Latina Adolescent Suicide: Examining the Effects of Cultural Status and Parental, Peer and Teacher Supports

LATINA ADOLESCENT SUICIDE: EXAMINING THE EFFECTS OF CULTURAL STATUS AND PARENTAL, PEER AND TEACHER SUPPORTS

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

Susan M. De Luca, MSW

Social Work Graduate Program

The Ohio State University

2009

Dissertation Committee:

Keith L. Warren, MSW, Ph.D., Advisor Thomas K. Gregoire, MSW, Ph.D. Lisa Raiz, MSW, Ph.D. Ernesto R. Escoto, PhD

Susan M. De Luca

2009

ABSTRACT

This study utilized data from the National Longitudinal Study of Adolescent

Health (Add Health) (Waves I and II) to examine variations in Latina adolescents’ level

of suicide proclivity as predicted by friendship networks, family support, cultural status

and behavioral health (Harris, Mullan, Halpern, Entzel, Tabor, Bearman & Udry, 2008).

By examining these correlates of suicide, this study concentrated on two specific

objectives set forth by Healthy People 2010: reducing the number of adolescents who experience suicide and reducing the number of future adults who ideate or attempt suicide (Centers for Disease Control [CDC], 2001). Add Health’s stratified cluster design and over-sampling obtained an appropriate number of Latinas (1,618) in the sample. The data include network-level variables to distinguish particular peer group factors that predict suicide. Measures used in this study explored the extent to which cultural status, self-reported parental support, friendship characteristics (reciprocity and suicidal friends) and behavioral health status (depression and substance use) were linked to suicide

proclivity (ideations and attempts).

Logistic and negative binomial regression analyses were used to observe any

predictive or mediating relationships the types of friendships and familial supports had

with suicide proclivity. The goal was to ascertain the way in which the cultural status of

the Latinas, the type of friends they reported and the level of social support they received

from their parents and teachers determined their incidence of suicidal ideation and

attempts. Determining the extent to which these factors help account for variation in suicidal thoughts and behaviors will aid in proposing effective strategies for prevention.

Empirically-based suicide prevention programs geared specifically for Latinas are important given that these females have higher suicide attempt rates when compared to

African American and White teens (Centers for Disease Control [CDC], 2008). Although the U.S Bureau of the Census does not provide information explicitly regarding Latina adolescents, the Latino population continuing to grow in the U.S. Almost half (48%) of the Latino population comprised of children, preventing suicide among Latina

adolescents is essential (CDC, 2006).

iii

Dedicated to Rosina & Antonio Arena for teaching me the importance of culture

and

Anna & Joseph De Luca,

for their guidance, love and support

ACKNOWLEDGEMENTS

First, I would like to thank my committee, Dr. Keith Warren for being interested from day-one about my research, Dr. Tom Gregoire for reminding me of my “rock star status” even when my confidence waned, Dr. Ernesto Escoto for sharing his immense knowledge of

Latino issues and taking time to be apart of this study and Dr. Lisa Raiz for coming up-to-bat when I needed her the most.

I would also like to acknowledge Dr. Celeste Burke for her guidance, support and hugs throughout my doctoral program.

My eternal gratitude goes to my parents, Anna and Joseph De Luca as well as the rest of my family, Rose-Marie, Bo, Logan and Sienna Kendall, Nick and Sandy Arena and special thanks to my cugino, Nicky. Your encouragement, love and laughs got me through the hard times. This is as much your PhD as mine.

I am also grateful to Dr. Michael S. Spencer for continually inspiring me throughout my graduate work. You have changed the way I think.

I would also like to acknowledge my fellow College of Social Work cohort members, especially Linda Helm, Nick Martt, Amy Mendenhall, Monica Mickles Ascar for keeping me on track and always reminding me that the end was not only possible, but around the corner.

I want to thank my friends who balanced my life outside of the classroom, namely

Karen Parker, my partner in crime at Michigan and beyond, Rosario Martinez for cheering me on every day and Jason Melton for his support and encouragement. Lastly, I would be

remiss to not thank in a special way Jeanne Bennett, for her help with this document and guidance throughout our long friendship.

DISCLAIMER

This document includes data from the National Longitudinal Study of

Adolescent Health which was developed under contract with the Carolina Population

Center at the University of North Carolina at Chapel Hill. The information and opinions expressed herein reflect solely the position of the author.

vii

VITA

August 28, 1973 ...... Born – Parma, OH

2001...... Master’s of Social Work,

The University of Michigan, Ann Arbor, MI

1997...... Bachelor’s of Arts,

Kent State University, Kent, OH

2004- Present ...... Teaching Assistant

Research Assistant

The Ohio State University, Columbus, OH

PUBLICATIONS

De Luca, S. (2007). Las hijas de Juan : Daughters Betrayed. [Review of the book Las hijas de Juan : Daughters Betrayed]. Families in Society, 8(4), 666-667.

Lewis, M., Phillips, G., Bowser, M., De Luca, S., Johnson, H. & Rosen, T. (2009). Cocaine-exposed infant behavior during Still-Face test: risk factor analyses. American Journal of Orthopsychiatry.79(1), 60-70.

FIELDS OF STUDY

Major Field: Social Work

viii

TABLE OF CONTENTS

ABSTRACT ...... ii

DEDICATION………………………………………………………….………………..iv

ACKNOWLEDGMENTS ...... v

DISCLAIMER ...... vii

VITA ...... viii

TABLE OF CONTENTS ...... ix

LIST OF TABLES ...... xii

Chapter 1

Introduction ...... 1

Dissertation Overview ...... 2

Rationale for Study ...... 2

Latino Demographics ...... 4

Summary…………………………………………………………………………………..5

Chapter 2

Literature Review...... 7

Introduction ...... 7

Theorectical Framework ...... 9

Suicide Proclivity ...... 13

Cultural Status ...... 15

Behavioral Health Status ...... 20

Social Supports ...... 24

Suicidal Friends…………………...... 34

Traditional Latino Family Structure ...... 35

Summary ...... 39

Chapter 3

Methodology ...... 42

Description of the Add Health Dataset ...... 42

Measures ...... 47

Hypotheses Tested ...... 52

Summary of Hypotheses ...... 54

Analytic Strategy ...... 55

Chapter 4

Results ...... 62

Univariate Analyses ...... 63

Latina Suicide Attempts and Social Supports ...... 70

Model Development ...... 70

Summary of Negative Binomial Regressions ...... 82

Latina Suicidal Ideations and Social Supports…………………………………………...82

Summary of Logistic Regressions ...... 9 1

Results Summary ...... 92

Hypotheses Tested ...... 92

Chapter 5

Discussion ...... 95

Suicidal Ideations and Attempts ...... 96

Cultural Status ...... 96

Family Support...... 96

School Support ...... 97

Suicidal Friends ...... 99

Frequency of Contact ...... 99

Depression...... 100

Substance Use ...... 101

Reciprocity ...... 102

Social Acceptance ...... 102

Latina Adolescents and Suicide Attempts……………………………………………...103

Latina Adolescents and Suicidal Ideations……………………………………………..106

Limitations……………………………………………………………………………...109

Strengths………………………………………………………………………………..110

Policy and Research Implications………………………………………………………111

Conclusion ...... 112

References ...... 113

LIST OF TABLES

Table Page

Table 1. Descriptive Statistics of Perceived Continuous & Count Family

Support Variables ...... 64

Table 2. Descriptive Statistics of Count School Support Variables ...... 65

Table 3. Descriptive Statistics of Frequency of Contact ...... 65

Table 4. Descriptive Statistics of Continuous Behavioral Health Variables ...... 66

Table 5. Descriptive Statistics of Social Acceptance Variable Table ...... 67

Table 6. Descriptive Statistics of Suicidal Ideations ...... 67

Table 7. Proportion Estimates of Count Variables ...... 70

Table 8. Summary of Negative Binomial Regression Analysis for Variables Predicting

Cultural Status (n = 1,618) Controlling for Depression ...... 73

Table 9. Summary of Negative Binomial Regression Analysis for Variables Predicting

Suicidal Attempts by Cultural Status and Family Support (n = 1,618) Controlling for

Depression...... 74

Table 10. Summary of Negative Binomial Regression Analysis for Variables Predicting

Suicidal Attempts by Cultural Status and Family Support (n = 1,618) Controlling for

Depression...... 74

xii

Table 11. Summary of Negative Binomial Regression Analysis for Variables Predicting

Suicidal Attempts by Cultural Status and Family and Teacher Support (n = 1,618)

Controlling for Depression ...... 75

Table 12. Summary of Negative Binomial Regression Analysis for Variables Predicting

Suicidal Attempts by Cultural Status and Family and Teacher Support (n = 1,618)

Controlling for Depression ...... 76

Table 13. Summary of Negative Binomial Regression Analysis for Variables Predicting

Suicidal Attempts by Teacher Supports and Suicidal Friends (n = 1,618) Controlling for

Depression...... 77

Table 14. Summary of Negative Binomial Regression Analysis for Variables Predicting

Suicidal Attempts by Cultural Status and Parental Support (n = 1,618) Controlling for

Depression...... 78

Table 15. Summary of Negative Binomial Regression Analysis for Variables Predicting

Suicidal Attempts by Cultural Status, Parental and Teacher Support (n = 1,618)

Controlling for Depression ...... 79

Table 16. Summary of Negative Binomial Regression Analysis for Variables Predicting

Suicidal Attempts by Cultural Status, Parental and Teacher Support and Friendship

Characteristics (n = 1,618) Controlling for Depression ...... 80

Table 17. Summary of Negative Binomial Regression Analysis for Variables Predicting

Suicidal Attempts by Parental and Teacher Support and Friendship Characteristics (n =

1,618) Controlling for Depression ...... 81

xiii

Table 18. Summary of Logistic Regression Analysis for Variables Predicting Suicidal

Attempts by Cultural Status (n = 1,618) Controlling for Depression ...... 83

Table 19. Summary of Logistic Regression Analysis for Variables Predicting Suicidal

Ideations by Cultural Status and Parental Support (n = 1,618) Controlling for

Depression...... 83

Table 20. Summary of Logistic Regression Analysis for Variables Predicting Suicidal

Ideations by Cultural Status and Parental Support (n = 1,618) Controlling for

Depression ...... 84

Table 21. Summary of Logistic Regression Analysis for Variables Predicting Suicidal

Ideations by Cultural Status, Parental Support (n = 1,618) Controlling for Depression ...85

Table 22. Summary of Logistic Regression Analysis for Variables Predicting Suicidal

Ideations by Cultural Status, Parental and Teacher Support (n = 1,618) Controlling for

Depression...... 85

Table 23. Summary of Logistic Regression Analysis for Variables Predicting Suicidal

Ideations by Cultural Status, Parental Support and Friendship Characteristics (n = 1,618)

Controlling for Depression ...... 86

Table 24. Summary of Logistic Regression Analysis for Variables Predicting Suicidal

Ideations by Parental Support and Friendship Characteristics (n = 1,618) Controlling for

Depression...... 88

Table 25. Summary of Logistic Regression Analysis for Variables Predicting Suicidal

Ideations by Cultural Status and Parental Support (n = 1,618) Controlling for

Depression...... 89

xiv

Table 26. Summary of Logistic Regression Analysis for Variables Predicting Suicidal

Ideations by Cultural Status, Parental and Teacher Support (n = 1,618) Controlling for

Depression...... 89

Table 27. Summary of Logistic Regression Analysis for Variables Predicting Suicidal

Ideations by Cultural Status and Parental Support and Friendship Characteristics (n =

1,618) Controlling for Depression ...... 90

Table 28. Summary of Logistic Regression Analysis for Variables Predicting Suicidal

Ideations by Parental Support and Suicidal Friends (n = 1,618) Controlling for

Depression...... 91

CHAPTER 1

INTRODUCTION

Suicide continues to be the third-leading cause of death among 10-24 year olds in the United States (Substance Abuse and Mental Health Services Administration

[SAMHSA], 2006). Adolescent suicide represents 12.3% of all deaths among 15-24 year olds and accounted for 11 deaths each day within this age group (American Association of Suicidology [AAS], 2004). Among 10-14 year olds, death rates by suicide have been estimated at 1.7/100,000 (SAMHSA, 2006b). Suicide is associated with a variety of risk factors in a child’s life including interpersonal issues such as family conflict, a mental health condition and/or drug use (Olvera, 2001).

Latina adolescents have the highest rates of hopelessness, suicidal ideations, plans, attempts and injuries sustained from an attempt when compared to their peers regardless of gender and race/ethnicity (CDC, 2008). These females strive to simultaneously navigate between their native Latin American and new American cultures but eventually acculturate (Vega, Zimmerman, Gil & Warheit, 1994). The literature is divided on the effects of acculturation and the risk of suicide among Latinos and Latinas.

While it is thought that adopting new cultural norms and traditions can result in feelings of despair, frustration and isolation, and eventually leading to suicide proclivity (Zayas,

1987; Vega et al, 1994; Turner, Kaplan, Zayas & Ross, 2002) others have found that

acculturation is a protective factor against suicide (Hines & Caetano, 1998; Marin,

Gomez, Tschann, Gregorich, 1997).

This dissertation explored the variables associated with Latina adolescents and their high rates and increased risk of suicidal ideations and attempts. In particular, the dissertation looked at how their choices in social supports translated to their incidence of suicide. Data from the National Longitudinal Study of Adolescent Health (Add Health) was analyzed. The data are compiled from a nationally representative school-based questionnaire of adolescents who attended U.S middle and high schools in 1994 through

1996. In-home questionnaires were also given to a proportion of students who completed the in-school survey where more sensitive information, such as suicidality and drug use, were addressed.

Dissertation Overview

This dissertation is divided into five chapters. The first chapter is an introduction to the study and its relevance to the field. The second chapter provides a review of the existing literature on adolescent suicide with an emphasis on Latinas. The third chapter explains the research methods. The fourth chapter presents the results of the analyses.

The final chapter offers a discussion of the implications for social work practice and future directions for research.

Rationale for the Study

Latinas are disproportionately at risk for suicide attempts in the United States.

Latinos constitute approximately 15% of the total U.S. population (Census, 2008) but represent disproportionally higher rates of maladaptive functioning compared to their

peers. Latina adolescents reported feeling sad or hopeless in the past year (42%), had

thoughts of attempting suicide (21%), made a plan (15%) or attempted (14%). These

rates were higher than their African American and White counterparts. Of those Latina

teens who had attempted in the past year, one out of 25 injured themselves severely

enough to seek medical care, twice the rate of White and African American females

(CDC, 2008).

One factor related to a Latina’s high rate of suicidal thoughts and actions is how

she transitions from her native culture to mainstream U.S. culture (Trautman, 1961a;

1961b; Zayas, 1987; Turner, Kaplan, Zayas & Ross, 2002). A Latina’s ability to

acculturate, but also the emotional support she receives from her family members, has

been found to be critical to her mental health status. Given that many Latinos do not

fully acculturate until the second or third generation (Fishman, 1966; Veltman, 1983),

there could be noticeable differences in potential suicidal risk factors contingent upon

their cohort and cultural status (Pena, Wyman, Brown, Matthieu, Olivares, Hartel &

Zayas, 2008).

More research is needed to effectively prevent Latina adolescent suicide. Latino males and females under the age of 18 accounted for 26% of all Latino suicides (CDC,

2006). The Latino population is projected to increase in the United States for the next fifty years (Census, 2002). Therefore, deciphering the suicidal risk factors among young

Latinas will continue to be a public health concern.

Latino Demographics

Latinos are the largest racial/ethnic minority group in the United States and they are projected to comprise 24% of the U.S population and over 31% of all children by

2050 (Census, 2002). In 2007 alone, 1.4 million Latinos arrived in the United States from the Caribbean, Central and South America (Census, 2008). Latino children are the fastest-growing portion of the U.S. population and accounted for over 48% of all Latinos.

In 2008, the median age of Latinos was 27.6 and 24% for those under 18 years old

(Census, 2008). Of the 12.1 million Latino children (age 18 and younger) included in the

2000 census, more than half (6.7 million) had foreign-born parents (Census, 2004). Over half of U.S. foreign-born children in the U.S. (54%) were of Latino origin (Census,

2004). The Latino population has increased significantly since the 2000 Census report.

Last year the Latino youth population totaled 15.3 million (roughly 1 out of 3 children under the age of 18 are Latino) (Census, 2008). Suicide is rare in pre-adolescence but the risk tends to increase dramatically with age (Jellinek & Snyder, 1998). Seeing how these youth comprise such a large and growing portion of the overall U.S. population, it is vital to fully understand and prevent Latina suicide.

Beyond the public health concerns of suicide, there are also economic factors that result from this phenomenon. Although economists have not specifically examined the financial costs of suicide with regard to adolescents, in general suicide expenditures in the U.S. reach $111.3 billion each year in terms of medical care, mental health services and loss of work (Leone, n.d). Of that amount, $3.7 billion were incurred by medical expenses and $80.2 billion estimated lost concerning the quality of life for these

individuals (Leone, n.d). Therefore, preventing suicide is not only a health issue but an economical threat to the country.

Estimating the economic cost of suicide is difficult. The human costs are just as daunting. According the SAMHSA (n.d) from 1952-1995 suicide adolescents nearly tripled. To understand the magnitude of suicide, more adolescents “die from suicide more than cancer, heart disease, AIDS, birth defects, stroke, pneumonia and influenza and chronic lung disease combined” (SAMHSA, n.d., para. 12).

Summary

Latina adolescent suicide is a multi-faceted issue for mental health, medical, educational and other systems serving teens. Antecedents associated with Latina suicide include factors linked with one’s cultural status, emotional health and substance abuse in addition to securing social and emotional support. These determinants present many challenges to those offering prevention services.

The study will focus on the following areas:

• Does a Latina’s cultural status matter in terms of her likelihood to ideate or

attempt suicide?

• Does a Latina’s behavioral status increase or decrease her risk of suicide

proclivity?

• Does a Latina’s social support from her parents, peers or teachers raise or lessen

her propensity for suicidal thoughts or attempts?

The social work profession has a long history of serving high-risk populations. A current trend in the field is emphasizing empirically-based, culturally competent mental

health services (Hoagwood, Burns, Kiser, Ringeisen & Schoenwald, 2001). The goal of this dissertation is to provide policy makers with empirical evidence to create culturally competent prevention efforts for Latina adolescents.

CHAPTER 2:

LITERATURE REVIEW

This chapter will concentrate on the risk and protective factors related to adolescent suicide with a focus on Latinas. Each section will deal with a particular variable in the conceptual model for the study. This study adds to the field of suicidology by investigating the applicability of attachment theory to suicidal thoughts and behaviors in terms of friendship networks and adult supports.

Introduction to the topic

Adolescent suicide. Adolescents who contemplate or attempt suicide may have

numerous and multifaceted problems, including issues with mental health, substance

abuse and poor social supports. Although rates of suicide have decreased for most age

groups in the United States, in the past 60 years suicide rates doubled for females ages

15-24 (AAS, 2004). Substance abuse and mental health problems among youth

contributed to the increased suicide rates (Shaffer, Gould, Fisher, Trautment, Moreau,

Kleinman, & Flory, 1996; Birckmayer & Hemenway, 1999).

Adolescents who report stronger support systems tend to have lower attempt rates.

Social isolation, typically associated with depression and hopelessness, decreases vital

social support and increases risky behaviors related to suicide. Females attempt suicide

more than males and this finding has been linked to females’ needs for more emotional

supports in their relationships than boys (AAS, 2006).

Suicide Among Racial/Ethnic Minorities. Racial/ethnic minority females are three

times more likely to attempt suicide than minority boys (Groves, Stanley & Sher, 2007).

Although some demographic groups appear to have a higher incidence of suicide (such as adolescents), there are subgroups within this population that have even more disproportionate rates of suicide. Latina suicide rates, ideations, plans and attempts,

consistently rank highest among adolescents (CDC, 2005; 2008). Latina adolescents are

especially vulnerable to suicidal thoughts and behaviors. Robles (1995) found that in a

Florida school system, Latinas accounted for 35% of the district’s suicide attempts but

constituted only 22% of the school population. In one Bronx hospital, more than 25% of

the patients admitted were Latina adolescents recovering from a recent suicide attempt

(Razin, O’Dowd, Nathan, Rodriguez, Goldfield, Martin, Goulet, Scheftel, Mezan &

Mosca, 1991). Ng (1996) reported that suicide attempts counted for 66% of Latina

adolescent hospitalizations. Research involving over 3,000 urban teens reported that

being Latina was a powerful indicator for suicide attempts (Garafolo, Wolf, Wissow,

Woods & Goodman, 1999). In the literature, Latina adolescents have consistently been

found to have been depressed, ideated, made a suicide plan, attempted suicide and

sustained an injury from an attempt more than their African American, Asian American

and Caucasian peers (CDC, 2008).

Theoretical Framework

John Bowlby and Mary Ainsworth developed the concept of attachment theory, which guides this study (Ainsworth & Bowlby, 1991). This theory draws from a number of disciplines including developmental psychology and psychoanalysis. Bowlby believed that the types of bonds children had to their attachment figure, typically their mother, can positively or negatively affect their mental health. An attachment figure can meet a variety of needs including support, protection, and care. By having a person a child feels safe and secure with, the child can discover his or her environment, which is essential to healthy development. Children who do not have such an attachment figure can potentially deal with various pathogenic conditions (Bowlby, 1969). Bowlby contended that in order for a child to be healthy “the young child should experience a warm, intimate, and continuous relationship with his mother (or permanent mother substitute) in which both find satisfaction and enjoyment” (Bowlby, 1951, p. 13).

Some have argued that Bowlby and Ainsworth discounted the importance of other adults and family members and supported a traditional view of women, in a child’s life

(Chodorow, 1978; Johnson, 1988; Marris, 1982). While mothers are typically thought of as primary caregivers, there is evidence that children can bond with a variety of individuals including fathers, grandparents and siblings (Schaffer & Emerson, 1964).

Somewhat overlooked, Bowlby also stressed the importance of social networks in the progression of the mother-child connection. Other studies have supported the view of social networks as attachment figures (Belsky & Isabella, 1988; Belsky, Rovine, &

Taylor, 1984). The aforementioned studies found that social networks were found to be

beneficial if they were consistent, quality relationships while number of contacts did not

appear to matter to a person’s mental health. Studies that include teens such as Perez-

Smith, Spirito and Boergers’s (2002) work, found that teens who reported tenuous social

networks also had higher levels of hopelessness, even when depression was controlled.

Another study examined teen’s peer networks and risk of suicide (Brent, Perper, Moritz,

Allman, Schweers, Roth et al., 1993). Brent and colleagues studied 26 suicide victim’s

social network and interviewed 146 teens 7 months after their friend’s death. These

participants’ level of depression, post-traumatic stress disorder and suicidal ideations

were assessed. The new-onset disorders including ideation occurred within one month of

their friend’s death.

The break of secure and continuous relationships in a child’s life can have

deleterious effects depending on how the child reacts to these changes.

Three stages of separation were classified by Robertson and Bowlby (1952): protest (a

result of separation anxiety), despair (associated with grief and bereavement), and

detachment (linked to defense mechanisms). Bowlby argued that infants and children are

subjected to separation anxiety. He believed that when a secure figure it not available, a

child might find others for their attachment needs. He continued that if a child had too

many substitutes of attachment-like figures, he or she may never fully learn how to

develop strong relationships with others. Without consistent figures in a child’s life, he

or she lacks the ability to know and have faith in predictable outcomes. This lack of

guidance increases their anxiety levels, rates of depression and lack of worth.

When caregivers respond positively, they model appropriate behaviors of social interaction (Schaffer & Emerson, 1964). This repetitive interchange provides a safe environment for a child to continually return to when reassurance is needed (Ainsworth,

1967; Schaffer & Emerson, 1964). If this person provides the child with comfort and security, the child creates a working archetype of self worth. If this does not occur, the child begins to view him or herself as undervalued and useless. The type of person the child envisions as part of his or her identity has long-lasting effects.

Researchers have found that adults’ level of attachment can be predicted from their childhood (Hazan and Shaver, 1987; Fraley, 2002; George, Kaplan, & Main, 1984).

The researchers asked adults about their own childhood attachments with their parents and how it affected their development (George, Kaplan & Main, 1984). Their research was consistent with Bowlby and Ainsworth’s findings. They discovered adults also had three levels of attachment (secure, ambivalent and avoidant) patterns. Young adults were also studied in terms of attachment, with similar results (Kobak & Sceery, 1988).

Therefore, the types of bonds children have, and the effects these bonds created with their caregivers, appear to continue into their adult years.

Individuals who thought of killing themselves came from families with high generational conflict (Joffe, Offord & Boyle, 1988). Females were especially at risk for ideations if they felt devalued by their family. Children who came from families who moved often also reported more suicidal thoughts than those who lived in a more stable home environment (Potter, Kresnow & Powell, 2001).

Suicide and Attachment Theory. Attachment Theory (Bowlby, 1980) contends that a lack of connection to others, especially parental figures, can result in issues of

rejection and separation. A child who experiences feelings of separation can trigger a

suicide attempt in an effort to obtain attachment (Zimmerman, 1991; Zimmerman&

Zayas, 1995; Debold et al., 1993). For example, attachment theory (Bowlby, 1980)

contends that the potential separation from their parent, whether physical or emotional,

signifies a major break in attachment. Fear of abandonment can trigger an attempt in

hopes that the bond will be secured once again (Lyon, Benoit, O'Donnell, Getson, Silber

& Walsh, 2000).

