10826084.2011.592432 .pdf



Nom original: 10826084.2011.592432.pdf
Titre: Culture and Context: Buffering the Relationship Between Stressful Life Events and Risky Behaviors in American Indian Youth
Auteur: Julie A. Baldwin1jbaldwin@health.usf.edu, Betty G. Brown2, Heidi A. Wayment3, Ramona Antone Nez2 and Kathleen M. Brelsford1

Ce document au format PDF 1.3 a été généré par dvips(k) 5.95a Copyright 2005 Radical Eye Software / Acrobat Distiller 6.0.1 (Windows), et a été envoyé sur fichier-pdf.fr le 24/09/2011 à 21:11, depuis l'adresse IP 174.89.x.x. La présente page de téléchargement du fichier a été vue 1256 fois.
Taille du document: 738 Ko (15 pages).
Confidentialité: fichier public




Télécharger le fichier (PDF)










Aperçu du document


Substance Use & Misuse, 46:1380–1394, 2011
C 2011 Informa Healthcare USA, Inc.
Copyright
ISSN: 1082-6084 print / 1532-2491 online
DOI: 10.3109/10826084.2011.592432

ORIGINAL ARTICLE

Culture and Context: Buffering the Relationship Between Stressful Life
Events and Risky Behaviors in American Indian Youth

Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

Julie A. Baldwin1 , Betty G. Brown2 , Heidi A. Wayment3 , Ramona Antone Nez2 and Kathleen
M. Brelsford1
1

Department of Community and Family Health, University of South Florida, Tampa, Florida, USA; 2 Department of
Health Sciences, Northern Arizona University, Flagstaff, Arizona, USA; 3 Department of Psychology, Northern Arizona
University, Flagstaff, Arizona, USA
The Sacred Mountain Youth Project (SMYP) sought to
address these issues through an in-depth investigation of
health risks and protective factors among rural AI youth
between the ages of 15 and 24 years.1 The project was
designed to identify social networks, support systems,
protective behaviors, and risk contexts of AI adolescents
residing in two boarding school dormitories located offreservation in the southwestern United States. The project
design combined systematic ethnographic data collection
with structured social network, behavioral, and psychosocial interviews to obtain data on peer networks, family
factors, cultural conditions, individual risk behaviors, and
protective factors for these adolescents in order to identify and measure the relationship of these various factors
to risk behaviors.

The Sacred Mountain Youth Project was conducted to
investigate risk and protective factors related to alcohol and drug use among American Indian youth. Findings indicated that stressful life events were positively
associated with depressed mood, substance use, and
risky behavior; cultural identity had no direct effects,
but a secondary model showed that social support and
protective family and peer influences were related to
cultural identity. These findings suggest that the relationships between stressors and their negative sequelae
are complex. Emphasis on protective processes that are
culturally specific to American Indian youth may lead
to effective alcohol and drug use prevention programs.
Keywords alcohol and substance use, risky behavior, American
Indian youth, cultural identity, depressed mood, risk factors,
protective factors

Ancestral History and Challenges

Changing social and cultural environments have had an
impact on the ways in which AI communities and individuals have maintained social relationships and cultural identity, perceived social support, and responded to
these influences. Although the AI/AN population comprises a diverse group of 562 tribal areas with different demographic characteristics and different health
issues (Federal Register, 2008), the AI/AN population is
united by an ancestral history of oppression, alienation,
and trauma. Loss of lands, loss of languages, genocide,
and forced alienation from cultural traditions have all
been a part of that history (Grandbois, 2005). For some
tribes, this history has contributed to significant community and family disintegration, as well as to unresolved

INTRODUCTION

Social scientists have long examined the stress process
and influences on psychological well-being and depression. For American Indians/Alaska Natives (AI/AN),
many studies have focused on depression and its relation
to alcohol abuse and suicide (Beauvais & Oetting, 1999;
Christensen & Manson, 2001; Green, Sack, & Pambrum,
1981; King, Beals, & Manson, 1992; Novins, Beals, &
Roberts, 1999). Very little research has examined the
perceptions of social support, cultural identity, and other
contextual influences on the mental health status and
protective and risk-taking behaviors of AI youth.

