do vigorous intensity leisure time physical activities which makes you breathe harder or puff and pant? Write in number of times_____

250

32. What do you estimate is the total time you spend doing vigorous intensity leisure time physical activities in a USUAL WEEK? (E.g., 3 times for 20 minutes = 60 minutes) Hours_____ Minutes_____

33. Apart from what you have already mentioned, in a USUAL WEEK do you do any other moderate intensity leisure time physical activities like gentle swimming, social tennis, golf or heavy gardening? Moderate intensity physical activities do not make you breathe harder or puff and pant. ___No (If checked, go to Section F) ___Yes

34. In a USUAL WEEK, how many times do you do moderate intensity leisure time physical activities which do not make you breathe harder or puff and pant? Write in number of times_____

35. What do you estimate is the total time you spend doing moderate intensity leisure time physical activities in a USUAL WEEK? (E.g., 1 time for 1 hour = 1 hour) Hours_____ Minutes_____

251

APPENDIX E

NEIGHBORHOOD SELF-SELECTION [ENGLISH]

FRANK ET AL., 2007

Please answer the following questions about why you chose to live in your neighborhood.

1. I live in this neighborhood because housing is affordable. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

2. I live in this neighborhood because it is easy to walk to places I need to get to. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

3. I live in this neighborhood because the transportation costs (e.g., public transportation) are low. 1 2 3 4 Strongly disagree Somewhat disagree Somewhat agree Strongly agree

252

APPENDIX F

MULTI-ETHNIC IDENTITY MEASURE – REVISED (MEIM-R) [ENGLISH]

PHINNEY & ONG, 2007

1. I have spent time trying to find out more about my ethnic group, such as its history, traditions and customs. 1 2 4 3 4 Strongly Somewhat Neutral Somewhat Strongly agree disagree disagree agree

2. I have a strong sense of belonging to my own ethnic group. 1 2 4 3 4 Strongly Somewhat Neutral Somewhat Strongly agree disagree disagree agree

3. I understand pretty well what my ethnic group membership means to me. 1 2 4 3 4 Strongly Somewhat Neutral Somewhat Strongly agree disagree disagree agree

4. I have often done things that will help me understand my ethnic background better. 1 2 4 3 4 Strongly Somewhat Neutral Somewhat Strongly agree disagree disagree agree

5. I have often talked to other people in order to learn more about my ethnic group. 1 2 4 3 4 Strongly Somewhat Neutral Somewhat Strongly agree disagree disagree agree

6. I feel a strong attachment towards my own ethnic group. 1 2 4 3 4 Strongly Somewhat Neutral Somewhat Strongly agree disagree disagree agree

253

APPENDIX G

SOCIODEMOGRAPHIC AND HEALTH QUESTIONS

1. Address A) Street number: ______B) Street name: ______apt/suite/unit______C) City: ______D) Zip code: ______2. How long have you lived at your current address? _____years 3. In what country were you born? ______4. What is your yearly income? a. ___Less than $20,000 c. ___Between $30,000-39,999 b. ___Between $20,000-29,999 d. ___$40,000 and above 5. Are you currently employed? (please circle) a. Yes b. No 6. What is the highest grade of school you completed? (please circle) a. No education or elementary school b. Below high school graduate c. High School d. College and above 7. How well do you speak English? (please circle) a. Not at all b. Not well c. I speak English well d. I speak English fluently 8. What is your height without shoes ______feet ___ inches (or ______cm) 9. What is your weight without shoes ______lbs (or ______kg) 10. Have you ever been told by your health care provider that you have any of the following conditions (check all that apply): a. ___Hypertension/high blood f. ___Sleep Apnea pressure g. ___Obesity b. ___Diabetes/high blood sugar h. ___Asthma c. ___High cholesterol i. ___Arthritis (rheumatoid, gout, d. ___Heart attack/Coronary heart fibromyalgia) disease j. ___Any type of cancer (please e. ___Stroke specify):______

254

APPENDIX H

MAP OF DISTRIBUTION OF ASIANS LIVING IN PHILADELPHIA (CENSUS DATA:

ARCMAP 10.7.1)

255

APPENDIX I

SUMMARY OF STUDY FINDINGS

Summary of Associations between the Built Environment and Engaging in Physical Activity (Count Models) Walking in AT Recreational Moderate Vigorous PA MVPA Overall PA neighborhood Walk/Bike PA Residential density ns ns - Land use mix-diversity ns ++ + Land use mix – access ns + + Street connectivity ns ns ns Infrastructure and safety ns ns ns RQ 1&2: Aesthetics ns ns ns Perceived Traffic hazardsª ns ns ns Neighborhoo Crimeª ns - ns d Walkability Parking difficulty shopping area ns ns ns (NEWS) Lack of cul-de-sacs - - -- Hilly streets in neighborhoodª ns ns ns Physical barriers to walkingª - ns ns Overall walkability ns ns ns ns ns ns Ethnic Identity x NEWS - ns + ns ns Walk Score ns ns ns Walk time to park + + ns + RQ 3&4: Walk time to gym ns ns ns ns Objective Walk time to specialty exercise ns + + ns Measures Drive time to park ns ns + + Drive time to gym ns ns ns ns Drive time to specialty exercise ns + + ns + significant positive coefficient at p<0.05 || + significant positive coefficient at p<0.001 || - significant negative coefficient at p<0.05 -- significant negative coefficient at p<0.001 || ns not significant || lack of symbol denotes that no regression models were ran

256

Summary of Associations between the Built Environment and Odds of Not Engaging in Physical Activity (Binary Models) Walking in AT Recreational Moderate Vigorous PA MVPA Overall PA neighborhood Walk/Bike PA Residential density + ++ ns Land use mix-diversity ns - -- Land use mix – access ns - - Street connectivity ns ns ns Infrastructure and safety ns ns ns RQ 1&2: Aesthetics ns ns - Perceived Traffic hazardsª ns ns ns Neighborhoo Crimeª ns ns ns d Walkability Parking difficulty shopping area ns - - (NEWS) Lack of cul-de-sacs + ns + Hilly streets in neighborhoodª ns ns ns Physical barriers to walkingª + ns ns Overall walkability ns ns - ns ns - Ethnic Identity x NEWS - ns + ns ns Walk Score + ns ns Walk time to park ns + ns ns RQ 3&4: Walk time to gym ns ns ns ns Objective Walk time to specialty exercise + ++ ns ns Measures Drive time to gym ns ns ns ns Drive time to park ns ns ns ns Drive time to specialty exercise ns ns ns ns + significant positive coefficient at p<0.05 || + significant positive coefficient at p<0.001 || - significant negative coefficient at p<0.05 -- significant negative coefficient at p<0.001 || ns not significant || lack of symbol denotes that no regression models were ran

257