A disturbance in the parent child relationship has been found to increase an

adolescent’s risk for suicide attempts (Sabbath, 1969). Lack of attention or mutuality can

decrease the parental bond. If a child is in competition for their parents’ attention, a

parent might become resentful of their child. This can continue until the parent has an

unconscious or conscious desire to rid themselves of their child. Failure to nurture or

provide consistent nurturing increased the risk of suicide attempts (Lyon et al, 2000). In

fact, Lyon et al, (2000) found that neglect more than physical or sexual abuse

distinguished those who had attempted suicide from those who had not attempted.

In a study focusing on the differences in bonding between mother and bonding to

father, significant relationships relating to the child’s self-esteem, depression and suicidal

attempts were found (McGarvey, Kryzhanovskaya, Koopman, Waite & Canterbury,

1998). Parents who were not affectionate had the highest rates of stress, while those with

most favorable bonding styles had the least distress. Fearful or preoccupied attachments

increased the likelihood of suicidal ideations (Lessard & Moretti, 1998). Children who had these weak bonds with their parents used more lethal means during their attempt.

This dissertation observed different forms of parental support to ascertain its applicability to Latinas. It also examined if self-reported support from peer networks played a role in suicide. A child who does not feel connected to important figures in her life might feel isolated from others, thus increasing the propensity for suicide. Parents or other attachment figures who fail to care for their children not only reported more neglect but also a mental health diagnosis (Gilbert, 1989).

Researchers have identified variables at various levels of analysis including individual(Turner et al, 2002; Razin, O’Dowd, Nathan, Goldfield, Martin, Goulet,

Scheftel, Mezan & Mosca, 1991), family (Zayas, 2006; Institutes of Medicine, [IOM],

2002; Zayas, 2000; Hovey & King, 1996; Guiao & Esparza, 1995; Zayas, 1987), neighborhood (Perez-Smith, Spirito & Boergers, 2002) and school (Turner et al, 2006;

Moskos, Olson, Halbern, Keller & Gray, 2005) contexts as predictors of adolescent suicide. These factors have consistently been shown to predict increased risk of suicide for all Latinas: acculturation status, age, sparse social networks and supports, drug use and mental health diagnosis (Zayas, 2006; Turner et al, 2002; Zayas, 1987)

Suicide Proclivity

Suicidal Ideations. Suicidal ideation is defined as a person’s serious consideration to take his or her own life (De Wilde, 2000). The CDC (2008) recently found in the general population that roughly 14% of adolescents questioned considered a suicide attempt in the past year. Females (19%) more than boys (10%) reported a

previous suicidal ideation in the last 12 months. Individuals who thought of killing

themselves came from families with high generational conflict (Joffe, Offord & Boyle,

1988). Family conflict has also been found to increase suicidality among Latina

adolescents (Zayas et al, 2000). Females were especially at risk for ideations if they felt

devalued by their family. Children who came from families who moved often also

reported more suicidal thoughts than those who lived in a more stable home environment

(Potter, Kresnow & Powell, 2001). Lifetime rates of ideation varied from 2% to 63%,

with a median value of 20%, and more common in females than males (Beautrais, 2002).

Suicidal thoughts increased during adolescence. In one study, females who were 15 years old reported 3.5 times as many ideations as boys of the same age (Fergusson,

Woodward & Horwood, 2000). Therefore according to the literature, conflicted interpersonal relationships, gender, emotional supports, behavioral health status and a

Latina’s cultural status predict suicidal ideations.

Suicide Attempts. A suicidal ideation always precedes an attempt, so many of the risk factors associated with a suicidal ideation mirror those for suicide attempts. The following is a discussion of suicide attempt data regarding the general population. The

National Institutes of Mental Health (NIMH; 2003) reported that between 30-40% of those who died by suicide made at least one previous attempt. Those who attempted suicide were 6-8 times more likely to try again compared to those who never attempted

(Pfeffer, Klerman, Hurt & Kakuma, 1993; Lewinsohn, Rohde, & Seeley, 1994). The probability of another attempt in the same year increased 100-fold compared to previous years (National Mental Health Association, 2006). The rates of teens reattempting ranged

from 6-15% per year (Brent, 1993). The greatest threat occurs within the first 3 months of

the initial attempt. Since mortality from suicide increases steadily through the teen years, identifying at-risk children can prevent future injuries and possible death (CDC, 1999;

Hoyert, Kochanek & Murphy, 1999).

Cultural Status

The debate continues on how best to define and measure cultural status. It is important to take into account differences that exist among Latinos pertaining to their individual and parents’ place of birth and language preference when examining critical behavioral health issues such as suicide. Cultural status is a controversial concept that has been defined in many ways. The point of contention occurs as to what defines one’s cultural status-a person’s place of birth or language use or a person’s values, beliefs, or traditions or a combination of the two. The term refers to one’s language, place of birth, citizenship, and strict family structure but has also included aspects of cultural values and traditions (Rogler, Cortes & Malgady, 1991; Zimmerman & Zayas, 1995; Tortolero &

Roberts, 2001). Efforts to measure this concept are equally diverse, including various combinations of objective and subjective criteria. Researchers agree that variation in the cultural status of Latinos is an important antecedent to a range of psychosocial and behavioral outcomes including suicide (Blake, Ledsky, Goodenow & O’Donnell, 2001;

Ebin, Sneed, Morisky, Rotheram-Borus, Magnusson & Malotte, 2001; Marin & Flores,

1994).

Cultural Status and Suicide. While Latina migrants tend to have lower rates of suicidal thoughts and behaviors, subsequent generations tend to be more at-risk (Pena et

al, 2008). One possible reason for this increase is that foreign-born Latinos engage in

more health-promoting behaviors (Ebin, Sneed, Morisky, Rotheram-Borus, Magnusson,

Malotte, 2001; Whaley & Francis, 2006; Unger, Ritt-Olson, Soto, Baezconde-Garbanati,

2007). These behaviors focus on the biopsychosocial aspects of a healthy lifestyle

including spirituality, family rituals, supportive emotional connections with relatives and

extended families. Collectivist homes with strong emotional bonds impact Latinos’

mental health, including suicidal ideations and attempts, depending on one’s level of

acculturation (Hovey, 1996; 1998; 2000).

Latina’s Cultural Status and Suicide. The longer a Latina resides in the U.S. the

higher her risk of suicide (Zayas, 1987, Sorenson & Golding, 1988; Swanson, Linskey,

Quintero-Salinas, Pumariega, & Holzer, 1992; Vega, Gil, Warheit, Apospori, &

Zimmerman, 1993; Zimmerman & Zayas, 1995; Vega, Kolody, Aguilar-Gaxiola,

Alderete, Catalano & Caraveo-Anduaga, 1998; Turner et al, 2006). Latina adolescents born in the United States are at a higher risk of suicide attempts than Latina immigrants

(Zayas, Kaplan, Turner, Romano & Gonzalez-Ramos, 2000; SAMHSA, 2003; Zayas,

Lester, Cabassa & Fortuna, 2005). This dissertation posits that the link between immigration status and suicidality suggests that Latina adolescent suicide may stem from different dynamics than suicide among other teens.

Parents’ Place of Birth. Children of immigrant parents face acculturative and migratory stressors increasing their risk for a variety of mental health conditions

(Szapocznik & Truss, 1978; Mindel, 1980; Landau, Griffiths and Mason, 1981; Vega et al., 1994; Boothby, 1994; Jablensky, 1994; Karno, 1994). Children’s mental health is

positively related to their parents’ mental health, especially their mother’s (Turner et al,

2006). The immigration experience can be filled with extreme anxiety. If families are immigrating illegally, the stress can be even worse. For example, due to traveling through difficult terrain, warlords, and border patrols, the level of uncertainty and safety can be debilitating to those crossing the U.S. border. Many Central American immigrants who fled from war-torn areas have witnessed various forms of violence including murders and rape (Cervantes, Salgado de Snyder, & Padilla, 1989). Salvadoran immigrants were found to suffer from Post-traumatic Stress Disorder (PTSD) at significantly higher rates within 5 years of their arrival into the U.S. (Plante, Manuel,

Menendez, and Marcotte, 1995). Instability, stress, and grief over the sudden loss of family and friends from their native country appear to occur after moving to a new country (Cervantes, Salgado de Snyder & Padilla, 1989; Irwin, 1991). Therefore immigrant parents might have a more difficult time dealing with acculturation, such as language acquisition, and their children might also be adversely affected by this transition.

When acculturating parents have poor coping skills, they can become less effective in caring for their children (Zayas & Pilat, 2008). In addition to childrearing responsibilities, these parents also must deal with their own separation from their native family members and create new support groups in the U.S. Building networks of social support was the most common response regarding stress by new immigrants (Mindel,

1980; Landau, Griffiths and Mason, 1981; Caplan, 2007). Healthy adjustment to a new country is related to the immigrant’s social relationships (Vega, Kolody & Valle, 1987).

Depression occurred when the individual believed moving to the U.S was a mistake and

experienced clashes between native and mainstream cultures (Vega, Kolody & Valle,

1987).

New immigrants face mental and emotional strains due to transitioning to a new culture but family can help with this stress (Choi, Meininger, J. & Roberts, 2006; Census,

2006a; 2006b; Bureau of Labor Statistics and Bureau of the Census, 2006). Selective

immigration, in which countries only allow certain immigrants based on skill level, and a

resilient family structure guarded individuals from depression (Delgado, 1995). The

longer a Latino immigrant resides in the U.S. and integrates however the more likely

these mental health protective factors will diminish (Vega, Kolody & Valle, 1987;

Abraido-Lanza, 1999). For example, Mexican-born immigrants have repeatedly reported

lower rates of mental illness compared to Mexican-Americans (Burnam, Hough, Karno,

Escobar & Telles, 1987; Escobar, Hoyos & Gara, 2000; Tortolero & Roberts, 2001;

Vega, Sribney, Aguilar-Gaxiola & Kolody, 2004).

It is projected that one out of every five Americans will receive a diagnosis of

depression in the U.S in their lifetime (US Department of Health and Human Services

[DHHS], 2000; Caplan, 2007). Latinos born in the U.S were at an increased risk of

depression, drug dependence and abuse when compared to native-born Latinos (Burnam,

Hough, Karno, Escobar & Telles, 1987; Escobar, Hoyos & Gara, 2000). Given that

depression can be debilitating, rendering a person incapable of performing routine

responsibilities, parental duties can easily be pushed aside.

Child’s Place of Birth. Although minority adolescents are prone to engaging in

dangerous behaviors (CDC, 2008), Latino children face the additional risk factors of

language proficiency and fewer social supports (Rogler, Malgady & Rodriguez, 1989).

Family is important to a Latino child’s functioning and a considerable amount of their

emotional support (Falicov, 1996). The longer Latino immigrants, adults or children,

reside in the US, the more the supportive function of the family structure weakens. For

example, the traditional Latino family incorporates a collectivist structure which is

different than the majority’s value of a more individualistic family structure. As the

immigrant acculturates, the need to adopt more mainstream values can reduce the collectivist paradigm commonly found in Latin culture (MCHB, 1999; Delgado, 1995).

When this happens, the extended family is expected to care for the children but networks might decrease in availability upon moving to a new country. Immigrant parents might choose an adult they are less familiar with for child care possibly exposing them to individuals engaged in risky behaviors (Garcia-Preto, 1996).

English Proficiency in Home. According to the U.S Census Bureau (2004a), 9.2 million Latino children spoke a language other than English in their home. Of the 12.1 million Latino children in the U.S, almost half (5.8 million) conversed in English “less than very well” (Census, 1999, p. 53). Over three-fourths (78%) of Latinos age 5 and older spoke Spanish in their home to maintain cultural ties and communicate with older, immigrant family members (Census, 2008).

Children’s English proficiency level impacts not only their educational attainment but also the quantity, quality and characteristics of their social networks. A Latino’s lack

of English skills can predict stress, depression and suicidal ideation (Hovey, 2000).

Young Latinas typically learn English more quickly than their older family members partially due to their involvement in English as a Second Language (ESL) classes at school. Due to exposure to English, opportunities to associate with a diverse group of people increase. The heterogeneity of her interpersonal contacts could be very different than the heterogeneity of her family members’ contacts. Since a child’s primary language with her friends might be English but Spanish with her family, there might be a disconnect between family and friends. Consequently, her family might not be able to monitor her peer groups as well as parents who speak English can (Allen, Elliot, Fuligni,

Morales, Hambarsoomian, & Schuster, 2008). Alternatively, her friends might find her traditional culture and family difficult to understand and appreciate. Dissonance between her friends and family might occur. This tension might increase her sense of hopelessness due to the pressure of having to ultimately reject one important support group for the other.

Behavioral Health Status

Over 90% of children who commit suicide experience some type of mental health disorder (Shaffer & Craft, 1999). Depression, substance abuse, hopelessness and suicide attempts intensify during adolescence (Rotheram-Borus, Trautman, Dopkins, & Shrout,

1990; Campbell, Milling, Laughlin, & Bush, 1993; Kovacs, Goldstein, & Gatsonis, 1993;

Wagner, Cole, & Schwarzman, 1995; CDC, 1999, Hoyert, Kochanek & Murphy, 1999;

SAMHSA, 2000, Kessler, Berglund, Borges, Nick and Wang, 2005, Spirito & Esposito-

Smythers, 2006, AAS, 2007). Distinct links between major depression, prior suicide

attempts and substance abuse increase one’s risk of future suicide attempts (Shaffer &

Gould, 1987; Brent, Perper, Moritz, Allman, Friend, Roth, Schweers, Balach, Baugher,

1993).

Gould and colleagues (1998) wanted to ascertain how specific diagnoses impacted suicidal ideation and attempts. Their study investigated 1,285 randomly-selected children ranging from 9-17 years of age. Forty-two of the participants had attempted suicide while 67 had a history of ideations. Having a mood, anxiety and/or drug use disorder increased one’s risk of ideations and attempts after controlling for demographic

characteristics. Children with a mood or anxiety disorder exhibited three times more

suicidal behaviors compared to those without a diagnosis.

Depression. During adolescence, the incidence of depression and suicide appears

to peak (SAMHSA, 2000). Between 10-15% of children suffer one or more symptoms of

depression and roughly 5% of 9-17 year olds receive a diagnosis of major depressive

disorder (Smucker, Craighead, Craighead & Green, 1986; Shaffer et al, 1996).

Depression increases a female’s incidence of suicide 12-fold and a previous suicide attempt can raise their risk 3-fold (Shaffer et al., 1996). After age 15, depression rates double for females when compared to males (Weissman & Klerman, 1977; McGee,

Feehan, Williams, Partridge, Silva & Kelly, 1990; Linehan, Heard & Armstrong 1993).

In 2004, nearly 3.5 million youth suffered at least one major depressive episode (MDE) in their lives (SAMHSA, 2006b) and 7% or 1.8 million teens with a MDE ideated. Of

that group, 712,000 teens attempted suicide during their depressive episode (SAMHSA,

2006b). For those diagnosed with MDE who also engaged in alcohol or drug abuse, the likelihood of a suicide attempt or ideation rose (Kessler et al, 2005).

Depression and Culture. Culture affects the way individuals define and process mental health issues including suicide (Giddens, 1964). Although linked with pathology, suicide can be a response to acculturative stress (Institute of Medicine [IOM], 2002).

Given the rates of suicide among young Latinas, acculturation models need to include the constructs of depression and suicidality (Hovey & King, 1997). These additional antecedents take into account how change in cultural status can result in the disruption of social and family support networks.

Acculturation, or adopting mainstream values, beliefs or traditions consistently appears to elevate the risk of depression and suicidal risk among Latinos (Swanson,

Linskey, Quintero-Salinas, Pumariega & Holzer, 1992; Vega, Gil, Zimmerman &

Warheit, 1993; Hovey, 1996; Hovey & King, 1997; Rasmussen, Negy, Calrlson & Burns,

1997; Hovey, 1998; 2000; Oquendo, Lizardi, Greenwald, Weissman & Mann, 2004). For example, Mexican-American adolescents had 1.8 times the ideations as European-

American children (Tortolero & Roberts, 2001). High levels of familial conflict and negative beliefs about the future predicted stress. Among Latina adolescents, acculturation status and family problems were found to be risk factors for depression and suicidal ideations (Olvera, 2001; Roberts & Chen, 2005).

Substance Use. Substance use raises one’s propensity for risky behaviors.

Several studies indicate the long-term affects of illicit drug use including mental illness and suicidal proclivity (Greenblat, 1998; Lynskey, 2004; Duberstein, Conwell & Caine,

1993, Brent, 1995). A 16 year examination of individuals with no depressive symptoms at baseline, who then began using substances, reported escalated rates of depression. The frequency of depression increased four times from the initial to the follow-up interview

(Bovasso, 2001). Fergusson’s (2002) work, spanning 21 years, demonstrated that habitual marijuana use was linked with depression, suicidal ideations and attempts.

Brook (2002) also found that, over a 14 year period, illicit drug use predicted a future diagnosis of major depressive disorder which is closely linked with suicidal thoughts and behaviors.

Substance Abuse and Suicide. Individuals with a mental health diagnosis who abuse substances showed a higher propensity for suicide compared to those without a diagnosis or drug use (AAS, 2007; Borowsky, Ireland & Resnick, 2001, American

Psychiatric Association, 2005). Roughly 50% of those with a mental health condition who committed suicide were found to be intoxicated at the time of their death

(SAMHSA, 2006a; American Psychiatric Association, 2005). Children using alcohol or other drugs at the time of suicide attempt also required longer hospital stays (SAMHSA,

2006a).

Adolescents might self-medicate with alcohol and other drugs while others might communicate their stress through suicide (APA, 2005). Alcohol, readily available to many adolescents, is the most frequently used drug during the time of a suicide completion. Roughly 50% of the adolescents who committed suicide under the influence of substances were found to be intoxicated at their moment of death while 66% had at least a trace of alcohol in their system (SAMHSA, 2006a).

Latinos and Substance Use. Distinct gender differences between Latino and

Latina adolescents include their level of drug use. Latino males typically use inhalants,

alcohol, marijuana and PCP while Latinas typically use tranquilizers and heroin

(Hunsaker, 1985). Acculturated Latino youth were nearly 13 times as likely to report

using illegal drugs compared to less-acculturated Latinos (National Drug Control Policy,

2007). A Department of Health and Human Services report on rates of Latino substance

use indicates that Latinos begin using drugs earlier than either European-Americans or

African-Americans:

rates of crack and ecstasy use in the 12th grade, and their level of heroin

use is equivalent to that of whites. But in the 8th grade they tend to come

out highest of the three racial/ethnic groups on nearly all classes of drugs,

including alcohol (amphetamines being the major exception (Department

of Health and Human Services [DHHS], 2001, p. 9).

The relationship between suicide and substance use has been studied among

Latino teens. Queralt (1993a) studied 14 Latino students who completed suicide.

Autopsy records found that 25% of them had alcohol and 15% had other drugs in their

system (Queralt, 1993a). Given Latinos’ use of illicit substances, how these rates increase with age, and that the rates are higher than that for other adolescents until 12th

grade, understanding the link between drug use and this group is essential.

Social Supports

Individuals with close relationships handle stress more effectively and show gains

in physical and mental health compared to those who lack social contacts (IOM, 2001).

Social supports yield various protective factors for depression and they can decrease the acuteness of a mood disorder in high risk groups such as immigrants (Shen & Takeuchi,

2001). These networks guard against mental health issues via emotional support

(O’Grady and Metz, 1987; Vance et al, 1998) and can moderate suicidal ideation and attempts with different racial/ethnic groups (Borowsky et al, 1999; Kaslow et al, 1998;

Kotler et al, 2001; Nisbet, 1996; Rubenstein et al, 1989; Thompson et al, 2000) specifically with Latinos (Hovey, 1999). Completed suicides increased when individuals reported fewer opportunities for family or peer emotional support (Allebeck et al, 1988;

Drake et al, 1986). A study focusing on adolescents found that half of the variance in suicidal risk related to one’s perceived level of social support (D’Attilio et al, 1992). As a result, efforts should include increasing adolescents’ social supports in their suicide

prevention programs (Borowsky et al, 1999; Vance et al, 1998).

Social Supports and Cultural Status. Adolescents who reported more friends were less likely to attempt suicide. This finding reflects the idea that decreased support systems affect one’s mental health (O’Donnell, Stueve, Wardlaw and O’Donnell, 2003,

Bearman & Moody, 2004; AAS, 2006). Opportunities to obtain these interpersonal contacts can be difficult for immigrants, especially those outside of their ethnic enclaves

(Oetting, 1993). When immigrants arrive in the U.S., they can feel perpetually displaced and isolated due to the lack of cultural reinforcement from other Latinos. In fact, one’s nativity may be a less significant risk factor for psychiatric morbidity than originally thought and actually how one integrates into the US might be more important (Alegria,

Shrout, Woo, Guarnaccia, Sribney, Vila, Polo, Cao, Mulvaney-Day, Torres & Canino,

2007).

Relationships can buffer the incidence of maladaptive behaviors (Griffith, 1984).

A study of Caucasian and Mexican American adults indicated that a Latino’s cultural

status impacted social supports. Mexican Americans who spoke English had

significantly more social networking opportunities, cumulative contacts and reciprocity

when compared to Spanish-speaking Mexican Americans. Latinos who were proficient

in English had more friends resulting in more opportunities for emotional support.

Although English proficiency increased social networks in Griffith’s (1984) study

among adult Latinos, adolescent friendships are different. Most of the Mexican

Americans in the study were married and reported fewer friends than the non-Latino

adults in the study. Marriage tends to change friendships in-that the married partner is

typically the focus of one’s social networks (Milardo, 1982). The partner also replaces

other friends resulting in fewer but more intimate relationships compared to young, single

adults. The adults in the study also learned the social nuances of making friends in the dominant culture, where young Latinas might not be as sophisticated in their social skills.

Therefore, it is difficult to generalize from these results to Latina adolescents. This study will examine if the practice of speaking English in the home, among other cultural factors, protects Latina teens as it appeared to protect Latino adults.

Little research exists on how social networks impact suicide among youth, especially young Latinas (Bearman & Moody, 2004). Bearman and Moody’s study found that having a friend who had died from suicide increased the odds that an

adolescent would ideate or attempt suicide themselves. Female adolescents with few friends, or did not have mutual friends in their social network, reported more ideations.

The authors concluded friendship networks impacted suicide rates. While having a larger network of friends might be associated with reduced risk of suicidal thoughts and behaviors among Latina adolescents, such friendships might also pose unique challenges.

In particular, the composition and complexity of their friendship networks might be as important, or more important, than the number of friends in their network per se.

Social support appears to buffer the adverse effects of acculturation and suicide

among Latinos, although research has not examined this with Latina adolescents (Hovey,

2000). A female who primarily speaks Spanish might have fewer friends, but more

emotional support from her family increasing her protective factors. A female who aligns more with the mainstream culture might have more individuals in her friendship network, but these friends could pull her to risky behaviors (Prinstein, Boergers & Spirito, 2001).

This could be partially due to her parents’ lack of knowledge of U.S. customs and culture and ability to monitor their children effectively due to their limited English proficiency.

Ultimately, these networks can become a risk factor in terms of the composition and characteristics of their friendship group. Understanding the specific factors associated with Latina adolescents such as social supports and mental health is essential to prevent suicide among this group (Hovey, 1999).

Latinas and Adult Social Supports. O’Donnell, Stueve, Wardlaw and O’Donnell

(2003) studied 879 urban adolescents’ reported levels of social support and their

relationship with suicidal ideations and attempts. Over one-quarter (28%) of the Latinos

in the sample reported an ideation and 17% had attempted suicide. The attempt and

ideation rates were higher than the African American and White participants. Half of the

participants who attempted suicide spoke with an adult about their problems. Those who

attempted stated they would not go to an adult for help in the future. Therefore we

suggest linking children, especially those with a history of suicide attempts, with an adult mentor might not prevent an attempt.

Traditional Latino culture stresses respect or respeto for elders, whether it is one’s parent, teacher, extended family or neighbors. This practice might encourage Latino adolescents to value an adult’s advice (Trepper, Nelson, McCollum, McAvoy, 1997). If a

Latina has close bonds to her parents or older siblings, she might feel more comfortable speaking to them than someone outside of her family. A Latina’s respect for her elders

might encourage her to believe their opinions are more valuable than those of her peers.

Therefore, adult social supports might be more important to Latina adolescents than to her peers.

Teacher Support. A Latina’s perceptions of her acceptance at school affect her impression of available social supports. Connections to teachers and classmates are

important to consider since children are typically in school more often than at home each

day. Latinas who felt unaccepted by their peers and faculty reported less attachment to

their schools (Turner et al, 2006). This was true, for instance, if Latinas believed their mothers were not welcome in their school due to their limited English ability. If a Latina does not feel connected to her school community, she might become increasingly isolated and less likely to seek support in times of distress.

Friends Support. Gender differences occur in the type of friendships boys and

females initiate. Female routinely engage in close relationships involving one or two people in a primary support group that is also loosely associated with a larger group

(Savin-Williams, 1980). Young females also experience heightened pressure of social involvement than boys (Brown, 1982). Therefore, if a female is having difficulty making friends, the effects prove more detrimental for her than for a boy in terms of her ability to share secrets and personal concerns (Diamond & Muntz, 1967; Bigelow & LaGaipa,

1980, Blyth & Foster-Clark, 1987). Higher levels of intimacy were reported more often in same gender friendship networks than with mixed gender groups (Sharbany, Gershoni &

Hofman, 1981; Hunter & Youniss, 1982; Reis, Senchak & Solomon, 1985).

Gender-based suicide rates could be attributed to how females are socialized to value their friendships. This can result in females’ heightened vulnerability to a relationship loss compared to boys (Allgood-Merten, Lewinsohn & Hops, 1990).