The SMYP was funded by the National Institute of Mental Health (R01-MH59001).
Address correspondence to Professor and Chair Julie A. Baldwin, Department of Community and Family Health, College of Public Health,
University of South Florida, 13201 Bruce B. Downs Boulevard, MDC 56, Tampa, FL 33612; E-mail: jbaldwin@health.usf.edu.
1
The SMYP protection of human subjects procedures included four original IRB reviews (Northern Arizona University, Indian Health Service
(federal and regional), and a formal tribal IRB), plus approvals from the tribal councils and research organizations, school boards, and health
agencies of all participating tribes. At the request of the tribal IRB’s and governments, we are not identifying either the dormitories or the specific
tribal affiliations of the students.

1380

Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

ROLE OF CULTURE AND CONTEXT IN BUFFERING STRESS

grief (Beauvais, 2000; Brave Heart, 1998; Mehrabadi
et al., 2008; Pearce et al., 2008; Novins et al., 1999). While
the effects of this “historical trauma” (Brave Heart, 1998)
may be difficult to quantify (Walters, Simoni, & EvansCampbell, 2002), it seems logical that sequelae would include greater distress and negative health outcomes. In a
recent study conceptualizing “historical loss,” Whitbeck,
Adams, Hoyt, and Chen (2004) found that anger, avoidance, anxiety, and depression were significantly linked to
emotional distress. Among the people of one tribe, evidence has been found of a “historical trauma response,”
including depression, unresolved grief, survivor guilt, excesses in cardiovascular disease and mortality, and violent deaths (Brave Heart, 1999). Although empirical
support for the connection among historical trauma,
stressful events, and negative health outcomes is often
descriptive (Strickland, Walsh, & Cooper, 2006; Walters
et al., 2002), many researchers, nevertheless, lend great
importance to the contribution of history to current AI
health issues, including deleterious health effects (Beauvais, 2000; Brave Heart, 1998; Mehrabadi et al., 2008;
Novins et al., 1999; Pearce et al., 2008; Whitbeck et al.,
2004).
It might also be worth noting the role that perception
of historical and current discrimination can have on one’s
health. For example, researchers (Whitbeck, Hoyt, McMorris, Xiaojin, & Stubben, 2001) found a relationship
between perceived discrimination and early substance
abuse2 among AI children. Perceived discrimination significantly contributed to internalizing symptoms among
youth, and the effects of perceived discrimination on early
substance use were mediated by adolescent anger and
delinquent behavior. In another study (Whitbeck, McMorris, Hoyt, Stubben, & LaFromboise, 2002), researchers
found that perceptions of discrimination were strongly
associated with depressive symptoms among AI adults.
American Indian Adolescents: A Population at Risk

The health-related issues and personal concerns facing
AI adolescents, such as peer pressure, alcohol and drug
use, violence, and premature sexual activity, are not unlike those facing other adolescents (Beauvais, 2000); however, AI adolescents face additional concerns. Almost
one-half of AI adolescents live below the poverty line,
and studies of high school students have identified dropout rates ranging from 20% to 80% (Greene & Forster,
2003; Novins et al., 1999). Boarding schools for AIs have
previously been characterized as having a high concentration of youth exhibiting risk behaviors (King et al., 1992;
Robbins, Colmant, Dorton, Schultz, & Ciali, 2006;
Shaughness & Jones, 2003). In one study of AI youth
residing in boarding schools, the frequency and severity
of comorbidity of depression, “suicidality,” and substance
use were found to be higher than for non-AIs and this trend
increased with age (Dinges & Duong-Tran, 1993).
2
The journal’s style utilizes the category substance abuse as a diagnostic category. Substances are used or misused; living organisms are and
can be abused. Editor’s note.

1381

In general, the health status of AI teenagers is below
that of the overall U.S. adolescent population. AI youth
have especially high rates of depression, suicide, anxiety, substance use, school dropout, and alcohol misuse
(Cameron, 1999; Waldvogel, Rueter, & Oberg, 2008; Wallace et al., 2002). AI youth are at a heightened risk for
chronic distress as a result of poverty, cultural trauma,
and violence (Grandbois, 2005; Rieckmann, Wadsworth,
& Deyhle, 2004). Distress and alienation from cultural
traditions, in turn, may lead to or be associated with depression, learning problems, conduct disorders, substance
use, running away, and suicide attempts (King et al., 1992;
Shaughnessy, Doshi, & Jones, 2004; Walters et al., 2002).
Depression is the most common diagnosis among
teenage girls at Indian Health Service mental health outpatient clinics, and consequences of depression include
suicide attempts (Rieckmann et al., 2004). According
to Dinges and Duong-Tran (1993), overall comorbidity
rates between depression and suicide and substance use
are higher for AI youth than for youth of other ethnicities, and depression is more frequently correlated with
suicide than with substance use. In another study, 16%
of the AI students attending a Bureau of Indian Affairs
(BIA)-funded high school attempted suicide one or more
times in the past 12 months. Students who attempted
suicide were more likely to engage in several risky
behaviors, including unintentional injury and violent
behavior, sexual risk behavior, and tobacco, alcohol, and
other drug use (Shaughnessy et al., 2004). There is also
high comorbidity between psychiatric disorders, such as
depression and alcoholism, among the general AI population (Westermeyer, 2001). A link between depression and
tobacco use has also been documented. In a three-year
longitudinal study of 743 AI youth, Whitbeck, Yu,
McChargue, and Crawford (2009) found that cigarette
smoking increased through time for all youth, but depressive symptoms were only associated with an increase in
cigarette smoking among girls. According to one tribe’s
data from the Youth Risk Behavior Surveillance System
Report (Benally, Werito, Begay, Jones & Yabeny, 2000),
almost one-third of high school students (39% of females
and 25% of males) reported that within the past year,
they “felt so sad or hopeless almost every day for two
weeks or more in a row that they stopped doing some
usual activity” (p. 8). In the United States, the AI suicide
rate among those aged 15–24 years is three times that
of the overall U.S. rate for that age group: 13%–23% of
AI youth self-reported suicide attempts and 14%–41%
have reported suicide ideation (Beauvais, 2000; Novins
et al., 1999). Alcohol use is a major concern among many
AI nations (May, 1994). According to Mitchell and colleagues (1996), AI youth on average reported greater use
of alcohol than did any other racial or ethnic subgroup;
however, certainly not all AI youth use or abuse alcohol
(May, 1994). Experimentation with alcohol consumption
may begin as early as 10 years old (Mitchell, O’Neill, &
Beals, 1996), and in a survey of 9th–12th grade students
at BIA-funded schools, lifetime alcohol use was reported
by more than 80% of students; approximately 50% of

Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

1382

J. A. BALDWIN ET AL.

students reported current alcohol use and 38% of students
reported episodic “heavy drinking” (Shaughness & Jones,
2003). One recent study found that drinking among AI
youth was not related to traditional Indian culture (Spicer,
Novins, & Mitchell, 2003); on the contrary, peer influence
appears to be an increasingly important factor in alcohol
initiation (Kunitz, 2006; Spicer et al., 2003). Low social
support from family was correlated strongly with high
alcohol use, and in boarding schools, alcohol use has been
reported to be at “near epidemic proportions” (Wiegman
Dick, Manson, & Beals, 1993).
According to the 1997 National School-Based
Youth Risk Survey, AI youth engaged in risky (“selfdestructive”) behaviors more often overall than white
and black youth (Frank & Lester, 2002). AI youth in
school were significantly more likely to attempt suicide
and use cocaine, while AI male youth were more likely
to carry weapons, smoke cigarettes, and use marijuana
as well (Frank & Lester, 2002). In addition, a higher
percentage of AI youth use alcohol and marijuana
at younger ages than other youth. According to one
study, alcohol and drug use begins with (often “heavy”)
drinking at a young age among AIs (Vernon & JumperThurman, 2002), and in a comparison of AI to non-Indian
fourth to sixth graders, Miller, Beauvais, Burnside, and
Jumper-Thurman (2008) found that 12% of AI and 3% of
non-Indian youth reported marijuana use in the past year.
The literature provides strong support for a link between stressful events and negative outcomes such as depression, substance use, and other risky behaviors among
adolescent AI populations. At the same time, there are
many strengths of AI families, communities, and culture
that might mitigate these deleterious outcomes. Thus, we
also sought to examine potential variables that might ameliorate these relationships, such as social support, protective family and peer influences, and cultural identity.
Social Support

Findings regarding the relationship between social support, depression, substance use, and risky behavior are
mixed. King and colleagues (1992) found no direct relationship between social support and drug abuse or depression among AI boarding school students. In the same
study, however, family support mediated rates of alcohol
use for these students, possibly because of parental modeling of appropriate behaviors. Age was positively related to
both peer and family support, with no gender differences
(King et al., 1992). Wiegman Dick and colleagues (1993)
also found that levels of familial support are inversely correlated with levels of teenage alcohol use, while peer influence was linked with alcohol use. Other studies have
found that peer influence is highly associated with substance use (Bates, Beauvais, & Trimble, 1997; Swaim,
Nemeth, & Oetting, 1995; Swaim, Oetting, Edwards, &
Beauvais, 1989).
Protective Influences: Family and Peer Sanctions
Against Alcohol and Drug Use