Tenuous connections or the termination of a relationship can lead to isolation and feelings of hopelessness. Isolated teen females or those with weak friendship networks are at a higher risk of suicidal ideation (Bearman & Moody, 2004). These feelings of loss threatened females more than males given the intensity of the emotional exchange females tend to have with their friends compared to boys.

Connections to others have also been examined in racial and ethnic samples. A convenience sample of 95 Latino immigrants perceived depression as an issue with interpersonal and social factors (Cabassa, Lester & Zayas, 2007). These immigrants reported that “Depression represented ‘being in a Labyrinth’ of interpersonal problems

and economic strains, and lacking the emotional support system to help them cope with

these difficulties in their everyday life” (p. 12). Sin apoyo (lack of support) refers to not

experiencing strong connections to individuals, feelings of isolation and little emotional

support to deal with one’s problems. Lacking close and committed relationships

positively added to one’s propensity for depression, especially in females. Furthermore, if

individuals did not believe they could trust or confide in their friends, depression rates

further increased. Isolation and interpersonal conflicts can also lead to the breakdown of

the cultural norm of personalismo, where relationships are considered imperative to effective functioning. Therefore, Latinas lacking social supports could have a higher risk of suicidal ideation and attempts compared to those with access to friends.

Reciprocated Friendships. Reciprocity occurs when individuals agree upon the

level of commitment they have to one another (Youniss, 1980). For example, if a child

states that “X” is their best friend, “X” would nominate that child as their best friend.

Children with many friend nominations had longer-lasting relationships than those with fewer nominations. Adolescents with higher reciprocity scores also ranked higher on

social competency (Asher, 1983).

Friendship mutuality can affect suicidality. Bearman and Moody (2004) discuss the distinction between transitive and intransitive relationships. Intransitive friendship networks occur when a person’s friends are friends with each other but do not nominate one another as a member of their respective network. Those who belong in intransitive networks typically maintain weak connections with others and are more prone to engage in risky behaviors including suicide.

Individuals possessing close relationships with others have more opportunities to self-disclose personal information in a safe place with minimal judgment (Vaux, 1985).

These friends also provide consistent companionship during difficult events. These intimate bonds also allow teens to share coping mechanisms that could protect them from unhealthy responses to stress. For example, African American adolescents in low- income families who were involved in reciprocated friendships had higher rates of self- competence (Cauce, 1986). Reciprocated relationships tend to endure longer than casual acquaintances and can become a consistent source of support.

Durkheim (1951) studied the phenomenon of suicide by observing an individual’s risk in relation to their level of social integration. He found that those who had died from suicide had very few, if any, social ties to others. This lack of social connection creates an inadequate support system for individuals (Lin, Simeone, Ensel & Kuo, 1979). As a

Latina transitions from her traditional ethnic norms to U.S. culture, she might have difficulty understanding the social differences in terms of social cues. Friendships are unique due to the manner in which they benefit one’s personality development. These interactions are distinctively different than those found in the parent-child relationship

(Sullivan, 1953). Friends depend upon each other for advice and validation (Smollar &

Youniss, 1983; Selman & Shultz, 1990). Peer interaction enables adolescents to share and clarify their feelings and thoughts (Smollar & Youniss, 1983; Selman & Shultz,

1990). As friendship becomes an “interactive phenomenon” throughout a child’s life, understanding which types of friendship networks are more prone to suicide proclivity is imperative (Youniss & Haynie, 1992, p. 61).

Networks also provide social integration. Popular children interact with more individuals and this gives them more opportunities for interpersonal contact. Teens rejected by their peers exhibited higher rates of inappropriate behavior such as being physically and verbally aggressive compared to well-liked individuals (Coie &

Kupersmidt, 1983; Dodge, 1983; Rubin, 1983; Reisman, 1985; Ginsberg, Gottman &

Parker, 1986). Teens also spend more time with their peers (Larson, 1983) and their acceptance is critical to a child’s mental health (O’Brien & Bierman, 1988). As a child matures, friends become more influential in terms of their impressions and decision making (Douvan & Adelson, 1966).

Although other racial and ethnic minorities encounter majority-minority issues, the rates of suicide among Latina adolescents are far higher when compared to other groups. Factors such as family dissonance and lack of mutuality have been theorized as possible sources of strain for Latina adolescents (Trautman, 1961a; 1961b, Zayas, 1987;

Queralt, 1993a; 1993b, O’Donnell, O’Donnell, Wardlaw & Stueve, 2004; Turner,

Kaplan, Zayas & Ross, 2002). As a child acculturates, her friends might play a more important role in her likelihood to ideate and attempt suicide.

Frequency of Friend Contacts. When a person’s group of friends are not accessible, negative health effects including depression can occur. Close friendships involve more frequent contact. Suicide prevention programs now target friends as gatekeepers due to their ability to observe changes in their peer’s behavior and mood

(Moskos, Olson, Halbern, Keller & Gray, 2005). Since teens arguably spend more time

with their friends than family (Larson, 1983), especially during school, they can offer

emotional support more so than perhaps their family members.

Building from the results of a 2000 study (Zayas et al, 2000); Turner and

colleagues (2006) examined how community factors affected the attempt rates of Latinas.

Females who participated in after-school activities reported more often to think

optimistically about their future and had supportive relationships with their mothers.

These females shared feelings more comfortably and were supported in school due to

their involvement with adults. These mentoring relationships enhanced Latinas’

resiliency and motivation in school and beyond. For less acculturated Latinas, especially

first-generation with household responsibilities, options for garnering these social

supports in school were limited (Zayas, 1987, Razin et al, 1991, Zayas et al, 2000).

Therefore, Latinas might have fewer opportunities to create and maintain social contacts

compared to their peers. Since a lack of social connection is a risk factor for suicide,

children not allowed to attend after-school programs and make friends could be more at

risk for suicidal thoughts and behaviors. These studies indicate that acculturation does

play a role in how Latinas access their peer supports, but the results are not definitive

how acculturation protects or increases their likelihood of suicide proclivity. Our study

takes their research farther by examining the types of friendships these Latinas have and

how that may play a more active role in their risk for ideations and attempts.

Adult mentorship increases in importance as a Latina acculturates due to her

mother’s lack of ability to provide guidance from her own past experiences (Zayas et al,

2000). Zayas and colleagues (2000, p.58) stated that an “effective mentor” is dependable,

encouraging and a positive role model (Rhodes, Contreras & Mandelsdorf, 1994). As a

Latina continues to turn to people from the majority culture, tension might swell at home.

As a result, a Latina might attempt to find guidance from individuals at their school.

Females with a consistent level of stress from their mothers can become less resilient and less able to handle problems (Werner & Smith, 1992; Debold, Wilson, & Malave, 1993).

Suicidal Friends

Friends have a noticeable impact on suicide as evidenced by suicide clusters.

Adolescents appear to be most at-risk for suicide contagion compared to other age groups, although this is a low-base rate phenomenon (Gould, Wallenstein, Kleinman,

1990). There is some evidence indicating peers’ risk behaviors are causal (Kandel, 1978;

Keenan, Loeber, Zhang, Stouthamer-Loeber & Van Kammen, 1995), overtly influencing the adolescent to exhibit such behaviors (Bandura, 1973). Teens tend to self-select into a cluster of friends predisposed to suicidal thoughts and behaviors (Joiner, 2002). When a member of a friendship network attempted or completed a suicide, the remaining peers were more likely to mimic those same behaviors. These friends also share modalities or methods of how to attempt suicide (Joiner, 2002).

Therefore, the types of friends one socializes with appear to influence one’s likelihood of suicide (Jellinek & Snyder, 1998). Bearman and Moody (2004) studied the connection between friendship, suicidal ideation and attempts. Teens with a friend who had a suicidal history were more likely to ideate or attempt themselves. Those with fewer social networking opportunities or who had weak friendship networks were also more likely to ideate (Roberts & Chen, 1995, Bearman & Moody, 2004).

Traditional Latino Family Structure

Traditional Latino families utilize an authoritative approach to parenting (Turner et al, 2002). Households are organized by a hierarchical system with regard to age and gender, where males and elders hold the most respect. The husband makes most of the family decisions, at times with little or no consultation from his wife (Garrison, Roy &

Azar, 1999). This restrictive environment can hinder some children’s ability to cope with problems and decrease a female’s self esteem (Dornbusch, Ritter, Leiderman, Roberts &

Fraleigh, 1987). When teenage daughters challenge the prescribed gender roles supported by their parents, the family structure can be threatened. The turmoil that ensues from Latinas confronting these expectations leads to depression and an increase in suicidal risk. Zayas and colleagues (2000) found that Latinas who attempted suicide came from homes with higher generational conflict compared to households without a reported attempt.

Latino families can also follow a more authoritarian approach to rearing their children. Arzubiaga, Ceja & Artiles, 2000). While authoritative parents are both demanding and responsive, an authoritarian style is best described by a demanding yet responsive approach to behavior modification. Parents do monitor their children and provide their children with clear standards for the actions. These parents are assertive but not intrusive or restrictive. The goal of raising individuals in this supportive environment is that children become self-regulated, assertive and cooperative to those in-and-out of the family unit. Although much of the research focuses on more stringent parenting styles such as the authoritiative approach, some studies have also found that the authoritarian

style to parenting decreases a child’s likelihood of suicidal ideation (Martin & Waite,

1994). The authoritarian style indicated the child had more effective coping skills, higher self esteem and was more comfortable talking to their parents in times of crisis.

As a Latina individuates from her family her rate of suicide attempts increase

(Turner et al, 2002; Razin, O’Dowd, Nathan, Goldfield, Martin, Goulet, Scheftel, Mezan

& Mosca, 1991). Individuation occurs when a child separates themselves from their

parents which can cause familial strain. This process can create tension between

generations which has been found to increase a Latina’s risk of suicide (Zayas et al,

2000). In one study, Latina adolescents believed attempting simultaneously protected

and relieved their mother from the stress they created by disrupting the family unit

(Zimmerman, 1991; Zimmerman& Zayas, 1995; Debold et al., 1993). Given the

collectivist characteristics of Latino families, namely the importance of the family unit

over the individual, a daughter might rationalize that her death would help the family

function more efficiently.

If a Latina senses tension between her family’s Latino culture and her friends’

U.S. culture, the strength and quality of her relationship with her relatives, especially

with her mother, can weaken. This dissonance especially occurs when a Latina seeks

alternative forms of support and acceptance. This pull between cultures can intensify with

peer pressure. For example if familial expectations include returning home after school

to care for her family, her ability to go out with friends can be affected. If a Latina

decides to forgo these responsibilities, her family might be adversely affected. As a result, feelings of shame due to relinquishing her Latino culture for acceptance by the

majority can arise (Brindis, Wolfe, McCarter, Ball & Starbuck-Morales, 1995; Hovey &

King, 1996).

Family Support. Latina teens experience guilt for feeling the need to choose

between their family and peer group (Hovey & King, 1996). Such a choice contradicts

the Latino cultural norm of collectivismo. The foundation of a collectivist home places

the group higher than the individual. The strong bond between family members creates a

high level of emotional connectedness (Delgado, 1995). Incongruity ensues between the

cooperative focus in the family compared to the predominantly individualistic U.S.

culture. This difference can challenge new immigrants to comprehend and assimilate,

especially among adolescents.

Family structure can offer protection from suicide (Institutes of Medicine, [IOM],

2002). Those living alone, for example, indicate an increased likelihood for suicide

(Allebeck et al, 1988; Drake et al, 1986). A child’s perception of his or her family’s

connectedness also protects him or her from suicide (Borowsky et al, 2001; Resnick et al,

1997) especially among Latino teens (Guiao & Esparza, 1995). Often regarded as the unused resource by clinicians and researchers, nuclear and extended families not only provide support for Latinos, but an opportunity for cultural exchange (Maternal and Child

Health Bureau [MCHB], 1999; Espinosa, 1998). The collectivist structure of Latino culture serves as a protective factor from risky behaviors and physical disease (MCHB,

1999).

Latino families average 3.5 people per household compared to 2.6 members in all other American households (MCHB, 1999). The greater number of individuals in the

home can provide more opportunities for support. Although a larger number of family members to turn to can protect one from suicide, controlled studies indicate that a low level of communication between parents and children acts as a significant risk factor for suicide (Gould, Fisher, Parides, Flory & Shaffer, 1996)

Latina adolescents with immigrant parents experience pressure to adopt adult roles before they are mature enough to assume those responsibilities (Zayas, 2000).

These females turn into substitute mothers resulting in a loss of identity compared to their peers (Zayas, 2006; 1987). Due to many Latinas feeling culturally sandwiched between two cultures, feelings of hopelessness arise (Smokowsi, Buchanan & Bacallao, 2009).

Incorporating native and mainstream traditions and values can be simultaneously stressful and liberating. With increasing tensions between relatives and young Latinas, family support will likely to decrease (Zayas, 2006; Zimmerman & Zayas, 1995).

When a Latina daughter tries to individuate, the level of emotional support from her mother might decrease. Mutuality has been defined as thoughts and communication where support and emotions were shared in a safe and secure environment (Genero,

Miller, Surrey & Baldwin, 1992; Jordan, 1998). Mothers who showed more interest and empathy in their children’s lives appeared to foster decreased depressive symptoms and ultimately suicide with their daughters (Turner et al, 2006; Medina & Luna, 2006).

Teen females strive to be connected and validated by their mothers. As Latina mothers typically do not work outside the home, their likelihood of confronting their daughters on issues increase as does the strain in the relationship (Zayas et al, 2000;

Zimmerman & Zayas, 1995). The family structure does appear to change with each

generation as evidenced by increased rates of single motherhood in second and third generation Latinas (Census, 2000). Regardless of the household composition, the strain typically resulted from the Latina daughter’s desire for independence. Latino parents

viewed her desire for individuation as self-destructive to the family unit and feared drug

use and sexual activity among other concerns (Zayas, 2006; Zimmerman & Zayas, 1995).

Overall, associations between Latina suicide and connections with family are

apparent in the literature (Zayas, 1987, Queralt, 1993a; 1993b, O’Donnell, O’Donnell,

Wardlaw & Stueve, 2004, Turner et al, 2006). Trautman (1961a; 1961b) studied Puerto

Rican women with a history of suicide living in New York City. Moving to a new

geographic location made these Latinas more susceptible to suicide. The women had no

mental health diagnoses or symptoms but relationship conflicts were found to precede

their attempt. These clashes typically happened with their mothers or spouses, arguably

the main support system in their lives. Controlling for mental health status is important,

but it appears the types of social supports Latinas have in their lives impacted their

suicide rates (Zayas & Dyche, 1995; Zayas et al, 2005).

Summary

The literature suggests a Latina’s risk for suicide is associated with her cultural status. The longer a Latina resides in the U.S, the higher her rates of depression and suicidal thoughts and behaviors. In addition, third generation Latinas are more likely to ideate and attempt suicide when compared to immigrant children. These rates of suicide are attributed to the higher incidence of depression among Latinos born in the U.S compared to those born in their country of origin. One’s cultural status can affect the

quality and quantity of accessible social contacts. These relationships support an individual in times of distress. Individuals lacking social networks and supports could be more susceptible to suicide.

This dissertation addresses a missing piece of the literature-how a Latina’s friends and adults outside her family act as a protective factor for suicidal ideations and attempts.

The study is the only known examination of the characteristics of a Latina’s friendship network regarding suicide proclivity. This investigation could be a critical addition to the research of Latina adolescent suicide and their heightened risk compared to other teens.

Arguably, friends influence teens at times more than their families. In this study,

I will focus on Latinas’ parents, friendship networks and suicide rates. Primarily, I will analyze the quality of these friendships and whether particular constellations of friends protect Latinas from suicide or actually increase risk. The project will expand suicide theories and test the mechanisms connecting race/ethnicity, family and peer support to suicidality among Latina adolescents.

The study utilizes a nationally representative data set which includes generational status of the child and extensive data on family and friendship networks. These meso- level variables could be critical to suicide prevention among Latinas. Latinas are different from most teens in-that they are navigating between two cultures (family and friends) from varying acculturation levels. Acculturation alters these networks and imposes strains on families. As a result, friendships are essential to provide necessary support.

But, as Latinas acculturates and learn European-American social norms and mores, they may create new and unfamiliar ties with her friends that are possibly unreciprocated.

Suicide is a public health crisis (U.S. Public Health Service, 1999, IOM, 2002,

Suicide Prevention Resource Center, 2005, World Health Organization, 2008). The U.S.

Latino population is young and Latinas in particular are at risk for suicidal behaviors.

There appears to be a strong link between a Latina’s acculturation level and suicidality,

but the literature is limited in terms of an explanation. Currently empirical research has

determined those at the highest risk of ideation and attempts are females, have weak

interpersonal relationships, suffer from substance abuse and/or depression and are

acculturating to the U.S. Current research limits our understanding of the specific reasons

why Latinas attempt at such high rates. Since acculturation is linked to a person’s social

networks, examining a teen’s friendship network is vital to prevent Latina adolescent

suicide.

CHAPTER 3:

METHODS

By examining the types of relationships Latinas choose, researchers can enhance current strategies to reduce the high rates of suicide among adolescents. If specific risk and protective factors can be identified in children transitioning to the U.S., culturally competent prevention efforts can be instituted. As a result, this dissertation will advance our understanding of the specific factors associated with suicide with young Latinas.

Description of the Add Health Dataset

The data used in the study were from the National Longitudinal Study of

Adolescent Health (Add Health). Add Health is a nationally representative study of

various health-related behaviors of adolescents in the United States. Add Health data files

contained names, personal identifiers, or geocode information. The total number of

respondents in Wave I, in-home interviews was 20,745. A select number of variables

were also taken from the in-school questionnaire and the network questionnaire. A

subsample was created solely of Latina (females of Hispanic/Latino decent) that totaled

1,618 participants. The in-school and social network questionnaires, along with a

discussion of the data collection, are described later in this chapter.

Complex Survey Design. Add Health incorporated a stratified cluster design and

also over-sampled to incorporate a complex design. The complex design incorporates

disproportionate sampling as a result of the probability proportionate to the size sampling method. If this is not addressed during analysis, it may result in biased parameter and variance estimates (Chantala, 2001; Chantala & Tabor, 1999). Conducting any analysis without addressing the homogeneity from the dependent observations and the over- sampling of specific groups, such as disabled children and various ethnic groups, these biased results could lead to incorrectly rejecting a model when not appropriate (Stapleton,

2006). The complex design can hinder the ability to make conclusions about the population it is attempting to characterize. The grand sample weight accounts for the disproportionate sampling while the strata and cluster variables guarantee accurate variance estimates (Stapleton, 2006). To ensure that the results depicted the population,

STATA 10.0 (StataCorp, 2007) was used because of its ability to take into account the weights, primary sampling units (PSU) and strata variables in Add Health (Chantala,

2001). Weights were created for each wave of data while the PSU and strata variables were consistent no matter which wave was used. The PSU signified which school the respondent attended and the strata variable indicated which geographic region the respondent resided in at the time they answered the questionnaire (Chantala & Tabor,

1999). Any case that had a missing sample weight (grand sample weight for Wave 1) was eliminated from the working dataset by this researcher which was advised by Add

Health (Chantala & Tabor, 1999).

Add Health was particularly well-suited for this study due to data collected on

Latinas’ and their parents’ places of birth, primary language spoken at home and peer

social networks. Information on race and ethnicity, suicidal proclivity (ideations and

attempts) and a wide range of background characteristics were also available.

Data Collection. Add Health utilized a school-based stratified random cluster

sample of all high schools in the United States. An eligible school included an 11th grade

with an enrollment of 30 students or more. Feeder schools that sent graduates to an

identified high school and included a 7th grade were recruited as well. Schools were stratified into 80 clusters based on geographic region (Northeast, South, Midwest and

West), metropolitan status (urban, rural, suburban), population of school (125 or less,

125-350, 351-775, and 776 and more students), type of school (public, private, parochial), percent Caucasian (0, 1-66, 67-93, 94-100), percent African American (0, 1-6,

7-33, 34-100), grade span (K-12, 7-12, 9-12, 10-12), curriculum (general,

vocational/technical, alternative and special education). A total of 146 schools were recruited.

The manner in which data is gathered can result in measurement error. This can happen in many stages of the collection process including how the interviewer asks a question, how the participant answers a particular question or data management. Some measurement error issues were alleviated by training Add Health researchers on a number of interviewer-related errors. In addition questionnaires were answered either as questionnaires or through a hand-held device for more sensitive information. If the surveys were done via in-person interviews, there is the potential for different responses

(Safir, Scheuren & Wang, 2000).

The initial phase of Wave I data collection occurred in 1994-1995 through in-

school questionnaires administered to a nationally representative sample of 7-12th grade students. The second phase of Wave I included follow-up interviews in their respective homes (27,000) from a core sample of each school plus special over samples. Non- response is a potential problem in data analysis. The greater the response rate, the more comfortable one can feel about the results. There was an overall 79% response rate for

Wave I data.

Add Health researchers conducted analyses on Wave I to ascertain any predictors of non-response in the dataset. Four predictors were found: contactability, unwillingness, inability and participation. If a participant’s family’s income was above the poverty level for the time (1995), the teen felt safe in their neighborhood, not smoked in the past 30 days or had good grades, they were more likely to complete the in-home questionnaire.

Respondents who appeared more unwilling to participate in the in-home interview had smoked cigarettes in the past month, whose parents did not volunteer for the PTA or who had a high school diploma or less. Inability was found in participants who lived in rural areas or smoked cigarettes in the past month. Finally those who were more likely to participate included racial/ethnic minorities, had parents who went to college or volunteered with the PTA. None of these measures or traits were included in the models tested (Kalsbeek, Yang & Agans, 2002).

Social network data were also available in the Add Health dataset. Social

network data are similar to other forms of data in terms analysis, but have distinct

differences in the types of information that can be generated from the ties an individual

has with others (Biegel, McCardle and Mendelson, 1985). The major difference in social network data compared to conventional data is the focus. Social network data concentrates on individual and the relationships they have with others while conventional data still focuses on their unique attributes instead of relationship ties. Data were collected on all students in each participating school in order to observe individual and

school-level networks. This design allowed one to look at the structure of each adolescent’s extended friendship network and describe the overall social structure of the child’s school. Friendship network data were available in the in-school questionnaire.

Each participant was asked to nominate up to five female and five male friends from the roster of all students in that child’s school and sister school. A sister school is defined as a school in the same community as the individual’s school.

Not all students’ names were listed due to errors on the rosters. In addition, not all

students whose names were listed on the rosters finished their questionnaire due to being

absent or not agreeing to participate in the study. If the ego’s school contained grades 7-

12, there was no sister school. An ego is an actual social network with its own data and

social connections (Biegel, McCardle and Mendelson, 1985). When friends were

designated, the respondent recorded their friend’s identification number on the

questionnaire. If the friend was not on the list, the respondent reported if the friend went

to their own school, sister school of did not attend either institution. Network measures

were computed for schools in which more than 50 percent of the student body completed

the questionnaire. The vast majority of friendship nominations were with other students

enrolled in the same school. Roughly 15% of all respondents’ friends did not attend their

own or their sister school and 8% of the nominations were children not on the school

rosters.

Measures

Cultural Status: The measure of cultural status consisted of four separate questions: mother’s, father’s and child’s place of birth (“were you born in the U.S.?”), and language preference (“language usually spoken at home-Spanish/English/Other”).

All four variables were dichotomous with the responses “yes” coded as 1 and “no” coded as 0. These codes were the affirmative of their label. Therefore the interpretation of the variable “born in the United States” would have a value of “1” if the child was born in the

US and a “0” if they were born outside the US.

These concepts have been used frequently in Latino research and are typically the basis of acculturation/cultural status measures (Cervantes & Castro, 1985; Zayas, 1987;

Rogler, Cortes & Malgady, 1991; Swanson, Linskey, Quintero Salinas, Pumariega &

Holzer, 1992; Vega, Gil, Warheit, Apospori & Zimmerman, 1993; Zimmerman & Zayas,

1995; Zayas & Dyche, 1995; Zayas, Kaplan, Turner, Romano & Gonzalez-Ramos, 2000).

Social Support: Measures of social support consisted of three family indexes

(mother support, father support and parent support, one teacher support question and two variables associated with friend support).

Family Support: The family index included 6 separate questions based on Haynie et al’s (2006) study on adolescent mobility and Kid et al’s (2006) study on parent, peer, and school social relations. These parental measures were also created from similar questions used in Turner et al’s (2002) study on parent-child support examining Latino

families and Borowsky et al’s (2001) study on adolescent suicide attempts and their closeness to both parents. Latino mothers and fathers have different roles and relationships with their daughters. As a result, mothers, fathers and the parental unit were examined separately to ascertain any differences. For example, Turner et al’s (2002) study focused on mother-daughter mutuality and how it protected Latina’s from an attempt. Although the literature has not focused on the father’s role as a protective factor, the measure was included to understand if mutuality was present among these fathers and daughters. The questions in these measures include:

1) How close do you feel towards your mother? 2) How much do you think she cares about you? 3) How close do you feel towards your father? 4) How much do you think he cares about you? 5) Most of the time your mother is warm and loving towards you. 6)

Most of the time your father is warm and loving towards you. The respondents had the option of answering: 1-not at all; 2-very little; 3- somewhat; 4-quite a bit; 5-very much.

Mother and father support indexes respectively had a minimum value of 0 and maximum value of 15. A lower value would indicate less support from their mother or father, while a higher value would infer more support.

The parent support index included all six questions regarding mother and father support described above into one variable. Family (mother, father and parental support) indexes were created from the previously mentioned variables.

Latinas were asked one question on how much their parents care about them.