Some studies have found a negative association between
family and peer sanctions against substances and sub-

stance abuse. Swaim, Oetting, and Thurman (1993) stated
that family sanctions among AIs had both direct and indirect effects against drug use. Family sanctions were, in
fact, found to be more important for AI youths’ avoidance of drug use than for Anglo youth because no direct
effects were found for the latter. In a study looking exclusively at alcohol use among AI adolescents, family sanctions against alcohol were found to be predictive only for
females’ non-use (Bates et al., 1997).
Literature also suggests that Native American youth
with a non-parental role model are less likely to use alcohol, tobacco, and other drugs. Beebe and colleagues
(2008) found that AI youth who had a nonparental role
model were 4.4 times less likely to have used alcohol,
nearly 7.5 times less likely to have used tobacco, and 5
times less likely to have used other drugs in the past 30
days than youth without a nonparental role model. In the
literature, peers and peer groups are generally noted for
their positive associations with alcohol and substance use
(e.g., Swaim et al., 1989), but these authors have also
found that negative peer attitudes toward alcohol and substance use can help limit these behaviors in adolescents.
Cultural Identity

There is still little research on AI cultural identity as a protective factor3 (Walters et al., 2002; Whitbeck et al., 2002),
and when it has been incorporated in studies to predict risk
behavior in youth, it has not always emerged as a significant predictor. Some researchers have used collectivistic
orientation or a sense of communal belonging as a proxy
for traditional Native American cultural identity (Hill,
2009; Hofboll, Jackson, Hofboll, Pierce, & Young, 2002;
Mohatt et al., 2004; Richmond & Ross, 2008; Richmond,
Ross, & Egeland, 2007). In a study of AIs, researchers
found a negative correlation between sense of belonging
and suicidal ideation, suggesting that individuals who feel
connected are less likely to experience suicidal thoughts
(Hill, 2009). In other studies among AIs and ANs, individuals who felt a strong sense of connectivity with others
experienced protective health effects (Mohatt et al., 2004;
Richmond & Ross, 2008; Richmond et al., 2007); in one
such study, women with an increased sense of communal
mastery (social attachment) experienced less increase in
depressive mood and anger than women with lower levels
of communal mastery (Hofboll et al., 2002). In additional
studies among AI youth, youth scoring high on Native cultural identity were no more likely to experience depressive
symptoms, although youth who perceived individualistic
3

The reader is reminded that the concepts of risk and protective factors are often noted in the literature without adequately delineating their
dimensions (linear and nonlinear), their “demands,” the critical necessary conditions (endogenously as well as exogenously; from a micro to
a macro level), which are necessary for either one of them to operate
(begin, continue, become anchored and integrate, change as de facto
realities change, cease, etc.) or not to and whether their underpinnings
are theory-driven, empirically based, individual and/or systemic stake
holder-bound, based upon “principles of faith,” historical observation,
precedents, and traditions that accumulate over time, perceptual, and
judgmental constraints, “transient public opinion.” or what. This is necessary to clarify if the term is not to remain as yet another shibboleth in
a field of many stereotypes. Editor’s note.

Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

ROLE OF CULTURE AND CONTEXT IN BUFFERING STRESS

goals were more likely to experience increased depressive
symptoms, perhaps because they are given lower cultural
priority with a Native American context (Hamill, Scott,
Dearing, & Pepper, 2009; Scott et al., 2008).
While Rieckmann et al. (2004) found “cultural attitude” to be weakly protective against depression in Navajo
adolescents, other studies found that AI identity was unrelated to tobacco use (Yu, Stiffmen, & Freedenthal, 2005)
or alcohol involvement (Bates et al., 1997; Trimble, 1995).
Instead, peer alcohol and tobacco associations and family substance use associations were stronger predictors of
alcohol or tobacco use in these studies. However, these
“weak” correlations between cultural identity and risk behaviors might stem from the previous ways in which cultural identity has been operationalized.
Oetting and Beauvais (1991) provide a distinct model
for cultural identification, the orthogonal cultural identification theory. According to these authors, identification
with any culture is a source of personal and social strength
and can be protective. Dimensions of cultural identification are measured independently of each other, thereby not
precluding association with one culture instead of another.
For example, strong identification with traditional AI culture does not require weak identification with the mainstream Anglo culture. A protective effect, therefore, may
be derived from identification with either culture or both.
Orthogonal cultural identification relates significantly to
important personal and social characteristics of youth.
Nevertheless, greater strength of identification does not
translate automatically into better health habits such as
less drug use. Drug use is related to how much the culture
with which the person identifies approves or disapproves
of drugs (Oetting & Beauvais, 1991). Both the risk and
protective nature of cultural identification warranted exploration. In summary, based upon previous findings and
our a priori expectations, this study was carried out to test
two models. The first model depicts main effect relationships between stressful life events, AI identity, and white
identity as predictors of depressed mood, substance use,
and risky behavior. The second model not only includes
the same variables as the first model but also includes social support and protective family and peer influences as
moderating variables situated between the antecedent and
outcome variables. These analyses represent the first attempt, to our knowledge, to compare two competing models for their ability to explain the relationships between
stressors and risky behavior among AI adolescents.

METHODS
Participants and Procedure

Data used for these analyses were collected in the SMYP
Initial Baseline Interview and the “Family, Life, and
Friends” questionnaire. The first wave of data collection
began in September 1999 and the last wave of data collection period ended in January 2001. While longitudinal information was collected three times from individuals who
participated in all three waves, the data for which this pa-

1383

per’s analyses are based were obtained from individuals’
baseline interviews only.
High school students from two off-reservation dormitories were first interviewed during the fall 1999 semester.
To be eligible for this initial interview, students had to be
enrolled in one of the two target dormitories, be between
the ages of 15 and 24 years, be AI, and give informed assent. For those youth under the age of 18 years, the researchers also obtained parental consent for the youth to
participate. All students, regardless of risk or nonrisk behavior, were eligible to participate since we were interested in identifying protective- as well as risk-behavior
information. In the baseline interview, students identified
members of their social networks; full relational network
interviews were conducted on a smaller sample from each
dormitory. With the permission of the students, network
members were contacted and completed the same interviews. The data were largely collected from dormitory
students (81%), but other individuals who were personal
contacts of the students participated as well. The members of the respondents’ networks had to be 15–24 years
of age, but could be of any ethnicity and did not need to
reside in a dormitory to be eligible for the interview.
Measures

Table 1 presents information about the measures used in
this study, including mean scores, standard deviations, and
normality (skewness and kurtosis). Most of the variables
listed in the table (so-called “manifest” variables) were
subsequently used to create latent variables for structural
equation modeling analyses to test our hypothesized models. In the final model, only age, gender, and stressful
events were used as “manifest” variables.
Demographic Information
Participants were asked to provide information about their
age and gender, as these variables have been demonstrated
to be predictors of perceived social support and depression (LeMaster, Beals, Novins, & Manson, 2004; Wallace
et al., 2003; Whitbeck et al., 2009). In compliance with
the confidentiality agreements with the tribes, no specific
tribal affiliation is identified in these analyses and data are
presented in an aggregate form.
Stressful Life Events
Categorical items were used to evaluate whether a respondent had experienced a stressful life event within the past
12 months (Lewinsohn et al., 1994). Participants were
asked to respond “yes” (= 1) or “no” (= 0) to the following question: “In the past 12 months, did the following event happen to you?” There were 12 items, including
entering high school as a new or transfer student, being
pregnant or having a partner who was pregnant, giving
birth or having a partner give birth, having a family member attempt suicide, having an adult who is important to
you have an alcohol or drug problem, experiencing verbal
abuse from adults in your family, having a parent unable to
find employment, having people gossip about you, having
a friend attempt suicide, having a serious argument with a

1384

J. A. BALDWIN ET AL.

TABLE 1. Means, standard deviations, and normality of study variables by gender
Females (N = 112)

Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

Ageb
Stressful life eventsc
Social supportd
Protective family and peer
influencee
Native American identitye
White/Anglo identitye
Depressed moodd
Substance usef
Risky behaviore

Males (N = 109)