“Parents care” was measured by the question, “How much do you feel that your parents care about you?” The respondents had the option of answering: 1-not at all; 2-very little;

3- somewhat; 4-quite a bit; 5-very much to the question. The variable had a minimum value of 0 and maximum value of 5. A higher value indicates more support while a lower score indicates lower support.

Teacher Support: Teacher support was measured by the question, “How much do you feel that your teachers care about you?” The respondents had the option of answering: 1-not at all; 2-very little; 3- somewhat; 4-quite a bit; 5-very much to the question. The variable had a minimum value of 0 and maximum value of 5. The teacher support variable has also specifically been used as a measure of bonding for adolescent suicide attempters (Crosnoe, Johnson & Elder, Jr., 2004).

Peer Support: Three variables were associated with friend support: “how much do feel that your friends care about you?”, “during the past week, how many times did you just hang out with your friends?” and “I feel socially accepted” “How much do you feel your friends care about you” was coded as follows: 1-not at all; 2-very little; 3- somewhat; 4-quite a bit; 5-very much. “How many times did you hang out with your friends was coded 0-3 (0=not at all, 1=1-2 times, 2=3-4 times and 3=5+ times). These measures have been used in Kid et al (2006) and were adapted from Crosnoe, Johnson &

Elder (2004) regarding their study on adolescent who attempted suicide. The social acceptance variable “you feel socially accepted” was coded 1 (strongly agree) to 5

(strongly disagree). A higher score indicated stronger feelings of peer support.

Suicidal Friends: Latinas were asked “Have any of your friends tried to kill themselves in the past 12 months?” This variable is coded 0/1. A respondent who had a friend who attempted suicide in the past year would have a value of one. This measure

was also used in Feigelman & Gorman’s (2008) study assessing the effects of peer suicide on youth suicide. Russel and Joyner (2001) used the same measure when examining adolescent suicide using Add Health data.

Reciprocated Friendships: This variable was found in the network questionnaire.

Respondents were asked to nominate their best female friend. In turn, two variables were created. The first variable specified if the best friend nominated the respondent as her friend. The second reciprocity variable showed if the best friend nominated the respondent as her best friend as well. The lower the number of identified friends who nominated the respondent as a friend, the lower the reciprocity in one’s friendship network. Reciprocity was coded 0=no reciprocity and 1=reciprocity between said-friend.

Bearman and Moody (2004) used the Add Health social network variables extensively in their study on adolescents’ likelihood of suicidality. The authors did not report any specific findings respective to Latinas.

Behavioral Health Status: The depression index included 19 of the 20 measures found in the CES-D. The depression index ranged from 0-19. Four items assessed positive indicators and were reverse coded before the scores were calculated. The depression index is based on a self-report scale used to assess depressive symptoms. The

CES-D is based on other, longer scales that have successfully measured depression. The present scale was tested in both in-home and clinical environments and found to have high internal consistency and test-retest reliability. Validity was determined through correlations of other similar measures and by its connection with other measures which support its construct validity. Use with various demographic groups has also been found

to be reliable and valid. This scale is widely accepted and used in most clinical settings in

the United States and has been a valuable tool in epidemiological studies (Radloff, 1977).

The 19 variables found in Add Health have also been used with success in with studies

focusing on adolescent depression and substance use (Shrier, Harris, Sternberg, &

Beardslee, 2001) in addition to adolescent suicide (Hallfors, Waller, Ford, Halpern,,

Brodish & Iritani, 2004).

Substance abuse/use index was based on a similar study using Add Health data

found in Pena et al’s (2008) study of Latina’s suicidal propensity due to substance use. I

did diverge slightly from Pena’s measures though. The two outcome variables (ideations

and attempts) were both asked in the timeframe of the past year. I believed that using

substance abuse measures that accounted for use in the past 30 days would be limiting

given that the outcome measure’s timeframe was the past 12 months. I did ask the same

questions as Pena and others (Granillo, Jones-Rodriguez and Carvaial, 2005) who studied

Latina adolescent drug use. As a result, I used the 12 month timeframe for alcohol use and lifetime use for illicit drugs (illicit drug use was asked for the past 30 days and lifetime only).

The measure included information on the respondents’ alcohol consumption in

the past 12 months, as well as use of illegal drugs (LSD, PCP, ecstasy, mushrooms,

speed, ice, heroin, or pills without a doctor’s prescription) all in one question while cocaine, inhalant and marijuana use were all asked separately. Participants were instructed to give the number of times that used any illicit drug in their lifetime.

Reponses to number of times an illicit substance was used ranged from 0 to 950 times.

Alcohol consumption was also included in the suicide index by “how many days did you

drink alcohol in the past 12 months”. Respondents were not asked lifetime use of alcohol.

Pena et al 2008 used this as their measure of alcohol use as well. The alcohol use variable

was coded 6 (every day or almost every day) to 1 (1 or 2 days in the past 12 months).

The higher value found in the depression index, the more symptoms the participant reported. The higher value in the substance abuse index, the higher the level of drug use

the Latina reported.

Suicidal Proclivity: Suicidal proclivity was measured using two observed

variables, ideation and attempts. Ideation was measured by “During the past 12 months,

did you ever seriously think about committing suicide?” with possible responses of “yes”

or “no”. If a participant had ideated suicide in the past year, the value would be one. Both

the ideation and attempt measures have been used extensively in the literature using Add

Health data (Borowsky et al, 2001; Bearman and Moody, 2004; Haynie, South & Bose,

2006; Silenzio, Pena, Duberstein, Cerel & Know, 2007). Attempts were measured by

“During the past 12 months, how many times did you actually attempt?” with possible responses of: 0= never; 1 time, 2-3 times, 4-5 times, 6+ times. Therefore, the higher the value the more suicide attempts the participant experienced in the past year.

Hypotheses Tested

Cultural Status. As a Latina’s cultural status aligns more with the majority culture (speaks English in the home, she and/or her parents were born in the U.S) her risk

of suicide ideations and attempts increase.

Family Support. Latinas reporting that their mother or father do not care or

support them will have a higher propensity of suicide proclivity than those Latinas

sharing their parents do care.

Teacher Support. Adult supports outside of their family can also be linked to suicide. A Latina who believes that her teachers do not care about her will be more likely to ideate or attempt suicide.

Friends Care. Latinas who believe that their friends care about them will be less likely to report any ideations or attempts in the past year.

Social Acceptance. Latinas who report their peers at school do not favorably recognize them will be more prone to attempt and ideate compared to those who are accepted by their peers.

Reciprocity. Latinas who engage in friendships that are not mutually reported will have an increased probability of ideations and attempts. Networks that are self- perceived as strong, but not reciprocated, will have tenuous bonds. These weak connections increase one’s risk of suicide.

Suicidal Friends. Latinas with friends who previously attempted suicide will have a higher likelihood of ideating or attempting themselves.

Depression. Latinas who exhibit biological, emotive or physical characteristics associated with depression will have an increased risk of suicidal thoughts or behaviors.

Latinas who report more thoughts and behaviors associated with a depression diagnosis will have a higher risk of suicide proclivity compared to those with minimal or no symptoms.

Substance Use. Latinas who use alcohol, marijuana, cocaine, inhalants or other illicit drugs are at a higher risk of suicide proclivity. Like depression, the more drugs one uses or the higher the number of times one has used substances, their likelihood of suicide ideations and attempts increases.

Summary of Hypotheses

The study worked from the hypothesis that suicide proclivity (ideations or attempts) is affected by a Latina’s self-reported level of social supports (parents, peers and teachers). Furthermore, it is the types and characteristics of their friendship networks that will forecast which peer groups would be most susceptible to suicide ideations and attempts. The core hypothesis is that there are various forms of support in a Latina’s social network and these connections will protect them from suicide proclivity.

A Latina’s cultural status (place of birth of child and parent and household language preference) will affect her suicide proclivity. For example, a Latina who primarily speaks Spanish in the home will then be forced to negotiate a more complicated set of cultural barriers outside of the home. This could lead to added pressure to create and maintain social supports outside of the home. Cultural status is known to be related to depression and substance abuse, and these in turn, are known to be related to suicide

(Cervantes & Castro, 1985; Zayas, 1987; Rogler, Cortes & Malgady, 1991; Swanson,

Linskey, Quintero Salinas, Pumariega & Holzer, 1992; Vega, Gil, Warheit, Apospori &

Zimmerman, 1993; Zimmerman & Zayas, 1995; Zayas & Dyche, 1995; Zayas, Kaplan,

Turner, Romano & Gonzalez-Ramos, 2000).

Analytic Strategy

Preparation of Data

Data was recoded and/or transformed in SPSS 17.0. A working data file was then created where data errors were eliminated. All descriptive analysis was conducted in

SPSS as well. STATA 10.0 is not compatible with normality tests involving survey data

(W. Eddings, personal communication, August 6, 2009). Multiple imputation, logistic

and negative binomial regressions were carried out in STATA 10.0 due to the software’s

ability to accommodate complex sample designs.

Descriptive Statistics

Prior to model testing descriptive analyses were conducted on all variables

including means and ranges. Univariate and bivariate correlations will be presented in

Chapter 4. These introductory analyses enable the reader to have a basic understanding of the data and were later utilized to interpret the results used in model testing.

Negative Binomial Regression

A series of regression models were conducted. Issues associated with using dichotomous variables in OLS regression are the same as using count data (Cohen,

Cohen, West & Aiken, 2002). First, residuals would not be normally distributed, heteroscedastic and result in biased predictions. Scores would also not be interpretable if scores would be less than zero. The regression coefficients could also be biased and variable because they would not become more accurate by increasing the sample size

(Long, 1997). Finally, the standard errors will increase the t-tests for the coefficients and

overestimate the significance of the coefficient as well (Gardner, Mulvey & Shaw, 1995).

Given that the dependent variable represented a count of suicidal attempts, a sequence of negative binomial models was estimated.

Negative binomial is an extension of the poisson regression (Hilbe, 2007; Long,

1997; Cameron & Trivedi, 1998). Many of the assumptions of the poisson are essentially the same as the negative binomial regression. One important deviation from poisson is that the negative binomial regression assumes that the mean and variance are not equal

(Hilbe, 2007). Negative binomial regression calculates count models when the poisson estimation is deemed unsuitable as a result of overdispersion. Overdispersion occurs when the variance is greater than the mean. Using a poisson regression would be incorrect in this study because the standard errors for the poisson regression are biased in the presence of overdispersion.

Detection of overdispersion can be done a number of ways such as dividing the deviance and Pearson chi-square by the degrees of freedom. If the variance exceeds the mean, essentially that the values are greater than 1, there is overdispersion and values less than 1 suggest underdispersion, where the variance is less than the mean. If there is an indication of underdispersion or overdispersion there is inadequate fit of the poisson model (Cameron & Trivedi, 1998). I first observed the histograms of suicide attempts variable to make a visual interpretation of the distribution. The variable did not indicate a normal distribution due to its high kurtosis value, difference in variance and means made it clear that multiple regression and poisson regression techniques were not appropriate for our hypothesized models.

The development of the model included a variety of steps through forward

selection. The forward selection process began with the constant-only model and then

each variable was entered one at a time. First, all four cultural status variables were

regressed on the variable of ideation to detect any possible connections between culture

and the outcome variable. Afterwards the measures representing age, depression,

substance use, suicidal friends, parental supports and then a variety of peer supports and

contacts from their network were included. Variables that were entered but did not reach

statistical significance were not included in subsequent models.

As with any regression technique, outliers can bias the results by shifting the regression line and providing incorrect regression coefficients. Residual analysis can allow the researcher to ascertain if outliers are present in the data and if the results potentially created a Type I error. After the final models were decided upon, the residuals of each negative binomial regression were examined. Examining the residuals allows detection of outliers in addition to the confirmation of a linear relationship. The residuals were plotted against each of the independent continuous variables to test for any possible non-linear relationship among the variables (Hilbe, 2007).

Logistic Regression

With a dichotomous outcome variable such as suicidal ideations, logistic regression was used in this dissertation. The argument for using logistic regression instead of OLS regression is the same as my argument for employing negative binomial regression to count variables (Cohen, Cohen, West & Aiken, 2002). First, residuals would not be normally distributed, heteroscedastic and result in biased predictions if a

dichotomous variable was used in an OLS regression. The dependent variable also does not need to have the same variances within categories. In addition to not assuming a normal distribution, logistic regression does not assume error terms are normally distributed. Perfect or high multicollinearity is not allowable due to the inflation of standard errors (Hosmer & Lemenshow, 2000). Given that dichotomous variables have either a value of zero or one, no scores could be negative. The regression coefficients could also be biased and variable because they would not become more accurate by increasing the sample size (Long, 1997). Finally, the standard errors will increase the t- tests for the coefficients and overestimate the significance of the coefficient as well

(Gardner, Mulvey & Shaw, 1995).

Given that the dependent variable represented whether or not a Latina ideated suicide or not, a sequence of logistic models was estimated. By converting the dependent variable into a logit variable, logistic regression calculates the odds that an event might happen (Hosmer & Lemenshow, 2000). Logistic regression is a popular statistical technique because it does not require that a linear relationship exists between the dependent and independent variables. Logistic regression does assume independence among observations and that the variables have a linear relationship with the dependent variable (Hosmer & Lemenshow, 2000). However, it is quite robust to departures from linearity.

The process of creating a full model was identical to those implemented in the negative binomial regressions. Forward selection included all four cultural status variables, followed by age, depression index, substance use index, suicidal friends,

parental supports and then a variety of peer supports and contacts from their network.

Again the cutoff level was reached when variables included in the model reported a

significance of p<.05.

In the final model, statistically significant variables were reported in addition to

their odds ratios. For continuous variables such as the depression index, this ratio indicated the odds change for each unit change in the independent variable. An odds ratio lower than 1 represented a decrease in the odds of the event happening while a value higher than 1 signified an increase in the odds of the event occurring. A ratio of 1 suggested that the independent variable had no effect. The odds ratios of dichotomous independent variables were interpreted a bit differently. The difference is that for dichotomous independent variables there is only one value at which the odds change. For example, the variable “suicidal friends” was coded 0=no suicidal friends, 1=friends who have attempted suicide. The value of 0 represented the reference category. Therefore, an odds ratio > 1 indicated those with suicidal friends were more likely to ideate suicide where an odds ratio < 1 suggested Latinas with suicidal friends were less likely to ideate.

Residuals were also observed with the logistic models. The standardized residuals were plotted against independent variables in the model and identified any outliers. Again, this procedure allowed one to understand any variation between the observed and expected probabilities. Residuals that were two standard deviations from the mean indicated that the model was not properly identified (Hosmer & Lemenshow,

2000).

As with any regression technique, outliers can affect the results dramatically.

Therefore standardized residuals need to be analyzed as necessary. Large samples are important when using logistic regression due to maximum likelihood estimations (MLE).

MLE depends on large samples to maintain the reliability of the estimates. If one is using

a small sample for analysis, high standard errors are likely (Hosmer & Lemenshow,

2000; Cameron & Trivedi, 1998).

Other Statistical Considerations. Multicollinearity can be a concern when using

either negative binomial or logistic regression and was tested by using bivariate

correlations (Weston & Gore, 2006). Correlations higher than r=.85 will identify possible

issues (Kline, 2005). If variables are highly correlated, the redundant variable can be removed (Weston & Gore, 2006).

Normality was tested by looking at the distribution of every measured variable.

While normality of predictors is not a requirement in negative binomial regression, understanding the distribution can help one understand how the covariates might affect the outcome variables. Missing data can be an issue with longitudinal research and also

with questions related to mental health and suicide proclivity (Allison, 2001). For

variables with more than 20% of missing data, multiple imputation (MI) was utilized. For

data with a large number of missing cases (>40%), these variables were removed.

Multiple imputation is a Monte Carlo technique that replaces missing data

through simulation. This has become an increasingly common practice in the field of epidemiology which typically use large datasets or in studies with high percentages of missing values. Rubin (1987) contended that 50% missing data, five imputations will

suffice. Due to its efficiency, generality and the high number of missing values found in

the dataset, MI was used instead of listwise deletion

Estimation of Full Model. The model was tested by entering each variable separately until all variables were included, representing the full model. Once accomplished, the social support and network relationships of primary interest were tested within the initial model’s framework. The sample size was large (1,618 Latinas)

and had sufficient statistical power to detect a moderate effect at a probability of .80

(Cameron & Trivedi, 1998; Shumacker & Lomax, 2004).

Model Fit & Interpretation. Variables that were not statistically significant at the

p<.05 level were eliminated as well as those with high FMI (fraction of missing

information). The FMI is the proportion of the between- to within-imputation variance of

the estimated coefficient and its degrees of freedom, ranging from 0 to 1 (Royston, Carlin

& White, in press). Although Royston, Carlin and White do not specify specific cutoff

levels for a high FMI value, they conclude in their article that a large FMI (i.e. higher

percentage of missing information) should be used with caution in terms of interpretation

and inclusion in a model. Finally, alternative models were tested. If the p-value or slope

changed significantly then the less restrictive alternative model was determined as the

best-fitting model.

CHAPTER 4

RESULTS

Chapter four consists of three sections: univariate and bivariate analyses and model testing using negative binomial and logistic regressions. Univariate analyses included the mean, median, proportion estimates for each variable to illustrate the attributes of the sample used. The next section incorporates a series of regressions used to ascertain the relationships between the dependent and independent variables. The final section tests the study’s hypotheses using logistic and negative binomial regressions.

Covariates including cultural status and behavioral health status (depression and

substance use), self-reported social supports (adult and peer) were included. For the

logistic regressions, odds ratios were reported to help in the interpretation of the results.

The effect size of the negative binomial regressions was reported using the incidence risk

ratio (IRR).

This dissertation addressed how a Latina’s parents, friends and adults outside her family affect her propensity of suicidal ideations and attempts. The study specifically examined a Latina’s friendship network regarding suicide proclivity. I also included several cultural status measures to ascertain if language use and place of birth mattered in terms of suicidal risk.

Given the complex sample design of the Add Health dataset, stratification and

cluster variables in addition to grand sample weights were included in all regression

models presented and during the imputation process. Missing values were imputed to

allow for more precise inferences to the larger population the data was meant to

characterize. Cases that were missing weight stratification and cluster variables were

removed (Chantala & Tabor, 1999).

Univariate Analyses

Cultural Status. Cultural status represented a respondent’s level of English proficiency in the home and whether the adolescent or either of her parents was born in the United States. Each of the cultural status variables (English speaking, Born in the

U.S., Father born in the U.S., Mother born in the U.S.) were coded 0/1. The minimum value was 0 and maximum value of 1. Almost half, 46% of the Latinas in the study came from homes that spoke primarily English. The majority of Latinas were born in the

United States (62%). Very few mothers (4%) and fathers (15%) were born in the United

States.

Family Support. (Table 1). The mother support index included 3 questions

pertaining to mother-daughter mutuality and had a range of values of from 0-5. The

mother support index had a mean of 2.0 and median of 2.2. The average of 2.0 on the

mother support index indicated that Latinas agree “somewhat” that their mothers were

supportive to them. The father support index ranged from 0-5, with a mean of 1.35 and a

median of 1.8. The mean score of 1.35 suggested that Latinas that felt that their fathers

were supportive “very little”.

Latinas were asked one question concerning their perception of the level of

parental care toward them. The variable was coded on a Likert scale of 0-5. The average

score was 4.77 and a median of 5.00. The mean score reflected the degree of self-

perceived care Latinas received from their parents. A value of 4.77 reflected, by the

recoding of this variable, Latinas believed their parents “cared quite a bit” about them.

Table 1. Descriptive Statistics of Perceived Continuous & Count Family Support

Variables

Variable Mean Median

Mother Support 2.00 2.20

Father Support 1.35 1.80

Parents Care 4.77 5.00

School Supports. (Table 2). School supports documented connections Latinas

stated they had with their teachers and peers at school. “Teachers care” had a range of values from 1-5. The mean value was 3.46 and a median of 3.00. The average meant that Latinas believed their teachers cared “somewhat” about them.

“Friends care” had a minimum score of 0 and maximum value of 6. The average for friends who care about them was 4.26. A mean of 4.26 suggested that the average

Latina reported their friends cared “quite a bit” for them.

Table 2. Descriptive Statistics of Count School Support Variables

Variable Mean Median

Teachers Care 3.46 3.00

Friends Care 4.26 4.00

Suicidal Friends. Each Latina was questioned if she had or did not have a friend

who had attempted suicide in the past year. The dichotomous variable indicated that 22%

of the Latinas questioned had a suicidal friend.

Frequency of Contact. (Table 3). Latinas were asked how often they “hung out

with their friends” in the past week. The range of possible scores was 0-3 with an

average rate of 1.78 in-person contacts per week, a median of 2.00. A 1.78 mean

indicated that the average Latina hung out with her friends 1-2 times per week.

Table 3. Descriptive Statistics of Frequency of Contact

Variable Mean Median

Hung Out With Friends This Week 1.78 2.00

Behavioral Health. (Table 4). Latinas answered questions pertaining to the

psycho/social/behavioral aspects of their lives related to depression as determined by

CES-D measurement tool. Nineteen questions were compiled into an index that had a minimum score of 0 and a maximum score of 19. The mean for the depression index was

3.03, with a median of 2.81.

The respondents were asked about their use of alcohol and illicit drugs in their

lifetime (past year for alcohol use). I created a substance abuse/use index included based on the respondents’ alcohol consumption in the past 12 months, as well as use of illegal drugs (LSD, PCP, ecstasy, mushrooms, speed, ice, heroin, or pills without a doctor’s prescription) all in one question while cocaine, inhalant and marijuana use were all asked separately. Participants were instructed to give the number of times used in their lifetime.

The substance abuse index resulted in responses of 0 times of use to 1020 times of use.

The average number of times of use was 8.22 for any drugs, with a median of .200.

Examining each substance abuse variable allows one to understand the distribution better and marijuana use (mean of 5.27, median of .000) appeared to be the most common form of illicit drug used in a Latina’s lifetime. Latinas reported they had used alcohol an average of 1.63 times in the past year with a median of 1.00.

Table 4. Descriptive Statistics of Continuous Behavioral Health Variables

Variable Mean Median

Depression Index 3.03 2.81

Substance Abuse Index 8.22 .200

Marijuana Lifetime 5.27 .000

Alcohol Past 12 Months 1.63 1.00

Reciprocity. Two variables examined a Latina’s reciprocity of her friendships.

Both variables were dichotomous and measured if a Latina’s best friend reported her as a

friend and if the best friend stated she was also her best friend. A Latina’s best female

friend reporting her as a friend resulted in 68% of the friendship relationships

reciprocated. Only 39% of those whom participants reported as best friend reciprocated.

Social Acceptance. (Table 5). Latinas reported quite strongly on feeling socially accepted by their peers. With a range of possible scores of 1-5, Latinas averaged a score of 3.96 and a median score of 4.00. A 3.96 mean indicated a neutral response to the question.

Table 5. Descriptive Statistics of Social Acceptance Variable

Variable Mean Median

Feel Socially Accepted by Peers 3.96 4.00

Suicide Ideations and Attempts (Table 6). The ideation variable was coded 0/1

while attempts ranged from 0-6 points. Roughly 17% of the Latinas in the study reported

an ideation in the past year. The number of Latinas who reported an attempt in the past

year averaged .091 and a median of .000. Due to suicide attempts being a low base rate

phenomenon, with 93 Latinas reporting an attempt in the past year, the mean of .091 does

not provide much information.

Table 6. Descriptive Statistics of Suicidal Ideations

Variable Mean Median

Suicide Attempts .091 .000

Proportion estimates were reported only for continuous and ordinal variables

since the means of dichotomous variables were equal to their estimated population proportion.

Substance Use. The most common drug used by the Latinas in the past year was alcohol (58%) in the past year. Latinas reported their lifetime illicit drug use. Almost one-third (30%) of the Latinas reported marijuana use. Inhalants and cocaine were asked separately. Inhalants were used by 6% of the Latinas in the study while 5% stated they had used cocaine at least once in their life.

Nine percent of the sample used “other drugs” (LSD, PCP, ecstasy, mushrooms, speed, ice heroin or pills).

Suicide Attempts Most Latinas (94%) reported never attempting suicide. Of the

6% who stated they attempted suicide in the past year, 3.5% attempted once, 1.5% had attempted 2-3 times, .5% reported 4-5 attempts and .3% reported attempting 6 or more times in the past 12 months. Therefore, in this sample roughly 6% of the Latinas had attempted suicide at least once in the past year.

Teachers Care. Latinas were asked how much they believed their teachers cared for them. Four percent reported their teachers did not care at all for them, 12% stated their teachers cared very little, while 34% shared their teachers cared somewhat. About one-third (31%) of Latinas indicated their teachers cared quite a bit for them, while 18% indicated that their teachers cared very much for them.

Friends Care. Respondents were questioned about their perceptions of how much their friends care for them. Few Latinas (.7%) felt that their friends did not care for them

at all; three percent shared their friends cared very little, while 13% believed their friends cared somewhat for them. Over one-third (34%) of the Latinas in the study stated their friends cared quite a bit for them and 49% answered that their friends cared very much for them.

Parents Care. The vast majority of Latinas (85%) reported that their parents care for them very much; followed by 7% stating their parents cared quite a bit. Far fewer

(4%) of the Latinas indicated their parents cared somewhat for them, followed by 1% feeling parent cared very little and .5% stated parents did not care at all.

Hang out with Friends. The majority of Latinas (60%) spent time with their friends at least 3 times in the past week (27% reported 3-4 times and 33% reported 5 or more times). One-quarter (25%) of the Latinas stated they hung out with their friends 1-2 times in the past week, while 15% did not spend any time with their friends.