M

SD

Skewness

Kurtosis

M

SD

Skewness

Kurtosis

F (1, 121)a

16.0
2.20
4.27
3.49

1.50
1.63
0.66
0.49

3.08
0.82
−1.00
−1.63

13.88
0.67
0.18
3.43

16.6
1.83
3.92
3.25

1.55
1.73
0.79
0.58

0.84
1.15
−0.96
−0.81

0.28
1.06
0.85
−0.03

6.96∗∗
2.18
12.58∗∗∗
10.67∗∗

3.40
2.67
1.80
1.75
1.48

0.69
0.80
0.76
1.60
0.43

−1.45
−0.49
1.22
0.86
1.10

1.66
−0.10
0.74
0.05
0.92

3.21
2.36
1.47
1.79
1.54

0.86
0.79
0.56
1.73
0.50

−1.19
−0.13
1.69
0.92
1.68

0.35
−0.84
2.91
0.05
4.34

3.22
8.04∗∗
13.65∗∗∗
0.03
0.94

a

Statistic from the corrected model, multivariate analysis of variance (MANOVA).
Range = 15–24.
c
Range = 0–12.
d
1–5 scale.
e
1–4 scale.
f
Range = 0–7.5.
∗∗
p < .01, ∗∗∗ p < .001.
b

friend, breaking up with a boyfriend or girlfriend, and being in a serious car wreck. For the purposes of this study,
responses to this group of variables were summed and a
new variable, “Stress,” created.
Alpha reliability models provide an approximate
equivalent coefficient to the Kudor-Richardson 20 (KR20)
for dichotomous variables. Since several of the items for
stressful life events were gender specific, coefficients were
run for males and females separately (Cronbach’s α =
0.592 and 0.482, respectively). These results demonstrate
a low correlation, as we expected, since our measure of
stressful life events is an index, not a scale.
Social Support
Perceived social support was evaluated using selected
items from the Multidimensional Scale of Perceived Social Support (Procidano & Heller, 1983). Participants
were asked to describe how much they agreed or disagreed
with levels of social support they felt that they received
from family, peers, and a special person. A five-point Likert scale ranged from 1 = “Strongly Disagree” and 5 =
“Strongly Agree” for the following seven items used in
this analysis: “My family really tries to help me,” “I can
talk about my problems with my family,” “If someone
wants to fight me, my family will stand by me,” “I have
friends with whom I can share my joys and sorrows,” “I
can talk about my problems with my friends,” “There is a
special person who is around when I am in need,” and “I
have a special person who is a real source of comfort to
me.”
In this study, these items demonstrated strong intercorrelation (Pearson’s correlation significant at the 0.01
level for all items) and a Cronbach’s α reliability of 0.77.
Participants’ responses to items were then used to create a
latent variable called “Social Support” composed of three

indicators of support by group. The α reliabilities of each
of the subscales were 0.58 for family support, 0.87 for
support from friends, and 0.80 for support from a special
person.
Protective Family and Peer Influence
Protective family and peer influence was evaluated using
selected items from Beauvais and Oetting (1991). Participants were asked questions about the protective influence
of family and peers. To measure family’s caring about
the youths’ behaviors, participants were asked to describe
how much they believed that their families cared whether
they “smoked cigarettes,” “got drunk,” “sniffed something
like gas or glue,” “used marijuana,” or “used other drugs.”
Responses were based on a four-point Likert scale: 1 =
“Not at all,” 2 = “Not much,” 3 = “Some,” and 4 = “A lot.”
In two other items, they were asked to similarly rank how
much their families and friends would try to stop them
from engaging in these behaviors. A fourth item asked
how much their friends would try to get them to use alcohol, tobacco, and marijuana; this item was reverse-coded
so that a higher score supported more protective behavior
and correlations could be run with the three other items in
the group. Participants’ responses to items were then used
to create a latent variable called “Protective Family and
Peer Influence” composed of four indicators of influence.
The α reliabilities of each of the subscales were 0.91
for “family caring,” 0.95 for “family’s trying to stop behaviors,” 0.96 for ‘friends’ trying to stop behaviors,’ and
0.86 for the reverse-coded item, “friends not trying to
get participants to use.” Three of the four items demonstrated strong intercorrelation (Pearson’s correlation
significant at the p < .01 level for all items), with the
exception of the friends not encouraging substance use.
Thus, when interpreting the latent variable, “Protective
Family and Peer Influence,” it is important to note that

ROLE OF CULTURE AND CONTEXT IN BUFFERING STRESS

Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

the variance in “friends not trying to get participants to
use” did not significantly covary with the other indicators
of this latent variable nor did it make a significant contribution to the variance of the latent variable.

American Indian and White Identity
Orthogonal cultural identification measures explored
strength of identification with each cultural way of life
separately, thereby not precluding participation in or identification with one culture for another. An orthogonal
cultural identification scale developed by Oetting and
Beauvais (1991) has demonstrated previous reliability:
α = 0.89. These items include questions about life and
success in various cultures.
In our study, participants were asked to respond to the
following questions based on a Likert scale: 1 = “Not
at all,” 2 = “A little,” 3 = “Some,” and 4 = “A lot.”
Items selected for use in this analysis included: “Does
your family live by or follow . . . the American Indian way
of life/the white/Anglo way of life?”; “Do you live by or
follow . . . the American Indian/the white/Anglo way of
life?”; “In your family, how many activities or traditions
are based on . . . American Indian culture/white/Anglo
culture?”; “When you are an adult, how involved do you
think you will be in . . . American Indian traditions and
beliefs/white/Anglo traditions and beliefs?”; “When you
are an adult, will you be a success in . . . American Indian
way of life/white/Anglo way of life?”; “Is your family a
success in . . . American Indian way of life/white/Anglo
way of life?” These items were used to create two latent
variables termed “American Indian Identity” and “White
Identity.” Items in each group demonstrated strong intercorrelation (Pearson’s correlation significant at the p < .01
level for all items) and a reliability of 0.94 for AI identity
and 0.90 for white identity.

Depressed Mood
Depressed mood was measured using the Center for Epidemiologic Studies Depression Scale (CES-D) to assess
symptoms of depressed mood experienced within the past
week (Reynolds, 1988). This scale has exhibited strong
reliability with AIs (Manson, 1997; α = 0.82).
Participants identified how many times within the past
week they experienced depressive symptoms (11 items),
including: “I felt that I could not shake off the blues (sadness) even with help from my family or friends.” “I felt
depressed.” “I thought my life had been a failure.” “I felt
fearful.” “I felt lonely.” “I had crying spells.” “I felt sad.”
“I felt crabby.” “I felt hopeless.” “I felt like no one cared.”
and “I felt discouraged.” The Likert scale included the following range: 1 = rarely or none of the time (0–1 days);
2 = some or a little of the time (1–2 days); 3 = a moderate
amount of the time (3–4 days); and 4 = most or all of
the time (5–7 days). These items demonstrated a strong
intercorrelation (Pearson’s correlation significant at the
p < .01 and a reliability of 0.92. The items were randomly
divided into three indicators that were used to create a

1385

latent variable termed “Depressed Mood.”‘ The α reliabilities of each of the subscales were 0.74, 0.79, and 0.82.
Substance Use
Items selected for use in this analysis included questions
about alcohol and drug use from Tri-Ethnic Center for
Prevention (Oetting & Beauvais, 1991). Participants were
asked: “Have you ever used alcohol?” and “Have you used
alcohol in the past 30 days?” Responses to these two items
were recoded to create a new variable describing “alcohol
use” and were based on the following scale: 0 = “Never
used alcohol,” 1 = “Used alcohol, but not in the past 30
days,” and 2 = “Used alcohol at least one day in the past
30 days.” A second item was created in a similar manner
to describe “marijuana use.”
To determine “problems related to alcohol use,” participants were asked to respond “yes” ( = 1) or “no” ( = 0) as
to whether drinking ever caused any of 16 behaviors: “get
a traffic ticket,” “have a car accident,” “get arrested,” “have
money problems,” “have trouble in school,” “hurt school
work,” “fight with other kids and adults,” “fight with parents,” “damaged a friendship,” “pass out,” “not remember
what happened,” “break something (vandalism),” “injury
to self,” “drop out of school,” “go further sexually than intended to,” or “try drugs not intended to.” A new variable
was created as a sum of these experiences. A fourth item
was created in a similar manner to describe “problems related to drug use.” These four items were used to create
a latent variable called “Substance Abuse.” These items
demonstrated a strong intercorrelation (Pearson’s correlation significant at the p < .01 level for all items) and a
reliability of 0.72.
Risky Behavior
Participants identified how many times within the past
month they had taken risks with their friends (10 items),
including the following: They had “done something risky
or dangerous on a dare,” “broken a rule that your parents/guardians or dorm aides set for you to see if you could
get away with it,” “stolen or shoplifted,” “slipped out at
night when your parents/guardians or dorm aides thought
you were asleep,” “willingly ridden in a car with someone you knew was driving dangerously,” “physically hurt
yourself on purpose,” “vandalized or destroyed school
property, or someone else’s property,” “talked back to
your parents/guardians, or done something on purpose to
make your parents angry,” “ditched from school or work,”
and “jumped into someone’s car when there was a raid
at a party.” (These items were adapted from the work
of Alexander, Somerfield, Ensminger, Kim, & Johnson,
1995.) These items demonstrated a strong intercorrelation (Pearson’s correlation significant at the p < .01 level
for all items) and a reliability of 0.74. The items were
randomly divided into three indicators that were used to
create a latent variable termed “Risky Behavior.” The reliabilities of each of the subscales were 0.72, 0.57, and
0.61.