Table 7. Proportion Estimates of Count Variables

Variable %

Substance Use Index 30 Marijuana 9 Other Drugs 6 Inhalants 5 Cocaine Suicide Attempts 94 Never Attempted 3.5 Attempted Once 1.5 Attempted 2-3 Times 0.5 Attempted 4-5 Times 0.3 Attempted 6 or More Times Teachers Care 4 Didn’t Care At All 12 Cared Very Little 34 Cared Somewhat 31 Cared Quite a Bit 18 Cared Very Much Friends Care 0.7 Didn’t Care At All 0.3 Cared Very Little 13 Cared Somewhat 33 Cared Quite a Bit 49 Cared Very Much Parents Care 0.5 Didn’t Care At All 1 Cared Very Little 4 Cared Somewhat 7 Cared Quite a Bit 85 Cared Very Much Hang Out with Friends (past week) 15 0 Times 25 1-2 Times 27 3-4 Times 33 5 or More Times

Latina Suicidal Attempts and Social Supports

Negative binomial regressions were performed using the covariates previously described, and using suicide attempts as the outcome variable. It was decided to use the

English speaking variable in the model for two primary reasons. First, the variable was

the only one of the cultural status variables (mom, father and child born in the U.S and

English spoken in the home) that was a statistically significant predictor of attempts when modeled without any covariates. Secondly, English speaking approached statistical significance (p=.051) when modeled with depression and attempts.

Model Development

Using forward selection procedures, I entered each variable to the model in the following manner based upon the literature for both outcome variables (suicidal ideations and attempts):

Forward Selection Model 1

1) English

2) Depression

3) Mother Support

4) Father Support

5) Teacher Support

6) Suicidal Friends

7) Substance Abuse

8) Other peer supports (socially accepted, friends care, “hanging out with friends”,

reciprocity).

I also created a second model structure to test if the combined parental unit showed any relationship with suicide proclivity:

Forward Selection Model 2

1) English

2) Depression

3) Parents Care

4) Teacher Support

5) Suicidal Friends

6) Substance Abuse

7) Other peer supports (socially accepted, friends care, “hung out with friends”,

reciprocity).

Forward selection is a useful tool in exploratory research for two reasons. First this

method allows for systematic investigation with a number of variables. Second, even

slight changes can be observed when variables fall below statistical significance when

other variables are entered.

For each model discussed, a variety of test statistics were reported including the

confidence levels. The 95% confidence interval is indicated for all models described in

this study. The confidence interval represents what the values would be if the regressions

were repeated on multiple samples. The range of values presented would include the true

population parameter 95% of the time (Cox, 1974).

This first model used the previously-mentioned forward selection procedure that

included mother and father support. English spoken in the home (English speaking) was

entered first, followed by being depressed (Table 8). English speaking was statistically

significant at the .10 level (b=.887, t=1.98, p=.051) and depression (b=.134, t=6.76, p≤.001). Next, the measure of mother support was entered and was also found to be statistically significant at the .10 level (b=-.070, t=-1.96, p=.053) (Table 9). English speaking (b=.929, t=2.06, p=.042) and depression (b=.130, t=6.81, p≤.001) continued to be statistically significant. Father support was then entered into the existing model

(Table 10). English speaking (b=.912, t=2.00, p=.049) and depression (b=.127, t=-1.73, p≤.001) were significant while father support was not (b=-.036, t=-.92, p=.360) and mother support continued to be significant at the .10 level (b=-.067, t=-1.73, p=.086).

Table 8

Summary of Negative Binomial Regression Analysis for Variables Predicting Cultural

Status (n = 1,618) Controlling for Depression

Variable Coef SE t Sig (p) Confidence

Interval

English Speaking .887 .448 1.98 .051 -.002-1.77

Depression .134 .019 6.76 ≤.001 .094-.173

Table 9

Summary of Negative Binomial Regression Analysis for Variables Predicting Suicidal

Attempts by Cultural Status and Family Support (n = 1,618) Controlling for

Depression

Variable Coef SE t Sig. (p) 95% Confidence

Interval

English Speaking .929 .452 2.06 .042 .033- 1.826

Depression .130 .019 6.81 ≤.001 .092 -.168

Mom Support -.070 .036 -1.96 .053 -.141-.001

Table 10

Summary of Negative Binomial Regression Analysis for Variables Predicting Suicidal

Attempts by Cultural Status and Family Support (n = 1,618) Controlling for

Depression

Variable Coef SE t Sig. (p) Confidence

Interval

English Speaking .912 .457 2.00 .049 .004 - 1.819

Depression .127 .019 6.43 ≤.001 .088 - .167

Mother Support -.067 .039 -1.73 .086 -.145 - .009

Father Support -.036 .039 -.92 .360 -.114 - .042

When father support was added to the existing model of mother support, depression and speaking English in the home, father support (b=-.036, t=.360) had a p-value of .360 and

was eliminated from the model. Mother support (b=-.067, t=-1.73, p=.086) remained non-significant at the p<.05 level, but English speaking (b=.912 t=2.00 p=.049) and depression (b=.127 t=6.43 p≤.001) were significant. Next the measure of teacher support was entered into the model and found to be statistically significant (p=.016) (Table 11).

The English speaking variable’s (b=.951, t=1.83, p=.070) significance level did increase from the previous model. Mother support (b=-.066, t=-1.74, p=.085) remained not significant, but was also kept in the model. Depression (b=.122, t=6.08) continued to be significant (p≤.001) and teacher support (b=-.285, t=-2.45, p=.016) was significant.

Table 11

Summary of Negative Binomial Regression Analysis for Variables Predicting Suicidal

Attempts by Cultural Status and Family and Teacher Support

(n = 1,618) Controlling for Depression

Variable Coef SE t Sig. (p) Confidence

Interval

English Speaking .951 .464 1.83≤ .070 -.070-1.772

Depression .122 .020 6.08 ≤.001 .081- .161

Mom Support -.066 .038 -1.74 .085 -.140-.009

Teachers Care -.285 .116 -2.45 .016 -.515-.054

Suicidal friends was entered next in the model (Table 12) and found to be significant (b=-

.271, t-2.25, p=.003) along with teacher support (p=.027) and being depressed b=.115, t=5.72, p≤.001). English spoken in the home rose above the .10 level of significance

(b=.633, t=1.44, p=.154) while mother’s support remained below .10 but above .05 (b=-

.073, t=-1.80, p=.075).

Table 12

Summary of Negative Binomial Regression Analysis for Variables Predicting Suicidal

Attempts by Cultural Status and Family and Teacher Support

(n = 1,618) Controlling for Depression

Variable Coef SE t Sig (p) Confidence

Interval

English Speaking .633 .440 1.44 .154 -.241 -1.50

Depression .115 .020 5.72 ≤.001 .075 - .154

Mother Support -.073 .041 -1.80 .075 -.155 - .007

Teacher Support -.271 .120 -2.25 .027 -.510- -.031

Suicidal Friends 1.02 .334 3.07 .003 .361 - 1.68

Other Measures Entered-Model 1-Attempts

Various peer support variable were also entered into the existing model that failed

to attain statistical significance. Substance abuse (b=.007, t=1.86, p=.810) , being socially accepted (b=.286, t=1.21, p=.299), friends care (b=107, t=.63, p=.533), hung out

with friends (b= .097, t=.60, p=.549), best friend reciprocates as a friend (b= .097, t=.60,

p=.549), best friend reciprocates as a best friend (b= 1.06, t=1.64, p=.163) were all non-

significant.

Excluding all measures that were not statistically significant at the p≤.05 level, the final model included depression (b=.111, t=5.73, p≤.001), teacher support (b=-.311, t=-

2.68, p=.009) and suicidal friends (b=1.13, t=3.23, p=.002) (Table 13). The effect size measured by the incidence rate ratio indicated that for every point in the depression scale a Latina reports, her likelihood of attempts increases by 11%, for even level of support a

Latina reports her teachers care for her likelihood of attempts decreases by 26%, and if a

Latina reports she had a suicidal friend, her likelihood of attempting increased by 209%.

Table 13

Summary of Negative Binomial Regression Analysis for Variables Predicting Suicidal

Attempts by Teacher Supports and Suicidal Friends (n = 1,618) Controlling for

Depression

Variable Coef SE t Sig (p) Confidence

Interval

Depression .111 .019 5.73 ≤.001 .072 - .150

Teacher Support -.311 .116 -2.68 .009 -.541 -.080

Suicidal Friends 1.13 .352 3.23 .002 .440 - 1.83

The second model also used the forward selection procedure and included the measure

“parents care”. English spoken in the home and depression were previously entered (see

Table 8). The first variables entered were English spoken in the home, depression and now parents care (Table 14). All measures were statistically significant, depression

(b=.125, t= 6.10, p≤.001), English speaking (b=.899, t= 2.33, p=.022) and parents caring

(b=-.548, t= -3.34, p≤.001).

Table 14

Summary of Negative Binomial Regression Analysis for Variables Predicting Suicidal

Attempts by Cultural Status and Parental Support (n = 1,618) Controlling for

Depression

Variable Coef SE t Sig (p) Confidence Interval

English Speaking .899 .386 2.33 .022 .133- 1.665

Depression .125 .020 6.10 ≤.001 .084- .165

Parents Care -.548 .166 -3.34 ≤.001 -.875-.222

Next, the teachers care measure was entered into the model (Table 15). English speaking

(b=.827, t=2.08, p=.040), depression (b=.117, t= 5.53, p≤.001, parents care (b=-.500, t= -

3.04, p=.003) and teachers care (b=-.236, t= -2.22, p=.029) were all found to be

statistically significant.

Table 15

Summary of Negative Binomial Regression Analysis for Variables Predicting Suicidal

Attempts by Cultural Status, Parental and Teacher Support (n = 1,618) Controlling for

Depression

Variable Coef SE t Sig (p) Confidence Interval

English Speaking .827 .397 2.08 .040 .039- 1.615

Depression .117 .021 5.53 ≤.001 .075- .159

Parents Care -.500 .164 -3.04 .003 -.828 -.173

Teachers Care -.236 .106 -2.22 .029 -.447 -.025

Having a suicidal friend was then entered into the model (Table 16). English speaking

(b.572, t=1.48, p=.141) is no longer significant, while the remaining variables, depression

(b=.110, t=5.34, p ≤.001), parents care (b=-.524, t=-2.82, p=.006), teachers care (b=-.231, t= -2.21, p=.030) and having a suicidal friend (b=1.07, t= 3.58, p=001) were all found to be statistically significant.

Table 16

Summary of Negative Binomial Regression Analysis for Variables Predicting Suicidal

Attempts by Cultural Status, Parental and Teacher Support and Friendship

Characteristics (n = 1,618) Controlling for Depression

Variable Coef SE t Sig (p) Confidence

Interval

English Speaking .572 .385 1.48 .141 -.193 - 1.33

Depression .110 .020 5.34 ≤.001 .069 - .151

Parents Care -.524 .185 -2.83 .006 -.893 -.156

Teacher Support -.231 .104 -2.21 .030 -.439 - -.023

Suicidal Friends 1.07 .298 3.58 .001 .477 - 1.66

Variable Coef SE t Sig (p) Confidence Interval

English Speaking .572 .385 1.48 .141 -.193 - 1.33

Depression .110 .020 5.34 ≤.001 .069 - .151

Parents Care -.524 .185 -2.83 .006 -.893 -.156

Teacher Support -.231 .104 -2.21 .030 -.439 - -.023

Suicidal Friends 1.07 .298 3.58 .001 .477 - 1.66

Other Measures Entered-Model 2-Attempts

Various peer support variable were also entered into the existing model that failed to attain statistical significance. Substance abuse (b=.004, t=1.00, p=.330), being socially accepted (b=.265, t=1.25, p=.215), friends care (b=.051, t=.29, p=.773), hanging out with

friends (b=.081, t= .52, p=.608), best friend reciprocates as a friend (b= .161, t= .26,

p=.806) and best friend reciprocates as a best friend (b= .770, t= -.82, p=.432) were not

statistically significant.

The final model included all statistically significant measures (Table 17). Being

depressed (b=.104, t=5.43, p≤.001), having a parent who is caring (b=-.532, t=-3.19, p=.002), having a teacher who is caring (b=-.263, t=-2.36, p=.020) or having a suicidal friend (b=1.21, t=3.88, p=.001) were related to suicide attempts. The effect size measured by the incidence rate ratio indicated that for every point in the depression scale a Latina reports, her likelihood of attempts increases by 10%, a Latina’s likelihood of an attempt decrease 41% for every level she reports her parents care for her and 23% decrease for every level she reports her teachers care for her. If a Latina reports she had a suicidal friend, her likelihood of attempting increased by 235%.

Table 17

Summary of Negative Binomial Regression Analysis for Variables Predicting Suicidal

Attempts by Parental and Teacher Support and Friendship Characteristics (n = 1,618)

Controlling for Depression

Variable Coef SE t Sig (p) Confidence Interval

Depression .104 .019 5.43 ≤.001 .066- .143

Parents Care -.532 .166 -3.19 .002 -862--.201

Teachers Care -.263 .111 -2.36 .020 -.484--.042

Suicidal Friends 1.21 .312 3.88 ≤.001 .591-1.83

Summary of Negative Binomial Regressions

In the two models there was an apparent theme: adult support protected Latinas

from attempts and friend support did not. Although having suicidal friends correlated

with an increase in suicide attempts, having friends to count on or having strong bonds

with peers, did not have an affect on attempts. While their mother’s and father’s

mutuality independently did not protect these Latinas from attempts, knowing that their

parents collectively cared did decrease their chances of attempting. Teachers who cared

about their Latina students also decreased the chance they she might attempt suicide.

While depression played a role in whether or not a Latina attempted, her use of substances was not associated with an attempt.

Latina Suicidal Ideations and Social Supports

Logistic regressions were used for models with the dichotomous outcome variable of suicidal ideations. Again, English spoken primarily in the home was used as a predictor of suicidal ideations (Table 18). English and depression were then included in the initial model for forward selection. Depression (Odds ratio= 1.14, t=8.25, p<.001) was significant while English speaking was significant at the .10 level (Odds ratio=1.56, t=1.67, p=.098).

Table 18

Summary of Logistic Regression Analysis for Variables Predicting Suicidal Attempts by Cultural Status (n = 1,618) Controlling for Depression

Variable Odds Ratio SE t Sig (p) Confidence

Interval

English Speaking 1.56 .417 1.67 .098 .918- 2.656

Depression 1.14 .018 8.25 ≤.001 1.105- 1.177

Mother support was then entered into the model (Table 19). English speaking in the home remained significant at the .10 level (Odds ratio=1.62, t= 1.88, p=.062) while mother support (Odds ratio= .962, t=-.97p=.336) was non-significant and being depressed (Odds ratio=1.13, t= 8.11, p≤.001) was significant.

Table 19

Summary of Logistic Regression Analysis for Variables Predicting Suicidal Ideations by Cultural Status and Parental Support (n = 1,618) Controlling for Depression

Variable Odds SE t Sig (p) Confidence Interval

Ratio

English Speaking 1.62 .415 1.88 .062 .974 - 2.694

Depression 1.13 .018 8.11 ≤.001 1.103 - 1.177

Mother Support .962 .038 -.97 .336 .889 – 1.041

Father support was then entered into the model (Table 20). English speaking in the home

(Odds ratio=1.59, t= 1.76, p=.082) and mother support (Odds ratio=.967, t= -.83, p=.408)

were both found to be non-significant. Since mother support had very high p-values in the current and last model, it was eliminated from future models. Being depressed (Odds ratio= 1.13, t= 7.81, p≤.001) and father support (Odds ratio=.948, t= -3.08, p=.003) were related to suicidal ideations.

Table 20

Summary of Logistic Regression Analysis for Variables Predicting Suicidal Ideations

by Cultural Status and Parental Support (n = 1,618) Controlling for Depression

Variable Odds Ratio SE t Sig (p) Confidence Interval

English Speaking 1.59 .419 1.76 .082 .941 - 2.685

Depression 1.13 .018 7.81 ≤.001 1.098 - 1.171

Mother Support .967 .037 -.83 .408 .895 - 1.046

Father Support .948 .016 -3.08 .003 .916 - .981

Father support was then entered into the model again (Table 21). English speaking in the

home (Odds ratio=1.57, t= 1.69, p=.094) continued to be significant at the .10 level.

Being depressed (Odds ratio=1.15, t= 7.85, p≤.001) and having a supportive father (Odds

ratio=1.57, t-1.69, p=.003) were both significant.

Table 21

Summary of Logistic Regression Analysis for Variables Predicting Suicidal Ideations

by Cultural Status, Parental Support (n = 1,618) Controlling for Depression

Variable Odds Ratio SE t Sig (p) Confidence Interval

English Speaking 1.57 .423 1.69 .094 .924508 2.68433

Depression 1.15 .020 7.85 ≤.001 1.089 - 1.179

Father Support .947 .016 -3.05 .003 . 914 - .981

Teacher support was then entered into the model again (Table 22). English speaking in

the home (Odds ratio= 1.52, t= 1.60, p=.113) and having a teacher care (Odds ratio=

.921, t= -.89, p=.347) were found to be non-significant. Being depressed (Odds

ratio=1.134, t= 7.74, p≤.001) and having a supportive father (Odds ratio= .949, t= -3.00, p=.003) were related to suicidal ideations.

Table 22

Summary of Logistic Regression Analysis for Variables Predicting Suicidal Ideations

by Cultural Status, Parental and Teacher Support (n = 1,618) Controlling for

Depression

Variable Odds Ratio SE t Sig (p) Confidence Interval

English Speaking 1.52 .406 1.60 .113 .902 – 2.59

Depression 1.134 .018 7.74 ≤.001 1.097 - 1.170

Father Support .949 .016 -3.00 .003 . 916 - .982

Teachers Care .921 .084 -.89 .374 .767 - 1.105

Suicidal friends was then entered into the model again (Table 23) English speaking in the

home (Odds ratio=1.44, t= 1.34, p=.182) was found to be non-significant. Being

depressed (Odds ratio= 1.13, t= 7.69, p≤.001) and having a supportive father (Odds

ratio= .948, t= -.292, p=.005) and having a suicidal friend (Odds ratio= 2.09, t=3.22,

p=.002) were related to suicidal ideations.

Table 23

Summary of Logistic Regression Analysis for Variables Predicting Suicidal Ideations

by Cultural Status, Parental Support and Friendship Characteristics (n = 1,618)

Controlling for Depression

Variable Odds Ratio SE t Sig (p) Confidence Interval

English Speaking 1.44 .397 1.34 .182 .838 - 2.494

Depression 1.13 .018 7.69 ≤.001 1.095 - 1.167

Father Support .948 .017 -2.91 .005 . 914 - .983

Suicidal Friends 2.09 .483 3.22 .002 1.329 - 3.314

Other Measures Entered-Model 1-Ideations

Various peer support variable were also entered into the existing model that fell

below significance. Being socially accepted (Odds ratio=1.07, t= .41, p=.685) friends care (Odds ratio=1.12, t= 1.17, p=.245) hung out with friends (Odds ratio=1.12, t= 1.27,

p=.207) substance abuse (Odds ratio=1.00, t= 1.21, p=.243) best friend reciprocates as

friend (Odds ratio=.801, t=-.75, p=.581) best friend reciprocates as best friend (Odds

ratio=.777, t= -.58, p=.581) were not statistically significant.

The final model included three variables, depression, father support and suicidal

friends (Table 24). Being depressed (Odds ratio=1.12, t=7.35, p≤.001) and having a

supportive father (Odds ratio=.947, t=-2.98, p=.004) and having a suicidal friend (Odds

ratio=2.20, t=3.36, p=.001) were all related to suicidal ideations. The effect size is

measured by the odds ratio. There was a 12% increase in ideations if the Latina reported

depression, a 6% decrease if she reported that her father was supportive and a 120% increase if she had a suicidal friend.

The model’s fit was ascertained via the Hosmer-Lemenshow test in STATA 10.0.

The statistic did not approach significance (p=.333) therefore, the model appears to fit quite well. This test is typically used with caution due to its dependency upon sample size. For small samples, the test has a tendency to indicate good model fit incorrectly and for larger samples, even if the model actually fits, the test may indicate failure. Given that the test’s result was non-significant, it is fair to say that the model fits well given the large sample size (Hosmer & Lemenshow, 2000).

Table 24

Summary of Logistic Regression Analysis for Variables Predicting Suicidal Ideations

by Parental Support and Friendship Characteristics (n = 1,618) Controlling for

Depression

Variable Odds Ratio SE t Sig (p) Confidence Interval

Depression 1.126 .018 7.35 ≤.001 1.095 - 1.160

Father Support .947 .017 -2.98 .004 . 913 - .982

Suicidal Friends 2.20 .516 3.36 .001 1.381 -3.505

Variable Odds Ratio SE T Sig (p) Confidence Interval

Depression 1.126 .018 7.35 ≤.001 1.095 - 1.160

Father Support .947 .017 -2.98 .004 . 913 - .982

Suicidal Friends 2.20 .516 3.36 .001 1.381 -3.505

The second model of predictors of ideations was constructed in the same fashion as the previous three models. First English, depression and parents care were included in

the initial model (Table 25) English speaking in the home (Odds ratio=1.59, t=1.76,

p=.081) was found to be significant at the .10 level. Being depressed (Odds ratio=1.13,

t=7.64, p≤.001) and having parents care (Odds ratio=.730, t= -2.51, p=.014) were statistically significant.

Table 25

Summary of Logistic Regression Analysis for Variables Predicting Suicidal Ideations by Cultural Status and Parental Support (n = 1,618) Controlling for Depression

Variable Odds Ratio SE t Sig (p) Confidence Interval

English Speaking 1.59 .419 1.76 .081 .943 – 2.685

Depression 1.13 .018 7.64 ≤.001 1.098 - 1.173

Parents Care .730 .091 -2.51 .014 .569 –.936

Teachers care was then entered into the model (Table 26). English speaking in the home

(Odds ratio=1.54, t=1.68, p=.095) maintained significance at the .10 level while having a teacher care (Odds ratio=.928, t=-.77, p=.441) was found to be non-significant. Being depressed (Odds ratio=1.13, t=7.59, p≤.001) and having parents care (Odds ratio=.743, t=-2.23, p=.029) were significantly related to suicidal ideations.

Table 26

Summary of Logistic Regression Analysis for Variables Predicting Suicidal Ideations by Cultural Status, Parental and Teacher Support (n = 1,618) Controlling for

Depression

Variable Odds Ratio SE t Sig (p) Confidence Interval

English Speaking 1.54 .402 1.68 .095 .924 - 2.594

Depression 1.13 .018 7.59 ≤.001 1.097 –1.171

Parents Care .743 .098 -2.23 .029 .570 – .968

Teachers Care .928 .088 -.77 .441 .767 – 1.123

Suicidal friends was then entered into the model (Table 27) English speaking in the home

(Odds ratio=1.43, t=1.34, p=.183) was found to be non-significant. Being depressed

(Odds ratio=1.13, t=7.75, p≤.001), having a suicidal friend (Odds ratio=2.16, t= 3.49,

p≤.001) and having parents care (Odds ratio=.722, t=-2.64, p=.010) were related to

suicidal ideations.

Table 27

Summary of Logistic Regression Analysis for Variables Predicting Suicidal Ideations

by Cultural Status and Parental Support and Friendship Characteristics (n = 1,618)

Controlling for Depression

Variable Odds Ratio SE t Sig (p) Confidence Interval

English Speaking 1.43 .389 1.34 .183 .840 – 2.460

Depression 1.13 .018 7.75 ≤.001 1.097 –1.166

Parents Care .722 .088 -2.64 .010 .565 – .922

Suicidal Friends 2.16 .479 3.49 .001 1.396 – 3.359

Other Measures Entered-Model 2- Ideations

Various peer support variable were also entered into the existing model that fell

below significance. Social acceptance (Odds ratio=1.08, t=-.48, p=.634), friends care

(Odds ratio=1.16, t=1.53, p=.130), hung out with friends (Odds ratio=1.61, t=1.12,

p=.130), substance abuse (Odds ratio=1.00, t= 1.39, p=.178), best friend reciprocates as

friend (Odds ratio=.709, t=-.80p=.455), best friend reciprocates as best friend (Odds ratio=.775, t=-.80 p=.444) were not statistically significant. The final model included

only measures statistically significant at the p < .05 level (Table 28). Being depressed

(p≤.001), having a suicidal friend (p≤.001) and having parents care (p=.012) were related to suicidal ideations. Odds ratios indicate that there is a 12% increase in suicidal ideations if the Latina is depressed), a 28% decrease in the likelihood of ideations if her parents care for her and a 127% increase in the likelihood of an ideation if she has a suicidal friend.

Table 28

Summary of Logistic Regression Analysis for Variables Predicting Suicidal Ideations

by Parental Support and Suicidal Friends (n = 1,618) Controlling for Depression

Variable Odds Ratio SE t Sig (p) Confidence Interval

Depression 1.12 .018 7.24 ≤.001 1.089 - 1.163

Parents Care .721 .091 -2.56 .012 .560 –.930

Suicidal Friends 2.27 .517 3.61 ≤.001 1.448 - 3.575

The model’s fit was ascertained via the Hosmer-Lemenshow test in STATA 10.0.

The statistic was not significant (p=.946) therefore, the model appears to fit quite well.

Summary of Logistic Regressions:

Model 1

• Latinas who report depressive symptoms are 12% more likely to ideate.

• If a Latina has a suicidal friend she has a 120% increase in the likelihood of an

ideation than one who does not have a suicidal friend.

• Latinas who reported that her father was supportive were 6% less likely to

ideate suicide.

Model 2 • Latinas who report depressive symptoms are 12% more likely to ideate.

• If a Latina has a suicidal friend she has a 120% increase in the likelihood of an

ideation than one who does not have a suicidal friend.

• Latinas who reported that their parents are care were 6% less likely to ideate

suicide.

Results Summary

Chapter four consisted of univariate and model testing procedures using Add

Health data. Variables with missing values were imputed, resulting in five datasets.

Weights were included in the analysis to provide a more precise generalization to the population represented.

Negative binomial and logistic regressions were utilized for hypothesis testing.

Forward entry analysis was conducted in both regression types indicating the strength of specific predictors of suicide proclivity. Models were then depicted in various tables indicating statistically significant results.

Hypotheses tested

Cultural Status. As a Latina’s cultural status aligns more with the majority culture (speaks English in the home, she and/or her parents were born in the U.S) her risk of suicide ideations and attempts increase.

It was anticipated that a Latina’s cultural status would increase the incidence of suicide attempts based upon the literature. Although birth place was not found to be

significant in any of the models discussed, English spoken in the home was found to be a risk factor for suicide attempts but not ideations in this sample of Latinas in initial models. Once “suicidal friends” was entered into both ideation and attempt models,

English spoken in the home fell below significance at the .10 level.

Family Support. Latinas reporting that their parents cared for them were less likely to attempt or ideate suicide.

Although mother support was not found to be significant in any models, a father’s support was found to be a protective factor in a Latina’s propensity for suicidal thoughts.

Latinas who reported that their parents cared for them had a significant decrease in suicidal ideations and attempts.

Teacher Support. Adult supports outside of their family can also be linked to suicide. A Latina who believed that her teachers do not care about her was found to be more likely to attempt suicide. Teacher support was not found to be a factor for Latina suicidal ideations.

Friends Care. Although Latinas reported that they believed their friends cared about them very much, the measure of “friends care” was not significant with regard to ideations or attempts.

Social Acceptance. Social acceptance was not found to be a protective factor for suicide proclivity.

Reciprocity. A Latina’s level of reciprocity either from her friends or her best friends was not found to be a significant factor in suicide proclivity.

Suicidal Friends. Having a friend who had attempted suicide in the past year was found to be a risk factor for a Latina in terms of suicide attempts and ideations.

Depression. Depression was found to increase a Latina’s risk of suicidal ideations and attempts.

Substance Use. Substance use was not found to be a significant factor for

Latina’s suicidal risk.

The following chapter will include the theoretical and clinical applications of the research findings including implications for practice and also the strengths and weaknesses of the study.

CHAPTER 5

DISCUSSION

The final chapter will be divided into five sections. The first section will examine the descriptive statistics of the sample related to the suicidal risk outcomes including familial, peer and school support. The second section will summarize the results from the negative binomial and logistic regression models. The third section will include practice implications associated with each significant model. The fourth section will discuss the strengths and limitations for the dissertation. Policy and research implications will be discussed in the final section.

Suicidal Ideations and Attempts

A similar longitudinal study’s results for the same time period, the Youth Risk and Behavior Surveillance System (YRBSS), will be used as a comparison for this dissertation. The YRBSS was conducted in 1995 and involved complex sample methodologies similar to the Add Health dataset. Although most Latinas (94%) in the sample never attempted suicide in the past year, 17% reported an ideation and 6% reported at least one attempt in the past 12 months. This rate of attempts is far less than the YRBSS (1995) reported. Their survey found that 31% of Latinas had ideated in the past year and 21% had attempted in the past year. This study did not support the findings

of the YRBSS (1995) study. The results from this dissertation could be due to the large number of Latinas who came from immigrant households.

Cultural Status

The percentage of Latinas born in the United States in this sample was indicative of most young Latinos. More than half (61%) were born in the U.S. The US census

(2001) reported at the time the questionnaire was completed (1994-1996), that 60% of

Latinas were born in the U.S. Noting Zayas’ (1987; 2008) and other studies (Vega et al,

1993; Hovey, 1996; 1998; 2000) involving acculturating Latina adolescent’s propensity for attempts and ideations, having a higher proportion of US-born Latinas could have biased the results. An interesting finding is that only 47% of these Latinas stated they spoke English in the home predominately. Therefore, these females might be living in more traditional homes due to having immigrant parents since the number of Latinas having immigrant parents was very common (mother born in the US 4%, father 15%).

Again, by comparing U.S. Census reports when the questionnaire was disseminated, the numbers of immigrant Latino parents with children who are foreign-born were quite similar to our findings. The U.S. Census Current Population Survey (1998) found that

85% of families consisting of at least one immigrant parent had at least one child who was born in the U.S.

Family Support

Latinas in the sample generally felt cared and supported for by their parents.

When looking at their mothers and fathers separately, Latinas reported they had less support from their father (very little) than their mother (somewhat). This finding was

surprising given that the literature has stressed the importance of mother-daughter

mutuality and family support in general (Turner et al, 2002). This may be because they

anticipated more support from their mothers due to social expectations. Therefore a

Latina expects her mother to be supportive and assumes to receive less support from her

father. As a result, if her father is slightly supportive, she might weigh that more heavily

than her mother’s level of support. These measures were collected for biological fathers

and not necessarily a father figure who currently lives in the home.

Latinas were asked one specific question about parental caretaking. The average

Latina reported that parents care quite a bit about them. The responses signified that

parents do play an active role in their children’s lives which supports the literature review

findings (Turner et al, 2004; Borowsky et al, 2001; Zayas et al, 2000; Delgado, 1995).

This dissertation found strong bonds between parents and Latinas given that the majority

(86%) stated that their parents cared very much for them. Therefore, Latinas in this study might have seen the global sense of parental care as more important than mother/father- mutuality.

School Support

Latinas reported that they generally believed their friends cared (82%) and teachers cared (48%) at least quite a bit for them. This is important since children spend most of their time at school (U.S. Department of Education, 2000). Given that Latinas also feel very cared for at home, one might suppose that Latinas in this study appear to be supported both at home at school. In terms of peer relations at school, the average Latina

reported having problems at school “a few times” with 60% reporting few or no problems

at school.

However, Latinas had problems with their peers at school. Almost half (46%)

either agreed or strongly agreed that they did not feel close to people at school. As a

result, Latinas might believe they are cared for by their peers, but they do not have close

bonds with them. If a Latina does not feel close to her peers, the likelihood of her turning

to them during a crisis is doubtful. This finding could potentially affect school suicide

prevention programs that are common among adolescents. Latinas who might believe

that their friends care about them could be helpful to the group dynamic, but if they do

not feel close to their peers any self-disclosure of suicidal thoughts might be difficult for

a facilitator to obtain.

Due to the findings that elders matter, a possible therapeutic intervention, such as

Multisystemic Therapy (MST), could be used in suicide prevention. A feature of MST is

to incorporate the family unit and other systems in the child’s life into the therapeutic

process. Although having a child’s teacher physically present in the sessions could prove

difficult, other alternatives such as providing classroom progress notes could provide

parent and clinician important information. Likewise, teachers could receive clinical

notes from the therapist advising current issues to work on in the classroom. Since having

a teacher care decreased the likelihood of Latina suicide, including teachers could be an

important factor to consider. Creating a continual stream of information might be important for a child dealing with suicidal thoughts or behaviors whose behavior can

spiral quickly. If the child has immigrant parents, they might not know what specific

mood or behavioral changes to observe in their child. This process would also educate each parent on the warning signs of suicide.

Suicidal Friends

Latinas reported they had a friend who attempted suicide in the past year (22%).

These findings were similar to those of the Bearman and Moody (2004) study on adolescent friends and suicidal risk. Latinas are not any different than other adolescents, having a suicidal friend increases your risk of suicide proclivity. The percentage of suicidal ideations and attempts presented previously sheds some light on this finding.

Latinas in this study reported they had ideated (17%) and attempted (6%). It is not surprising that almost one-quarter of the respondents had a suicidal friend given the high rate of ideations. This is an interesting finding since these Latinas, coming from predominately immigrant households, most likely did not learn suicidal coping behaviors from their parents. Since suicide is a rare phenomenon in countries such as Mexico, these females are likely learning these coping skills elsewhere. Understanding the type of friends Latinas associate with is crucial. Due to the higher rates of suicidal contagion, among adolescents, having friends who are at more risk for suicidal thoughts and behaviors increases one’s risk of suicide proclivity. Research on suicidal clusters posits that children tend to mimic their friend’s coping skills and this could be reflected in suicide rates (Gould, Wallenstein, Kleinman, 1990, Joiner, 2002).

Frequency of Contact

Since isolation is a risk factor for suicide proclivity, understanding the number of contacts Latinas experience can help ascertain their risk. Somewhat surprising was that

85% of the Latinas who responded said they hung out with their friends outside of school

time. Since most of the Latinas come from immigrant households, females typically are

assigned household chores which tend to take them away from socializing with their

peers outside of school. The sample appeared to be well connected to their friends out of

the school environment which could indicate a possible protective factor for suicide.

Therefore, the quality and maybe not the quantity of contacts could be a factor in suicidal

risk.

Depression

Latinas in this study had a mean depression score of 3.03. This score is difficult to

interpret since one of the questions from the CES-D was omitted from the questionnaire.

Further, the CES-D is scored from 0-3 for each question while the Add Health

questionnaire provided yes/no answer options. Therefore interpreting the mean is

difficult. In any case, the more depressed a Latina was, the more likely she was to attempt

and ideate suicide. Having clinicians and other adults in their lives identify depressed

Latinas and assessing their risk of suicide is critical. Some individuals might be reluctant to ask a Latina if she has thought of suicide. If clinicians and teachers are not aware of

their own biases with suicide one can not provide an emotionally safe environment for

their client to express their thoughts. Intensive training needs to be implemented in

school systems, especially those in Latino neighborhoods, since the rates of suicide

among Latinas is so high.

100

Substance Use

Alcohol was the most common drug of choice (1.63 times in the past year), which

is not surprising given its accessibility. Marijuana (average number of times used in

lifetime was 5.27) was the most used illicit drug, which is similar to SAMHSA reports in

1995 (2008). The standard deviation was quite high in this study (23.46) which is not

surprising given that most teens never use illicit drugs. This finding also indicates that those Latinas, who do use substances, use them quite frequently. Of those Latinas stating they had used an illicit substance, 30% used marijuana, 9% used other drugs, 6% inhalants and 5% used cocaine.

Information from the YRBSS 1996 report was collected in 1995. Although some of their survey questions were different, their findings can shed some light on drug use among Latinas in the mid-1990s. First, Latinas were asked about their past-month’s use of alcohol. Over half (52%) answered they had at least one drink in the past 30 days. No

definitive results were supplied regarding Latinas’ use of marijuana in their report and

also through the Center for Substance Abuse Treatment (2002). CSAT found that in

1995 Latinos (both boys and females) had the highest rate of marijuana use (31%)

compared to White and African-American teens. The YRBSS discovered that Latinas

(11.6%) reported in 1995 that they had used cocaine in their lifetime. This was more than

double what was found with the Latinas in this study. The YRBSS’s report solely stated

results of Latino use of inhalants and other illicit drugs (LSD, PCP, ecstasy, mushrooms,

speed, ice and heroin) and did not report specific to females and boys’ usage. The

101

YRBSS found that in 1995, 23% of Latino adolescents used inhalants and 18% used other illicit drugs.

Again, educators can take the lead in their current prevention programs by simply addressing how substance use and suicide are often related. School-based drug prevention programs should focus on the relationship between depression, substance use and suicide. Many children have been under the influence of drugs at the time of their death resulting from a suicide attempt (Queralt, 1994).

Reciprocity

Latinas’ impressions of their relationships with peers were mixed. Latinas had a relatively high reciprocity rate among their friends (68%) but did not have a true sense of who their best friends were in their respective schools (39%). This is a meaningful finding because although previously the respondents stated they knew friends cared for them, they also had problems feeling close or connected to people at school. This finding could support their responses of feeling cared for but not feeling close to them given the lack of reciprocity in their peer relationships. These results also take the quality of these

Latinas’ peer relationships into question. The females in this dissertation had a less than

50% chance of reciprocation as best friends from their best friends. Further research regarding the reciprocity of African American, White females and Latino males would help put these results into better perspective regarding suicide proclivity.

Social Acceptance

Although Latinas reported they felt cared for by friends, they did not feel close to their peers and had little reciprocity with their best friends. Latinas in the sample stated

102

they neither agreed nor disagreed if they felt socially accepted at their school. Latinas

who do not have strong connections to their peers might believe the reason their

friendships are not strong is due to their lack of acceptance. It is not clear if a Latina’s

acceptance from her peers, the reciprocity of her relationships or the closeness she

internalizes from her friends might be derived from her cultural status or not, but further

inquiry is necessary to understand how they relate to suicide proclivity.

Latina Adolescents and Suicide Attempts

Three negative binomial models indicated that adult support, whether from their family or teachers, shielded Latinas from suicide attempts. Given that traditional Latino culture stresses respect for elders, this finding was not surprising in terms of parent

support. A somewhat unexpected result was that teachers also had an impact on suicide

attempts. Again, since one could assume Latinas in this study came from traditional

immigrant families, individuals outside of the family unit might not matter in terms of

support. Although this study did not examine if these Latinas were sharing any type of

emotional concerns or problems with their teachers, one might suppose that if they did

have problems they might turn to someone they believed cared for them, including their teachers.

A point to stress is that the findings differ in some aspects from previous studies.

Turner and colleagues (2004) discovered that the ability to feel close to their mothers and

confide in them protected Latina adolescents from suicide. We did not find any

connection between a Latina’s perceived level of maternal support guarding against

suicide attempts. We did find that if their parents cared for them, Latinas were less likely

103

to attempt suicide. These findings are important for future suicide prevention curricula.

Including parents and their teachers in suicide prevention programs might be more

effective for Latinas than traditional school-based initiatives traditionally facilitated by a

trained clinician or guest speaker. Teachers might be an un-tapped resource in suicide

prevention in new and exciting ways. For example, a teacher could help the marginalized

adolescent in the classroom by providing support in the classroom with short “check-ins”

daily to monitor any changes in mood but also allow the teen to know her teacher cares.

Since this measure correlated highly with Latina adolescents this could be a critical piece of school-based suicide prevention.

Role conflict can also be a factor with the Latinas in the study. Therefore, a mother’s sole support might not be a protective factor. Latinas with immigrant parents typically adopt the role of a cultural broker, mediating the relationship between their family and teachers. As a result, these Latinas have a different relationship with not only their parents but with their teachers. By brokering these adult relationships, her view of peer relationships might not actually be with those her age, but with those outside her household including her teachers. Therefore having friends who care or are supportive might not matter to Latinas from immigrant homes, or even materialize, when compared to the relationships she has with their teachers.

Although the literature review stipulated that adolescents tend to turn to friends when they were contemplating suicide, this might not be the case for Latinas. A variety of friendship variables were included in the models to ascertain any level of risk for suicide attempts. Given that a majority of these Latinas came from immigrant families,

104

one might assume that family connections were very strong for these females. Therefore, influences from their peers might not be as important as they might be for second or third generation females. For Latinas in immigrant families, her peer support might come from family members her age. For example, maybe a better predictor of peer support for

Latina adolescents in immigrant households would include their cousins rather than children at school or in their neighborhood.

The emphasis on peer-mentored suicide prevention groups might be a less effective tool for Latinas coming from immigrant households than for other adolescents.

Peer support should still be considered important with regard to Latinas and suicidal risk, but adults appear to matter more with girls from immigrant households. Although these girls report feeling cared for by their friends, they do not appear to protect them from suicide proclivity.

The presence of a suicidal friend increased one’s risk of suicide attempts and ideations remained consistent with the literature. Friends do appear to matter, but it is the type of friends Latina adolescents choose that makes an apparent difference. Physical or emotional connections per se did not alter either the likelihood of suicidal ideations or the number of attempts. Having a suicidal friend was a significant factor. Therefore, awareness of a child’s friends could become an early suicide prevention tool for parents, teachers and clinicians. For instance, alerting parents to their child’s friend’s attempt could help benefit a Latina. Providing documentation of this in Spanish and English would be imperative. Given that many of the children in the study came from immigrant

105

homes, where suicide is less common in their homeland, having school or parent-teacher

meetings on the topic could make parents aware of this phenomenon.

In addition to providing suicide prevention information in Spanish, material should also be disseminated in a variety of non-traditional locations in the community.

For example, continuing to post suicide hotline numbers in schools and clinics is important, but also providing information in other settings such as neighborhood restaurants or in places of worship. Since Latino immigrants are less likely to obtain traditional mental health services, bringing prevention information to areas where they might frequent could increase awareness to Latino parents and elders.

The suicidal friends measure was statistically significant in all suicidal ideation and attempt models presented. Furthermore, suicidal friends continually eroded the detrimental effects that speaking English in the home had for suicide attempts. This finding can also shed some light on the types of friends one has as they increase their language use in the home. For example, as a Latina increasingly uses more English with her household members, her social network might change. A Latina, whose family is not as dependent on their daughter’s English ability, might be able to increase her social network. This potentially larger network might not provide emotional supports but actually allow her to make friends with a more diverse and potentially riskier friendship network.

Latina Adolescents and Suicidal Ideations

Although speaking predominately English in the home was not significantly related to ideations, other variables protected Latinas from suicide proclivity. Lack of

106

self-perceived father support was found to increase a Latina’s risk of ideations. This

finding has not been previously found in the literature. In fact, father support has been

included in past models without much success. Past research concluded that the father

typically was not in the home enough to make an effect on their child’s mental health

status or that childrearing was seen as more of a mother’s responsibility in a traditional

Latino household. Therefore, a father’s level of emotional support sheds new light on the

needs of young Latinas and suicidal risk. It should be noted that this related to their

biological father and not necessarily the male figure living in the child’s home.

The mother-daughter relationship might also be different for these Latinas compared to those in more acculturated households. Young Latinas are exposed to

American viewpoints about women’s roles in school or in the media. Daughters might become frustrated by their immigrant mother’s continued practice of traditional gender roles in the home. These feelings might make her align more with her father, who has a higher likelihood of working outside the home and adopting more mainstream characteristics than the mother. The young Latina might have more contact with her mother as well while in the home. With this increased interaction, the daughter might see her mother as more of the disciplinarian while her father might appear to be more loving and nurturing. Therefore, a father’s support might be interpreted differently than the support of their mother.

In both the negative binomial and logistic regressions, Latinas who were depressed were more likely to ideate or attempt suicide. The odds ratios give a more precise idea of how likely Latina adolescents were to ideate. Latinas who were depressed

107

were more likely to ideate. This was consistent with Bearman and Moody’s (2004) study

on adolescents (OR=1.45). The more depressed an adolescent was, the more likely she

was to ideate suicide. Latinas who had suicidal friends were twice as likely to ideate

compared to Latinas who did not have a friend who attempted. Given the high risk of

suicide contagion among adolescents, parents and teachers should be keenly aware of

Latina’s friendship network. More attention should be focused on the impact their

father’s support had on these females (6% reduction in suicidal ideations).

Suicidal friends was statistically significant in all suicidal ideation and attempt

models presented and continually eroded the detrimental effects that speaking English in

the home had for suicide attempts. This finding can also shed some light on the types of

friends one has as they increase their language use in the home. For example, as a Latina

increasingly uses more English with her household members, her social network might

change. A Latina, whose family is not as dependent on their daughter’s English ability,

might be able to increase her social network. This potentially larger network might not

provide emotional supports but actually allow her to make friends with a more diverse and potentially riskier friendship network.

A father’s support functioning as a protective factor for his Latina daughter’s suicidal ideations has not been supported in the literature. While researchers have

focused on the mother-daughter relationship with Latinas, more attention needs to be

given to a father’s role regarding his daughter’s mental health. In addition, incorporating

the father in family therapy sessions is imperative. Furthermore, psychoeducational

108

programs need to include curriculums focusing on the Latino father and his role in

protecting his daughter from suicidal ideations.

Limitations

One limitation of this dissertation, often associated with self-report

questionnaires, is the risk of social desirability bias. This can affect the outcomes of the

study. For example, due to the stigma of mental health diagnoses and suicide in general,

respondents might have minimized their actual occurrence in their lives. The statistical

methods accommodated for this weakness somewhat by utilizing earphones for sensitive

questions and then entering their answers in provided computers.

Children who were not in school the day the questionnaire was distributed, or who had dropped-out of school, did not have their own unique experiences represented in this study. These children arguably would be without the positive emotional supports of their

teachers, likely making them more susceptible to suicidal thought and behaviors.

The models themselves were limited in terms of missing data. Although multiple

imputation was utilized, the software still had some difficulty in assigning values to some

variables namely the peer, friendship network and substance use variables. As a result, I was not able to make inferences to the population on important peer supports and risk factors such as substance use.

The Add Health dataset also had a limited number of cultural status variables.

Although the place of birth of a Latina and her parents, along with English spoken in the

home, can provide some insight into the level of acculturation of an individual, there are

109

other qualitative measures that would have provided a more conclusive depiction of the

construct of acculturation.

The data was also collected over 10 years ago. The Latino population has

changed in terms of acceptance and perception by the majority in recent years. Therefore, changes in cultural status and suicidal behavior patterns over the past decade could potentially limit the validity of the results to present-day Latinas.

The sample used for this study was not representative of the population.

Although 62% of the sample was born in the United States, only 4% of mothers and 14% of fathers were born in the U.S. Therefore, the sample was essentially a study of mostly

immigrant families. In addition, only 6% of the Latinas in the study had reported an

attempt in the past 12 months. Latinas questioned in other national studies, such as the

YRBSS, reported they (21%) had attempted suicide in the same year the Add Health data was collected.

Strengths

The dissertation also had a number of strengths due to methodological and

statistical considerations. The Add Health dataset incorporated a nationally

representative sample of adolescents. The inclusion of weighted analysis, logistic and

negative binomial regressions improved the precision of the results.

The study was also unique in a number of ways. First, this dissertation examined

not only a Latina’s cultural status and familial supports, but also a variety of peer

supports. Friendship networks are seldom studied among Latinas and their suicidal

proclivity. This study is the first of its kind to observe possible protective factors

110

associated with a Latina’s peer group along with cultural status and risk of suicide.

Finally, the study’s results were compared to a similar school-based sample, the Youth

Risk Behavior Surveillance System when possible.

Policy and Research Implications

Policy Implications

Latinas continue to have incredibly high rates of suicide yet Latina-focused prevention programs are rare. Since the rates have not decreased, mental health professionals, teachers and parents all need to play an active role in developing new, culturally competent prevention tools. Although research has indicated for over 60 years that the Latina suicide experience is different when compared to other females, those in direct service and policy arenas have been slow to act, especially in areas where the

Latino population rate is high.

Using past research can help individuals recognize and empower Latino families, immigrant and non-immigrant, and acknowledge the unique factors associated with

Latina adolescent suicide. The study concluded that adult supports mattered to these

Latinas. Therefore, creating programs that focus on reinforcing family bonds and enhancing parental, especially father, communication with young Latinas can make a difference with her coping mechanisms. One option for clinicians is to utilize family- centered therapies in order to work with the entire family unit. The collectivist nature of traditional Latino families could potentially benefit from dealing with the family unit as opposed to the child dealing with suicidal tendencies. Given the projected increase in the

111

Latino population in the next 40 years, finding effective suicide prevention programs through mental health, local, state and federal levels is imperative.

Research Implications

Additional research is necessary to understand why adult supports, especially with those outside the family unit, protect Latinas from suicidal thoughts and behaviors.

Future research directions should include the following questions: How do interventions utilizing the cultural and familial dynamics introduced in this study enhance prevention efforts for Latina teens? Does incorporating these suggested factors enrich the parent- child relationship and drive down suicidal ideations and attempts with Latina adolescents? Do specific types of friends, such as peers who have used illicit substances, increase one’s risk of suicide proclivity and not their own personal usage?

Conclusion

This chapter assessed significant determinants and found that Latinas who have adult supports fair better in terms of suicide proclivity. Theoretical implications for

Attachment Theory were considered together with significant findings of how social supports have a fundamental influence on a Latina adolescent’s ideations and attempts in this sample. Subsequent sections discussed the strengths and limitations of this dissertation, in particular methodological techniques used and the novel approach to including peer networks. Finally policy and research implications and were suggested such as the creation of interventions that cater to Latino communities in addition to the incorporation of adult supports.

112

REFERENCES

Abraido-Lanza, A. (1999). The Latino Mortality Paradox: A test of the “salmon bias” and healthy migrant hypothesis. American Journal of Public Health, 89, 1543- 1548.

Alegria, M., Shrout, P., Woo, M., Guarnaccia, P., Sribney, W., Vila, D., Polo, A., Cao, Z., Mulvaney-Day, N., Torres, M. & Canino, G. (2007). Understanding differences in past year psychiatric disorders for Latinos living in the U.S. Social Science and Medicine, 65(2), 214-230.

Allebeck, P., Allgunalnder, C. & Fisher, L. (1988). Predictors of completed suicide in a cohort of 50, 465 young men: Role of personality an deviant behaviour. British Medical Journal 297, 6642, 176-178.

Allen, M., Elliot, M., Fuligni, A., Morales, L. Hambarsoomian, K., Schuster, M. (2008). The relationship between Spanish language use and substance use behaviors among Latino youth: a social network approach. Journal of Adolescent Health, 43(4), 372-279.

Allgood-Merten, B., Lewinsohn, P. A., & Hops, H. (1990). Sex differences and adolescent depression. Journal of Abnormal Psychology, 99, 55–63.

Allison, P. 2001 Missing Data. Thousand Oaks, CA: Sage Publications.

American Association of Suicidology. (2004). Youth Suicide Fact Sheet Based on Current (2004) Statistics. Retrieved October 23, 2006 from http://www.jedfoundation.org/documents/YouthSuicide.pdf

American Association of Suicidology. (2006). Suicide in the U.S.A. Based on 2003 Statistics. Retrieved on October 20, 2006 from http://www.suicidology.org/associations/1045/files/SuicideInTheUS.pdf

American Association of Suicidology. (2007). Facts about suicide and depression. Retrieved on February 5, 2007 from http://www.suicidology.org/associations/1045/files/Depression.pdf

113

American Psychiatric Association. (2005). Let’s Talk: Facts about teen suicide. Retrieved on October 20, 2006 from http://www.healthyminds.org/multimedia/teensuicide.pdf

Asher, S. (1983). Social competence and peer status: Recent advances and future directions. Child Development, 54, 1427-1434.

Arzubiaga, A., Ceja, M. & Artiles, A. (2000). Transcending deficit thinking about Latinos' parenting styles: Toward an ecocultural view of family life. Retrieved on August 5, 2009 from http://courses.coe.asu.edu/artiles/pubs/LatinosParentingStyles.pdf

Bandura, A. (1973). Aggression: A social learning analysis. Oxford, England: Prentice-Hall.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182.

Bearman, P. & Moody, J. (2004). Suicide and friendships among American adolescents. American Journal of Public Health, 94(1), 89-95.

Berndt, T. (1989). Contributions of peer relationships to children’s development. In Berndt, T & Ladd, G. (Eds.). Peer Relationships in Child Development. New York: Wiley & Sons, Ltd., 407-416.

Berry, J. (1997). Immigration, acculturation, and adaptation. Applied Psychology, 46(1), 5–34.

Beautrais, AL. (2002). Gender issues in youth suicidal behaviour. Emergency Medicine, 14(1), 35-42.

Biegel, D., McCardle, E., Mendelson, S. (1985). Social Networks and Mental Health: An annotated bibliography. Sage Publications: Beverly Hills, CA.

Bigelow, B. & LaGaipa, J. (1980). The development of friendship values and choices. In Foot, H., Chapman, A. & Smith, R. (Eds.). Friendship and Social Relations in Children. New York: Wiley & Sons, Ltd., 16-44.

Birckmayer, J., & Hemenway, D. (1999). Minimum-age drinking laws and youth suicide, 1970–1990. American Journal of Public Health, 89, 1365–1368.

114

Blake, S., Ledsky, R., Goodenow, C. & O’Donnell, L. (2001). Recency of immigration, substance use and sexual behavior among Massachusetts adolescents. American Journal of Public Health, 91(5), 794-798.

Blyth, D. & Foster-Clark, F. (1987). Gender differences in perceived intimacy with different members of adolescent social networks. Sex Roles, 17, 689-718.

Boothby, N. (1994). Trauma and violence among refugee children. In. A. J. Marsella, T. Bornemann, S.Ekblad, & J. Orley (Eds.),Amidst peril and pain: The mental health and well-being of the world’s refugees pp. 239–259). Washington, DC: American Psychological Association.

Borders, T., Brannon-Goedeke, A., Arif, A., & Xu, K. (2004). Parents’ reports of children’s medical care access: Are there Mexican-American versus non-Hispanic white disparities? Medical Care. 42(9), 884–892.

Borowsky, I., Resnick, M., Ireland, M. & Blum, R. (1999). Suicide attempts among American Indian and Alaska Native youth: Risk and protective factors. Archives of Pediatrics and Adolescent Medicine, 153, 6, 573-580.

Borowsky, I., Ireland, M. & Resnick, M. (2001). Adolescent suicide attempts: Risks and protective factors. Pediatrics, 107, 3, 485-493.

Bovasso, GB. (2001). Cannabis abuse as a risk factor for depressive symptoms. The American Journal of Psychiatry, 158, 2033-2037.

Brent, D., Perper, J., Moritz, G., Allman, C., Schweers, J., Roth, C. et al. (1993). Psychiatric sequelae to the loss of an adolescent peer to suicide. Journal of the American Academy of Child & Adolescent Psychiatry, 32(3), 509-17.

Brent, D. (1993). Depression and Suicide in Children and Adolescents. Pediatrics in Review, 14, 380-388.

Brent DA, Perper JA, Moritz G, Allman C, Friend A, Roth C, Schweers J, Balach L, Baugher M. (1993). Psychiatric risk factors for adolescent suicide: a case-control study. Journal of the American Academy of Child and Adolescent Psychiatry, 32(3), 521-529.

Brent, D. (1995). Risk factors for adolescent suicide and suicidal behavior: mental and substance abuse disorders, family environmental factors, and life stress. Suicide and Life Threatening Behavior, 25 Supplement, 52-63.

Brindis, C., Wolfe, A., McCarter, V., Ball, S. & Starbuck-Morales, S. (1995). The Associations Between Immigrant Status and Risk-Behavior Patterns

115

in Latino Adolescents. Journal of Adolescent Health, 17(2), 99-105. Brindis, C., Park, M., Ozer, E.,&Irwin, C. (2002). Adolescents’ access to health services and clinical preventive health care: Crossing the great divide. Pediatric Annals, 31(9), 575–581.

Brook,D., Brook J., Zhang, C., Cohen, P. & Whiteman, M. (2002). Drug use and the risk of major depressive disorder, alcohol dependence, and substance use disorders. Archives of General Psychiatry, 59(11), 1039-1044.

Brown, B. (1982). The extent and effects of peer pressure among high school students: A retrospective analysis. Journal of Youth and Adolescence, 11, 121-133 . Brown, B., Lohr, M. & McClenahan, E. (1985). Early adolescents’ perceptions of peer pressure. Journal of Early Adolescence, 6, 139-154.

Brown, B. (1989). The role of peer groups in adolescents’ adjustment to secondary school. In Berndt, T & Ladd, G. (Eds.). Peer Relationships in Child Development. New York: Wiley & Sons, Ltd., 188-215.

Bureau of Labor Statistics and Bureau of the Census. (2006) Annual Demographic Survey- Health Insurance Coverage Status by Nativity, Citizenship, and Duration of Residence for Hispanic Population. Retrieved on October 24, 2006 from http://pubdb3.census.gov/macro/032004/health/h09a_000.htm

Burnam, M., Hough, R., Karno, M., Escobar, J. & Telles, C. (1987).Acculturation and lifetime prevalence of psychiatric disorders among Mexican Americans in Los Angeles. Journal of Health and Social Behaviors, 28, 89-102.

Cabassa, L., Lester, R. & Zayas, L. (2007). "It's like being in a labyrinth:" Hispanic immigrants' perceptions of depression and attitudes toward treatments. Journal of Immigrant and Minority Health, 9(1), 1-16.

Cameron, A.C. & Trivedi, P.K. (1998). Regression Analysis of Count Data, Cambridge University Press.

Campbell, N.B., Milling. L., Laughlin, A., & Bush, E. (1993). The psychosocial climate of families with suicidal pre-adolescent children. American Journal of Orthopsychiatry,63(1), 142–145.

Caplan, S. (2007). Latinos, acculturation, and acculturative stress: a dimensional concept analysis. Policy Politics and Nursing Practice, 8(2):93-106.

Cassel, J. (1974). Psychosocial processes and “stress”: Theoretical formulations. International Journal of Health Services, 4, 471-482.

116

Cauce, A. (1986). Social networks and social competence: Exploring the effects of early adolescent friendships. American Journal of Community Psychology, 14, 607-628.

Census. (1998). Current Population Survey. March 1998.

Census. (1999). The statistical abstract of the united states, 1996. Retrieved on June 30, 2009 from http://www.census.gov/prod/2/gen/96statab/96statab.html

Census. (2000). U.S. Census 2000. Retrieved March 16, 2008, from http://www.census.gov/main/www/cen2000.html

Census. (2001). Economic Challenges Wane With Time Among The Nation's Foreign- Born Population. Retrieved on June 19, 2009 from http://www.census.gov/Press- Release/cb97-55.html

Census. (2002). Demographic Trends in the 20th Century: Census 2000 Special Reports. Retrieved November 22, 2006 from http://www.census.gov/prod/2002pubs/censr- 4.pdf

Census. (2004). Facts and Figures. Retrieve on May 20, 2009 From http://www.census.gov/PressRelease/www/releases/archives/facts_for_features_s pecial_editions/002270.html

Census. (2006a). Income, Poverty, and Health Insurance Coverage in the United States: 2005. Retrieved on October 25, 2006 from http://www.census.gov/prod/2006pubs/p60-231.pdf

Census. (2006b). Health Status, Health Insurance, and Health Services Utilization: 2001. Retrieved on October 24, 2006 from http://www.census.gov/prod/2006pubs/p70-106.pdf

Census. (2008). Hispanic Population in the United States. Retrieved on May 20, 2009 from http://www.census.gov/population/www/socdemo/hispanic/hispanic_pop_present ation.html

Centers for Disease and Control. (1999). The Surgeon General’s Call to Action to Prevent Suicide 1999. Centers for Disease Control website. Retrieved on January 9, 2006 from http://www.surgeongeneral.gov/library/calltoaction/calltoaction

117

Centers for Disease Control. (2001). APPENDIX A: Selected Healthy People 2010 Objectives Related to Child and Adolescent Unintentional Injury, Violence, and Suicide Prevention. Retrieved on August 4, 2009 from http://www.cdc.gov/mmwr/preview/mmwrhtml/rr5022a2.htm

Centers for Disease Control. (2005). Health, United States: with Chartbook on Trends in the Health of Americans. Retrieved on October 20, 2006 from http://www.cdc.gov/nchs/data/hus/hus05.pdf

Centers for Disease Control. (2006). Overview of the Youth Risk Behavior Surveillance System. Retrieved on January 5, 2007 from http://www.cdc.gov/HealthyYouth/yrbs/pdf/2005YRBSOverview.pdf

Centers for Disease Control. (2008). Youth Risk Behavior Surveillance --- United States, 2007. Morbidity and Mortality Weekly Report, 57 (SS04), 1-131.

Center for Substance Abuse Treatment. (2002). Adolescent Marijuana Use by Race/Ethnicity in Three National Surveys. Retrieved on June 20, 2009 from: http://www.icpsr.umich.edu/SAMHDA/NTIES/NTIES- PDF/SHEETS/122_marijuana_use.pdf

Cervantes, R. C., Salgado de Snyder, V. N., & Padilla, A. M. (1989).

Posttraumatic stress in immigrants from Central America and Mexico. Hospital and Community Psychiatry, 4, 615–619.

Chantala, K. & Tabor, J. (1999). Strategies to Perform a Design-Based Analysis Using the Add Health Data. Carolina Population Center University of North Carolina at Chapel Hill.

Chantala, K. (2001). Introduction to analyzing Add Health data. Retrieved on March 12, 2007 from www.cpc.unc.edu/projects/addhealth/files/analyze.pdf

Choi, H., Meininger, J. & Roberts, R. (2006). Ethnic Differences in Adolescents’ Mental Distress, Social Stress and Resources. Adolescence, 42(162), 263-283.

Cohen, J., Cohen, P., West, S. & Aiken, L. (2002). Applied multiple regression/correlation analyses for the behavioral sciences (3rd Edition). Lawrence Erlbaum: New York.

Coie, J. &Kupersmidt, J. (1983). A behavioral analysis of emerging social status in boys’ groups. Child Development, 54, 1400-1416.

Cox, H. (1974). Theoretical Statistics. Chapman & Hall: New York.

118

Crosnoe, R., Johnson, M. K., & Elder, Jr., G. H. (2004). Intergenerational Bonding in School: The Behavioral and Contextual Correlates of Student-Teacher Relationships. Sociology of Education, 77(1), 60-81.

D’Atillio, J., Campbell, B., Lubold, P., Jacobson, T. & Richard, J. (1987). Social support and suicide potential: Preliminary findings for adolescent populations. Psychological Reports, 70, 1, 76-78.

De Wilde EJ. (2000). Adolescent suicidal behaviour: a general population perspective. In: Hawton K, van Heeringen K, editors. The international handbook of suicide and attempted suicide. Chichester: John Wiley & Sons Ltd.

Debold, E., Wilson, M., & Malave´, I. (1993). Mother daughter revolution. New York: Bantam Books.

Delgado, L. (1995, March-April). Meeting the health promotion needs of Hispanic communities. American Journal of Health Promotion. 9(4), 300-11.

Department of Health and Human Services. (2000). Healthy people 2010: leading health indicators. Retrieved on October 19, 2008 from http://www.healthypeople.gov.proxy.lib.ohiostate.edu/document/html/uih/uih_bw /uih_4.htm#mentalhealth.

Department of Health and Human Services. (2001). Mental health: culture, race, and ethnicity. a supplement to mental health: a report of the Surgeon General. Washington, DC: Department of Health and Human Services. U.S. Public Health Service.

Diamond, A. & Muntz, R. (1967). Ordinal position of birth and self disclosure in high school students. Psychological Reports, 21, 829-833.

Dodge, K. (1983). Behavioral antecedents of peer social status. Child Development, 54, 1383-1385.

Dornbusch, S., Ritter, P., Leiderman, H., Roberts, D., Fraleigh, M. (1987). The Relation of Parenting Style to Adolescent School Performance. Child Development, 58, 5, 1244-1258.

Douglas-Hall, A., & Koball, H. (2004). Children of recent immigrants: National and regional trends: National Center for Children in Poverty. Columbia University. Mailman School of Public Health.

119

Drake, R., Gates, C. & Cotton, P. (1986). Suicide among schizophrenics: A comparison of attempters and completed suicides. British Journal of Psychiatry, 149, 784- 787.

Douvan, E. & Adelson, J. (1966). The Adolescent Experience. New York: Wiley & Sons, Ltd.

Duberstein, P., Conwell, Y. Caine, E. (1993): Interpersonal stressors, substance abuse, and suicide”. Journal of Nervous and Mental Disease, 181(2), 80-85.

Durkheim, E. (1951). Suicide. New York: Free Press. (Original work published 1897).

Ebin, V., Sneed, C., Morisky, D., Rotheram-Borus, M., Magnusson, A. & Malotte, C. (2001). Acculturation and interrelationships between problem and health promoting behaviors among Latino adolescents. Journal of Adolescent Health, 28(1), 62-72.

Escobar, J., Hoyos, N. & Gara, M. (2000). Immigration and Mental Health: Mexican Americans in the United States. Harvard Review Psychiatry, 8,2, 64-72.

Espinosa, M. (1998). School Involvement and Hispanic Parents. The Prevention Researcher. Winter, 5.

Falicov, C. J. (1996). Mexican families. In M. McGoldrick, J. Giordano, & J. K. Pearce (Eds.),Ethnicity and family therapy. (2nd ed., pp. 169–182). New York: Guilford Press.

Feigelman, W., & Gorman, B. S. (2008). Assessing the Effects of Peer Suicide on Youth Suicide. Suicide and Life-Threatening Behavior. Retrieved September 25, 2008, from http://www3.ncc.edu/faculty/soc/feigelb/friendsuicide2.pdf

Fergusson, DM., Woodward, LJ., Horwood, LJ. (2000). Risk factors and life processes associated with the onset of suicidal behaviour during adolescence and early adulthood. Psychological Medicine, 30, 23-29.

Fishman, J. (1966). Language loyalty in the United States; the maintenance and perpetuation of non-English mother tongues by American ethnic and religious groups. Mouton & Company: New York.

Flores, G., Abreu, M., Olivar, M., & Kastner, B. (1998). Access barriers to health care for Latino children. Archives of Pediatric and Adolescent Medicine, 152, 1119– 1125.

120

Garafolo,R., Wolf, C., Wissow, L., Woods, E. & Goodman, E. (1999). Sexual orientation and risk in suicide attempts among a representative sample of youth. Archives of Pediatric and Adolescent Medicine, 153, 487-493.

Garcia, C. & Saewyc, E. (2007). Perceptions of mental health among recently immigrated Mexican adolescents. Issues in Mental Health Nursing, 28(1), 37-54.

Garcia-Preto, N. (1996). Latino families: An overview. In McGoldrick, Monica Ed); Giordano, Joe (Ed); Pearce, John K. (Ed). Ethnicity and family therapy (2nd ed.), 141-154. New York, NY, US: Guilford Press.

Gardner, W., Mulvey, E. & Shaw, E.. (1995). Regression analyses of counts and rates: Poisson, overdispersed poisson and negative binomial models. Psychological Bulletin, 118(3), 393-404.

Garrison, E., Roy, I. & Azar, V. (1999). Responding to the mental health needs of Latino children and families through school-based services. Clinical Psychology Review, 19, 2, 199–219.

Genero, N.P., Miller, J.B., Surrey, J., & Baldwin, L.M. (1992). Measuring perceived mutuality in close relationships: Validation of the Mutual Psychological Development Questionnaire. Journal of Family Psychology, 6(1), 36–48.

Giddens, A. (1964). Suicide and family constellation in India. American Journal of Psychotherapy, 37,2, 273-278.

Ginsberg, D., Gottman, J. & Parker, J. (1986). The importance of friendship, in Gottman, J. & Parker, J. (Eds.). Conversations of Friends: Speculations on Affective Development. Cambridge, MA: Cambridge University Press, 3-48.

Gonzalez-Ramos, G.,& Sanchez-Nester,M. (2001). Responding to immigrant children’s mental health needs in the schools: ProjectMiTerra/My Country. Children & Schools, 23(1), 49–62.

Gould, M., Wallenstein, S. & Kleinman, M. (1990). Time-space clustering of teenage suicide. American Journal of Epidemiology, 131,1, 71-78

Gould, M. S., Fisher, P., Parides, M., Flory, M., & Shaffer, D. (1996). Psychosocial risk factors of child and adolescent completed suicide. Archives of General Psychiatry, 53, 1155–1162.

121

Gould, M., King, R., Greenwald, S., Fisher, P., Schwab-Stone, M., Kramer, R., Flisher, A., Goodman, S., Canino, G. & Shaffer, D. (1998). Psychopathology Associated with Suicidal Ideation and Attempts Among Children and Adolescents. American Academy of Child and Adolescent Psychiatry, 37, 9, 915-923.

Granillo, T., Jones-Rodriguez, G., & Carvajal, S. C. (2005, March). Prevalence of Eating Disorders in Latina Adolescents: Associations with Substance Use and Other Correlates. Journal of Adolescent Health, 36(3), 214-220.

Greenblatt, J. (1998). Adolescent self-reported behaviors and their association with marijuana use. National Household Survey on Drug Abuse, 1994-1996, SAMSHA.

Greig, R. (2003). Ethnic identity development: Implications for mental health in African- American and Hispanic adolescents. Issues in Mental Health Nursing, 24(3), 317– 331.

Griffith, J. (1984). Emotional support providers and psychological stress among Anglo- and Mexican Americans. Community Mental Health Journal, 20(3), 182-201.

Groves, S., Stanley, B. & Sher, L. (2007). Ethnicity and the relationship between adolescent alcohol use and suicidal behavior. International Journal of Adolescent Medical Health, 19(1), 19-25.

Guiao, I. & Esparza, D. (1995). Suicidality correlates in Mexican American teens. Issues in Mental Health Nursing, 16, 5, 461-479.

Hallfors, D., Waller, M. W., Ford, C. A., Halpern, C. T., Brodish, P. H., & Iritani, B. (2004). Adolescent Depression and Suicide Risk: Association with Sex and Drug Behavior. American Journal of Preventive Medicine, 27(3), 224-231.

Hanneman, R. & Riddle, M. (2005). Introduction to social network methods. Riverside, CA: University of California, Riverside (published in digital form at http://faculty.ucr.edu/~hanneman/ )

Haynie, D. L., South, S. J., & Bose, S. (2006). Adolescent Mobility and Suicide Attempts. Sociological Quarterly, 47, 693-721.

Harris, Kathleen Mullan, Carolyn Tucker Halpern, Pamela Entzel, Joyce Tabor, Peter S. Bearman, and J. Richard Udry. 2008. The National Longitudinal Study of Adolescent Health: Research Design [WWW document]. URL: http://www.cpc.unc.edu/projects/addhealth/design.

122

Hilbe, J. (2007). Negative Binomial Regression. Cambridge, UK: Cambridge University Press.

Hines, A. & Caetano, R. (1998). Alcohol and AIDS-related sexual behavior among Hispanics: acculturation and gender difference. AIDS Education and Prevention, 10(6), 533-547.

Hirsh, B. (1978). Social networks and the coping process: Creating personal communities. In Social networks and social support. Gottlieb, B. (Ed.). Beverly Hills: Sage., 149-170.

Hoagwood, K., Burns, B., Kiser, L., Ringeisen, H. & Schoenwald, S. (2001). Evidenced-based practice in child and adolescent mental health services. Psychiatric Services, 52(9), 1179-1189.

Hosmer, D. & Lemenshow, S. (2000). Applied Logistic Regression, 2nd Edition. Wiley-Interscience Publication.

Hovey, J. (1996). Acculturative stress, depression and suicidal ideation among immigrant and second-generation Latino adolescents. Journal of American Academy of Child and Adolescent Psychiatry, 35(9), 1183-1192.

Hovey, J., & King, C. (1996). Acculturative stress, depression, and suicidal ideation among immigrant and second-generation Latino adolescents. Journal of the American Academy of Child & Adolescent Psychiatry, 35(9), 1183–1192.

Hovey, J. (1998). Acculturative stress, depression and suicidal ideation among Mexican- American adolescents: implications for the development of suicide ideation programs in schools. Psychological Reports, 83(1), 249-250.

Hovey, J. D. (1999). The moderating influence of social support on suicidal ideation in a sample of Mexican immigrants. Psychological Reports, 85, 78-79.

Hovey, J. (2000). Acculturative stress, depression and suicidal ideation among Central American immigrants. Suicide and Life-Threatening Behavior, 30(2), 125-139.

Hoyert, D. L., Kochanek, K. D., & Murphy, S. L. (1999). Deaths: Final data for 1997. National Vital Statistics Reports, 47(9). Hyattsville, MD: National Center for Health Statistics.

Hunsaker, A. (1985). Chicano Drug Abuse Patterns: Using Archival Data to Test Hypotheses. Hispanic Journal of Behavioral Sciences, 7(1), pp. 93-104.

123

Hunter, F. & Youniss, J. (1982). Changes in the function of three relations during adolescence. Developmental Psychology, 18, 806-811.

Institute of Medicine. (2001). Health and Behavior: The interplay of biological, behavioral and societal influences. Washington, DC: National Academy Press.

Irwin, R. (1991). The newest North Americans: Why we need to help them. [Monograph]. Los Angeles: National Immigration Law Center.

Jablensky, A. (1994). Refugee mental health and well-being: Conclusions and recommendations. In A. J. Marsella, T. Bornemann, S. Ekblad, & J. Orley (Eds.), Amidst peril and pain: The mental health and well-being of the world’s refugees (pp. 327–339). Washington, DC: American Psychological Association.

Jellinek, M. & Snyder, J. (1998). Depression and Suicide in Children and Adolescents. Pediatrics in Review, 19, 255-264. Joffe RT., Offord DR. & Boyle MH. (1988). Ontario Child Health Study: suicidal behavior in youth age 12-16 years. American Journal of Psychiatry, 145, 1420- 1423. Joiner, T. (2002). The Clustering and Contagion of Suicide. Current Directions in Psychological Science, 8, 3, 89-92.

Jordan, J. (1998). Toward Connection and Competence. Wellesley, MA: Stone Center Publications.

Judd, C. M., & Kenny, D. A. (1981). Process analysis: Estimating mediation in treatment evaluations. Evaluation Review, 5, 602-619.

Kalsbeek, W., Yang, J. & Agans, R. (2002, August). Predictors of Nonresponse in a longitudinal survey of adolescents. Presented at the Annual Meeting of the American Statistical Association-Joint Statistical Meetings, New York, New York.

Kandel, D. (1978). Similarity in real-life adolescent friendship pairs. Journal of Personality and Social Psychology, 36, 306-312.

Kandel, D., Kessler, R. & Margulies, R. (1978). Antecedents of adolescent initiation into stages of drug use: a developmental analysis. Journal of Youth and Adolescence, 7, 13-40.

Karno, M. (1994). The prevalence of mental disorder among persons of Mexican birth or origin. In C. Telles & M. Karno (Eds.), Latino mental health: Current research and policy perspectives (pp. 3–16). Los Angeles, CA: University of California Neuropsychiatric Institute.

124

Kaslow, N., Thompson, M., Meadows, L., Jacobs, D., Chance, S., Gibb, B., Bornstein, H., Hollins, L., Rashid, A. & Phillips, K. (1998). Factors that mediate and moderate the link between partner abuse and suicidal behavior in African American women. Journal of Consulting and Clinical Psychology, 66, 3, 533- 540.

Keenan, K. Loeber, R. Zhang, Q. Stouthamer-Loeber, M. Van Kammen, W. B. (1995). The influence of deviant peers on the development of boys' disruptive and delinquent behavior: A temporal analysis. Development and Psychopathology, 7, 4, 715-726.

Keller, T., & Gray, D. (2005). Utah youth suicide study: psychological autopsy. Suicide and Life-Threatening Behavior, 35(5), 536-46.

Kessler, R., Berglund, P., & Demler, O. (2005). Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry. 62(6), 593–602.

Kidd, S., Henrich, C. C., Brookmeyer, K. A., Davidson, L., King, R. A., & Shahar, G. (2006, August). The social context of adolescent suicide attempts: Interactive effects of parent, peer, and school social relations. Suicide and Life-Threatening Behavior, 36(4), 386-395.

Kotler, M., Iancu, I., Efroni, R. & Amir, M. (2001). Anger, impulsivity, social support and suicide risk in patients with posttraumatic stress disorder. Journal of Nervous and Mental Disease, 189, 3, 162-167.

Kovacs, M., Goldstein, D., & Gatsonis, C. (1993). Suicidal ideation and childhood-onset depressive disorders: A longitudinal investigation. Journal of the American Academy of Child & Adolescent Psychiatry, 32, 8–20.

Ku, L., & Matani, S. (2001). Left out: Immigrants’ access to health care and insurance. Health Affairs, 20(1), 247–256.

Landau, J., Griffiths, J. A., & Mason, J. (1981). The extended family in transition: Clinical implications. Psychotherapeia,7, 370–381.

Larson, R. (1983). Adolescents’ daily experience with family and friends: Contrasting opportunity systems. Journal of Marriage and Family, 45, 739-750.

Leone, T. (n.d.). Suicide Prevention: A Strategy for the Nation. Retrieved from The University of South Florida Louis de la Parte Florida Mental Health Institute website on November 20, 2006 from http://www.fmhi.usf.edu/institute/pubs/newsletters/policybriefs/issue013.pdf

125

Lessard, J. & Moretti, M. (1998). Suicidal ideation in an adolescent clinical sample: attachment patterns and clinical implications. Journal of Adolescence, 21(4), 383-395.

Lewinsohn, P., Rohde, P. & Seeley, J. (1996).Adolescent Suicidal Ideation and Attempts: Prevalence, Risk Factors, and Clinical Implications. Clinical Psychology: Science and Practice, 3, 1, 25-46.

Lin, N., Simeone, R., Ensel, W. & Kuo, W. (1979). Social support, stressful life events, and illness: A model and an empirical test. Journal of Health and Social Behavior, 20, 108-119.

Linehan, M. M., Heard, H. L., & Armstrong, H. E. (1993). Naturalistic follow-up of a behavioral treatment for chronically parasuicidal borderline patients. Archives of General Psychiatry, 50, 971–974.

Long, J. (1997). Regression Models for Categorical and Limited Dependent Variables (Advanced Quantitative Techniques in the Social Sciences). Sage: New York.

Lowenthal, M. & Haven, C. (1968). Interaction and adaptation: Intimacy as a critical variable. American Sociological Review, 33, 20-30.

Lynskey, M. (2004). Major depressive disorder, suicidal ideation, and suicide attempt in twins discordant for cannabis dependence and early-onset cannabis use. Archives of General Psychiatry, 61, 1026-1032.

Lyon, M., Benoit, M., O'Donnell, R., Getson, P., Silber, T. & Walsh, T. (2000). Assessing African American Adolescents’ Risk for Suicide Attempts: Attachment Theory. Adolescence,35(137), 121-134

McGarvey, E,, Kryzhanovskaya, L., Koopman, C., Waite, D. & Canterbury, R. (1999). Incarcerated adolescents' distress and suicidality in relation to parental bonding styles. Crisis, 20(4), 164-170.

McGee, R., Feehan, M., Williams, S., Partridge, F., Silva, P. A., & Kelly, J. (1990). DSM-III disorders in a large sample of adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 29, 611–619.

Manaster, G., Chan, J., & Safady, R. (1992). Mexican-American migrant students’ academic success: Sociological and psychological acculturation. Adolescence, 27(105), 123–136.

Marin, B. & Flores, E. (1994). Acculturation, sexual behavior, and alcohol use among Latinas. International Journal of Addiction, 29(9), 1101-1114.

126

Marin, B., Gomez, C., Tschann, J., Gregorich, S. (1997). Condom use in unmarried Latino men: A test of cultural constructs. Healthy Psychology, 16(5), 458-467.

Martin, G. & Waite, S. Parental bonding and vulnerability to adolescent suicide. Acta Psychiatry Scandinavia, 89(4), 246-54.

Maternal & Child Health Bureau. (1999). Youth Violence Prevention in Latino Communities: A Resource Guide for MCH Professionals. Newton MA: Children’s Safety Network, Education Development Center, Inc.

Medina, C & Luna, G. (2006). Suicide attempts among adolescent Mexican American students enrolled in special education classes. Adolescence, 41(162), 299-312.

Milardo, R. (1982). Friendship Networks in Developing Relationships: Converging and Diverging Social Environments. Social Psychology Quarterly, 45(3), 162-172.

Miller, P. & Ingham, J. (1976). Friends, confidants and symptoms. Social Psychiatry, 11, 51-58.

Mindel, C. H. (1980). Extended familism among urban Mexican Americans, Anglos, and Blacks. Hispanic Journal of Behavioral Sciences, 2, 21–34.

Mitchell, R. & Trickett, E. (1980). Social network research and psychological adaptation: Implications for community mental health practice. In P. M. Insel (Ed.). Environmental Variables and the Prevention of Mental Illness, Lexington, MA. (43-68).

Moskos, M., Olson, L., Halbern, S., Keller, T., & Gray, D. (2005). Utah youth suicide study: psychological autopsy. Suicide and Life-Threatening Behavior, 35(5), 536- 46.

National Drug Control Policy. (2007). Hispanic Teens & Drugs: A Special Report- Office of National Drug Control Policy. Retrieved on September 27, 2007 from http://www.mediacampaign.org/hispnc_yth/HispanicTeens_and_Drugs.pdf

National Institute of Mental Health. (2003). In Harm’s Way. Retrieved October 20, 2006 from www.nimh.nih.gov/publicat/schizkids.cfm

National Mental Health Association (2006). Fact Sheet: Suicide. Retrieved October 20, 2006 from www.nmha.org/infoctr/factsheets/81.cfm

Ng, B. (1996). Characteristics of 61 Mexican American Adolescents Who Attempted Suicide. Hispanic Journal of Behavioral Sciences, 18(1), 3-12.

127

Nisbet, P. (1996). Protective factors for suicidal black females. Suicide and Life- Threatening Behavior, 26, 4 325-341.

O’Brien, S. & Bierman, K. (1988). Conceptions and perceived influence of peer groups: Interviews with pre-adolescents and adolescents. Child Development, 59, 1360- 1365.

O’Donnell, L., Stueve, A., Wardlaw, D. & O’Donnell, C. (2003). Adolescent suicidality and adult support: the reach for health study of urban youth. American Journal of Health Behavior, 27(6), 633-644.

O’Donnell, L., O’Donnell, C., Wardlaw, D. & Stueve, A. (2004). Risk and resiliency factors influencing suicidality among urban African American and Latino youth. American Journal of Community Psychology, 31(102), 37-49.

O’Grady, D. & Metz, J. (1987). Resilience in children at high risk for psychological disorder. Journal of Pediatric Psychology, 12, 1, 3-23.

Oetting, E.R. (1993). Orthogonal cultural identification: theoretical links between cultural identification and substance use. NIDA Research Monograph [NIDA Research Monograph] 130, 32-56.

Olvera, R. (2001). Suicidal Ideation in Hispanic and Mixed-Ancestry Adolescents. Suicide and Life-Threatening Behavior, 31(4), 416-427.

Oquendo, M., Lizardi,D., Greenwald, S., M. M. Weissman, M. & Mann, J. (2004) Rates of lifetime suicide attempt and rates of lifetime major depression in different ethnic groups in the United States Acta Psychiatrica Scandinavica 110 (6) , 446– 451.

Padilla, A., & Perez, W. (2003). Acculturation, social identity, and social cognition: A new perspective. Hispanic Journal of Behavioral Sciences, 25, 35–55.

Parker, J. (1986). Becoming friends: Conversational skills for friendship formation in young children. In Gottman, J. & Parker, J. (Eds). Conversations of friends: Speculations on Affective Development. Cambridge, MA: Cambridge University Press, 103-138.

Pena, J., Wyman, P. Brown, H., Matthieu, M., Olivares, T., Hartel, D. & Zayas, L. (2008). Immigration Generation Status and its Association with Suicide Attempts, Substance Use, and Depressive Symptoms among Latino Adolescents in the USA. Prevention Science, 9, 4, 299-310.

128

Perez-Smith, A., Spirito, A., & Boergers, J. (2002). Neighborhood predictors of hopelessness among adolescent suicide attempters: preliminary investigation. Suicide and Life-Threatening Behavior, 32(2), 139-45.

Pfeffer, C. R., Klerman, G. L., Hurt, S. W., & Kakuma, T. (1993). Suicidal children grow up: Rates and psychosocial risk factors for suicide attempts during follow-up. Journal of the American Academy of Child and Adolescent Psychiatry, 32, 106- 113.

Plante, T. J., Manuel, G. M., Menendez, A. V., & Marcotte, D. (1995). Coping with stress among Salvadoran immigrants. Hispanic Journal of Behavioral Sciences, 17, 471–479.

Portes, P., & Zady, M. (2002). Self-esteem in the adaptation of Spanish-speaking adolescents: The role of immigration, family conflict, and depression. Hispanic Journal of Behavioral Sciences, 24(3), 296–318.

Potter LB., Kresnow M., Powell KE. (2001). The influence of geographic mobility on nearly lethal suicide attempts. Suicide Life Threatening Behavior, 32 (1 Suppl.): 42-8.

Prinstein, M., Boergers, J., Spirito, A. (2001). Adolescents’ and their friends health risk behavior: Factors that alter or add to peer influence. Journal of Pediatric Psychology, 26, 5, 287-298.

Queralt, M. (1993a). Risk Factors Associated with Completed Suicide in Latino Adolescents. Adolescence, 28(112), 831-851.

Queralt, M. (1993b). Psychosocial Risk Factors Associated with Suicide in a Small Community Sample of Latino Adolescent Attempters. Social Work in Education, 15(2), 91-103.

Radloff, L. (1977). The CES-D Scale. Applied Psychological Measurement, 1(3), 385- 401.

RAND. (2001). Research highlights: Mental health care for youth. Retrieved on November 14, 2008 from http://www.rand.org/pubs/research_briefs/RB4541/index1.html

Rasmussen, K., Negy, C., Carlson, R. & Burns, J. (1997). Suicidal Ideation and Acculturation Among Low Socioeconomic Status Mexican American Adolescents. Journal of Early Adolescence, 17(4), 390-407.

129

Razin, A., O’Dowd, M., Nathan, A., Goldfield, A., Martin, C., Goulet, L., Scheftel, S., Mezan, P. & Mosca, J. (1991). Suicidal Behavior Among Inner-City Hispanic Adolescent Females, 13(1), 45-58.

Reis, H., Senchak, M. & Solomon, B. (1985). Sex differences in intimacy of social interaction: Further potential explanations. Journal of Personal and Social Psychology, 48, 1204-1217.

Reisman, J. (1985). Friendship and its implications for mental health or social competence. Journal of Early Adolescence, 5, 383-391.

Resnick, M., Bearman, P., Blum, R., Bauman, K., Harris, K., Jones, J., Tabor, J., Beuhring, T., Sieving, R., Shew, M., Ireland, M., Bearinger, L. & Udry, J. (1997). Protecting adolescents from harm. Findings from the National Longitudinal Study on Adolescent Health. Journal of the American Medical Association, 278, 10.

Rhodes, J. E., Contreras, J. M., & Mangelsdorf, S. C. (1994). Natural mentor relationships among Latina adolescent mothers: Psychological adjustment, moderating processes, and the role of early parental acceptance. American Journal of Community Psychology, 22, 211-227.

Roberts, R., & Chen, Y. (1995). Depressive symptoms and suicidal ideation among Mexican-origin and Anglo adolescents. Journal of the American Academy of Child & Adolescent Psychiatry, 34(1), 81–90.

Robles, F. (1995,December 18). Concern mounts over middle school suicides. Miami Herald, p. 4B.

Rogler, L. H., Malgady, R. G., & Rodriguez, O. (1989). Hispanics and mental health: A framework for research. Melbourne, FL: Krieger.

Rotheram-Borus, M.J., Trautman, P.D., Dopkins, S.C. & Shrout, P.E. (1990). Cognitive style and pleasant activities among female suicide attempters. Journal of Consulting and Clinical Psychology, 58(5), 554–561.

Royston, P., Carlin, J. & White, I.(In press, 2009) Multiple imputation of missing values: new features for mim. STATA Journal.

Rubenstein, J., Heeren, T., Housman, D., Rubin, C. & Stechler, G. (1989). Suicidal behavior in “normal” adolescents: Risk and protective factors. American Journal of Orthopsychiatry, 59, 1, 59-71.

130

Rubin, K. (1983). Recent perspectives on social competence and peer status: Some introductory remarks. Child Development, 54, 1383-1385.

Rubin, D. (1987). Multiple imputation for nonresponse in surveys. New York: John Wiley.

Russell, S. T., & Joyner, K. (2001). Adolescent Sexual Orientation and Suicide Risk: Evidence from a National Study. American Journal of Public Health, 91(8), 1276- 1281.

Sabbath J. (1969). The suicidal adolescent--The expendable child. Journal of the American Academy of Child Psychiatry, 8, 272-289.

Savin-Williams, R. (1980). Social interactions of adolescent females in natural groups. In Foot, H., Chapman, A. & Smith, J. (Eds.). Friendship and Social Relations in Children. New York: John Wiley & Sons, Ltd., 343-364.

Schafer, J. (1999). Multiple imputation: a primer. Statistical Methods in Medical Research, 8, 3-15

Selman, R. & Shultz, L. (1990). Making a friend in youth. Chicago, IL: University of Chicago Press.

Shaffer, D., & Gould, M. (1987). Study of completed and attempted suicides in adolescents. Progress Report: Department of Health and Human Services, Public Health Service, National Institute of Mental Health, Washington, DC.

Shaffer, D., Gould, M. S., Fisher, P., Trautment, P., Moreau, D., Kleinman, M., & Flory, M. (1996). Psychiatric diagnosis in child and adolescent suicide. Archives of General Psychiatry, 53, 339–348.

Shaffer, D., & Craft, L., (1999). Methods of adolescent suicide prevention. Journal of Clinical Psychiatry, 60 (Suppl. 2), 70–74.

Sharbany, R., Gershoni, R. & Hofman, J. (1981). Girlfriend, boyfriend: Age and sex differences in intimate friendship. Developmental Psychology, 17, 800-808.

Shen, B. & Tackeushi, D. (2001). A structural model of acculturation and mental health status among Chinese Americans. American Journal of Community Psychology, 29,3, 387-418.

Shrier,L., Harris, S., Sternberg, M. & Beardslee, W., (2001). Associations of Depression, Self-Esteem, and Substance Use with Sexual Risk among Adolescents. Preventive Medicine, 33, 179–189.

131

Silenzio, V. M., Pena, J. B., Duberstein, P. R., Cerel, J., & Know, K. L. (2007). Sexual Orientation and Risk Factors for Suicidal Ideation and Suicide Attempts Among Adolescents and Young Adults. American Journal of Public Health, 97(11), 2017-2019.

Smokowsi, P., Buchanan, R. & Bacallao, M. (2009). Acculturation and adjustment in Latino adolescents: how cultural risk factors and assets influence multiple domains of adolescent mental health. Journal of Primary Prevention, 3(4), 371- 393.

Smollar, J. & Youniss, J. (1983). Parent-adolescent relations in adolescents whose parents are divorced. Journal of Early Adolescence, 5, 129-144.

Smucker, M. R., Craighead, W. E., Craighead, L. W., & Green, B. J. (1986). Normative and reliability data for the Children’s Depression Inventory. Journal of Abnormal Child Psychology, 14, 25–39.

Snyder, J., Dishion, T. & Patterson, G. (1986). Determinants of associating with deviant peers during preadolescence. Journal of Early Adolescence, 6, 29- 43.

Sorenson, S. B., & Golding, J. M. (1988). Suicide ideation and attempts in Hispanics and non-Hispanic whites: Demographic and psychiatric disorder issues. Suicide and Life-Threatening Behavior, 18, 205-218.

Spirito, A. & Esposito-Smythers, C. (2006). Attempted and completed suicide in adolescence. Annual Review of Clinical Psychology, 2, 237-266.

Stapleton, L. M. (2006). An assessment of practical solutions for structural equation modeling with complex sample data. Structural Equation Modeling,13, 28-58

StataCorp. 2007. Stata Statistical Software: Release 10. College Station, TX: StataCorp LP.

Steinberg, L. & Silverberg, S. (1986). The vicissitudes of autonomy in early adolescence. Child Development, 57, 841-851.

Stokes, J. (1983). Predicting satisfaction with social support from social network structure. American Journal of Community Psychology, 11, 141-152.

Substance Abuse and Mental Health Services Administration. (n.d). Suicide: Cost to the Nation. Retrieved August 1, 2009 from http://mentalhealth.samhsa.gov/suicideprevention/costtonation.asp

132

Substance Abuse and Mental Health Services Administration. (2000). Mental Health: a Report of the Surgeon General. Retrieved October 23, 2006 from http://mentalhealth.gov/features/surgeongeneralreport/chapter3/sec5.asp

Substance Abuse and Mental Health Services Administration. SAMHSA Office of Applied Studies. (2003). Risk of Suicide Among Hispanic Females Aged 12 to 17. Retrieved from www.oas.samhsa.gov/2k3/LatinaSuicide/LatinaSuicide.cfm

Suicide Prevention Resource Center. (2005). Evidence-based Practices in Suicide Prevention Program (EBPP). Retrieved on February 14, 2007 from http://www.sprc.org/featured_resources/ebpp/ebpp_factsheets.asp

Substance Abuse and Mental Health Services Administration. (SAMHSA). (2006a). National Strategy for Suicide Prevention: At a Glance-Suicide Among the Young. Retrieved on October 23, 2006 from http://mentalhealth.samhsa.gov/suicideprevention/young.asp

Substance Abuse and Mental Health Services Administration. (2006b). Suicidal Thoughts among Youths Aged 12 to 17 with Major Depressive Episode. Retrieved on May 20, 2009 from http://www.oas.samhsa.gov/2k5/suicide/suicide.htm.

Substance Abuse and Mental Health Services Administration. (2008). Office of Applied Studies, National Household Survey on Drug Abuse Advance Report # 18. Retrieved on June 22, 2009 from: http://www.oas.samhsa.gov/nhsda/ar18t020.htm#A1

Sullivan, H. (1953). The Interpersonal Theory of Psychiatry. New York: Norton and Co.

Swanson, J. W., Linskey, A. O., Quintero-Salinas, R., Pumariega, A. J., & Holzer, C. E. (1992). A binational school survey of depressive symptoms, drug use, and suicidal ideation. Journal of the American Academy of Child and Adolescent Psychiatry, 31, 669-678.

Szapocznik, J., & Truss, C. (1978). Intergenerational sources of conflict in Cuban mothers. In C. Monteil (Ed.), Hispanic families: Critical issues for policy and programs in the human services (pp. 41–65). Washington, DC: COSSMHO.

Tardy, C. (1985). Social support and measurement. American Journal of Community Psychology, 13, 187-202.

133

Thompson, E., Eggert, L. & Herting, J. (2000). Mediating effects of an indicated prevention program for reducing youth depression and suicide risk behaviors. Suicide and Life-Threatening Behaviors, 30, 3, 252-271.

Tortolero, S & Roberts, R. (2001). Differences in Nonfatal Suicide Behaviors Among Mexican and European American Middle School Children. Suicide and Life- Threatening Behavior, 31(2), 214-223.

Trautman, E. (1961a). The suicidal fit: A psychobiotic study of Puerto Rican immigrants. Archives of General Psychiatry, 5, 98-105.

Trautman, E. C. (1961b). Suicide attempts of Puerto Rican immigrants. Psychiatric Quarterly, 35, 544- 545.

Trepper, T., Nelson, T., McCollum, E., McAvoy, P. (1997). Improving substance abuse delivery service to Hispanic women through increased cultural competencies: A qualitative study. Journal of Substance Abuse Treatment, 14, 225-234.

Tudor, C., Petersen, D. & Elifson, K. (1980). An examination of the relationship between peer and parental influences and adolescent drug use. Adolescence, 60, 783-798.

Turner, G., Kaplan, C., Zayas, L. & Ross, R. (2002). Suicide Attempts by Adolescent Latinas: An Exploratory Study of Individual and Family Correlates. Child and Adolescent Social Work Journal, 19(5), 357-374.

Unger, J., Ritt-Olson, A., Soto, D. & Baezconde-Garbanati, L. (2007). Parent-Child acculturation discrepancies as a risk factor for substance use among Hispanic adolescents in Southern California. Journal of Immigrant Minority Health.

U.S. Public Health Service. (1999). The Surgeon General's Call To Action To Prevent Suicide. Washington, DC.

Vance, J., Fernandez,G. &Biber, M. (1998). Educational progress in a population of youth with aggression and emotional disturbance: The role of risk and protective factors. Journal of Emotional and Behavioral Disorders. 6, 4, 214-221.

Vaux, A. (1985). Variations in social support associated with gender, ethnicity, and age. Journal of Social Issues, 41, 89–110.

Vaux, A. (1988). Social Support: Theory, research and intervention. New York: Praeger Publishers.

134

Veltman, C. (1983). Language Shift in the United States. Moulton and Company: New York.

Vega W., Kolody B. & Valle J. (1987). Migration and mental health: an empirical test of depression risk factors among immigrant Mexican women. International Migration Review, 21, 512–530.

Vega, W. A., Gil, A. G., Warheit, G. J., Apospori, E., & Zimmerman, R. (1993). The relationship of drug use to suicide ideation and attempts among African American, Hispanic, and White non-Hispanic male adolescents. Suicide and Life- Threatening Behavior, 23, 110-119.

Vega, W. A. Zimmerman, R. Gil, A. Warheit, G. J. (1994). Acculturation Strain Theory: Its Application in Explaining Drug Use Behavior Among Cuban and Other Hispanic Youth. NIDA Research Monograph, 130, 144-166

Vega, W. A., Zimmerman, G., Warheit, G., Jackson, D., Gill A., & Sokolkatz, J. (1994). The role of cultural factors in mental health problems of Hispanic adolescents. In C. Telles & M. Karno (Eds.). Latino mental health: Current research and policy perspectives (pp. 41–61). Los Angeles, CA: University of California Neuropsychiatric Institute.

Vega, W., Kolody, B., Aguilar-Gaxiola, S., Alderete, E., Catalano, R. & Caraveo- Anduaga, J. (1998). Lifetime prevalence of DSM-III-R psychiatric disorders among urban and rural Mexican Americans in California. Archives of General Psychiatry, 55, 771-778.

Vega, WA.; Sribney, WM.; Aguilar-Gaxiola, S.; Kolody, B. (2004). “12-month prevalence of DSM-III-R psychiatric disorders among Mexican Americans: nativity, social assimilation, and age determinants.” The Journal of Nervous and Mental Disease, 192(8), 532-41.

Wagner, B.M., Cole, R.E. & Schwarzman, P. (1995). Psychosocial correlates of suicide attempts among junior and senior high school youth. Suicide and Life- Threatening Behavior, 25(3), 358–372.

Wall, J., Power, T., & Arbona, C. (1993). Susceptibility to antisocial peer pressure and its relation to acculturation in Mexican-American adolescents. Journal of Adolescent Research, 8(4), 403–418.

Weissman, M. M., & Klerman, G. L. (1977). Sex differences and the epidemiology of depression. Archives of General Psychiatry, 34, 98–111.

135

Welfel, E. & R. Ingersoll, E. (2001). The Mental Health Desk Reference: A Practice- Based Guide to Diagnosis, Treatment, and Professional Ethics. New York : Wiley

Werner, E. & Smith, R. (1992). Overcoming the odds: High risk children from birth to adulthood. Ithaca, NY: Cornell University Press.

Whaley, A. & Francis, K. (2006). Behavioral health in multiracial adolescents: the role of Hispanic/Latino identity. Public Health Reports, 121(2), 169-174.

World Health Organization. (2008). World Suicide Prevention Day. Retrieved on February 29, 2008 from http://www.who.int/mediacentre/news/statements/2007/s16/en/index.html

Youniss, J. (1980). Parents and Peers in Social Development: A Sullivan-Piaget Perspective. Chicago, IL.: University of Chicago Press.

Youniss, J. & Haynie, D. (1992). Friendship in adolescence. Journal of Developmental Behavioral Pediatrics, 13, 1, 59-66.

Zayas, L. H. (1987). Toward an understanding of suicide risks in young Hispanic females. Journal of Adolescent Research, 2, 1-11.

Zayas, L. H., & Dyche, L. A. (1995). Ethnic issues in the treatment of Hispanic adolescent suicide attempters. In J. K. Zimmerman & G. M. Asnis (Eds.), Treatment approaches with suicidal adolescents (Vol. 12, Einstein Psychiatry Monograph Series, pp. 203-218). New York: John Wiley & Sons.

Zayas, L. H., Kaplan, C., Turner, S., Romano, K., & Gonzalez-Ramos, G. (2000). Understanding suicide attempts by adolescent Hispanic females. Social Work, 45, 53-63.

Zayas, L.H., Lester, R., Cabassa, L. & Fortuna, L. (2005). Why do so many Latina teens attempt suicide? A conceptual model of research. American Journal of Orthopsychiatry, 75(2), 275-287.

Zayas, L.H. &Pilat, A. (2008). Suicidal behavior in Latinas: explanatory cultural factors and implications for intervention. Suicide and Life Threatening Behavior, 38(3), 334-342.

136

Zimmerman, J.K. (1991). Crossing the desert alone: An etiological model of female adolescent suicidality. In C. Gilligan, A.G. Rogers, & D.L. Tolman (Eds.). Women, girls, and psychotherapy: Redefining resistance (pp. 223–240). New York: Haworth Press.

Zimmerman, J. K., & Zayas, L. H. (1995). Suicidal adolescent Latinas: Culture, female development, and restoring the mother-daughter relationship. In S. Canetto & D. Lester (Eds.), Women and suicide (pp. 120-132). New York: Springer.

137