1386

J. A. BALDWIN ET AL.

RESULTS

Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

Preliminary Analyses

Of the 269 students who participated in this study, a sample of AI youth (N = 221) was selected based on their
completion of both the Initial Baseline Interview and the
“Family, Life, and Friends” questionnaire from which the
study variables were obtained. The average age of the participants in this sample was 16.3 years (ages ranged from
15 to 24 years), though the mean age for girls (16.0 years)
was approximately six months younger than that for boys
(16.6 years). In this sample, participants were distributed
almost equally by gender (49% males and 51% females)
and all were AIs. Most respondents (93%) were currently
enrolled in school and had successfully completed 9th,
10th, or 11th grade (38%, 22%, and 24%, respectively).
Seventy-two percent of this sample were residents of one
of the target dormitories. Of those youth who knew their
parents’ educational attainment, 56% of the mothers and
54% of the fathers had completed high school or a GED;
23% of mothers and 15% of fathers had completed one or
more college degrees.
In order to examine gender differences on the study
variables, a one-way multivariate analysis of variance
(MANOVA) was conducted with gender as the betweengroup variable and age, stressful life events, social support, protective family and peer influence, AI identity,
white identity, depressed mood, substance use, and risky
behavior as the dependent measures. A significant main
effect for gender was found (Pillai’s trace, F (9, 221) =
6.82, p < .001). Inspection of univariate tests revealed that
girls had higher scores than boys on social support, positive family and friend influence, identification with white
culture, and depression. There were no significant gender
differences on stressful life events, AI identification, substance use, or risky behaviors.
Confirmatory Factor Analysis

The first step in the structural modeling analysis was
to test a measurement model using a confirmatory factor analysis (CFA) using the EQS structural modeling
program (Bentler, 1995). It was hypothesized that items
from 29 manifest variables would reflect seven latent constructs: “Social Support,” “Protective Family and Peer Influence,” “American Indian Identity,” “White Identity,”
“Depressed Mood,” “Substance Use,” and “Risky Behavior.” For identification purposes, the variances of the seven
latent factors were fixed at 1.00 and their respective indicators were allowed to load freely. The latent variables
were allowed to correlate. Criteria used to assess model fit
were based on established standards (Kline, 1998).
As predicted, the seven-factor CFA model fit moderately well (χ 2 (476) = 816.77, p < .001, CFI = 0.91,
NNFI = 0.89, SRMR = 0.09). The analysis revealed that
the measures loading onto each of the latent variables
were reliable at the p < .0001 level with one exception.
The item “friends not trying to get participants to use alcohol or drugs” (reverse-coded) on the “Protective Family and Peer Influence” latent variable only loaded 0.11

TABLE 2. Latent variables and factor loadings in confirmatory
factor analysis (N = 221)
Latent variables
Social support
Support of a “special person”
Support from family
Support from friends
Protective family and peer influence
Family’s concern about substance use
Family’s attempt to stop substance use
Friends’ not influencing substance use
Friends’ attempt to stop substance use
American Indian identity
Family lives by/follows culture
Individual lives by/follows culture
Activities/traditions of culture
Individual’s future involvement in culture
Individual’s future success in culture
Family’s future success in culture
White identity
Family lives by/follows culture
Individual lives by/follows culture
Activities/traditions of culture
Individual’s future involvement in culture
Individual’s future success in culture
Family’s future success in culture
Depressed mooda
CESD-1
CESD-2
CESD-3
Substance abuse
Level of alcohol use
Level of marijuana use
Alcohol use causes problems
Drug use causes problems
Risky behaviora
Risky behaviors 1
Risky behaviors 2
Risky behaviors 3

Factor loadings
0.56∗∗∗
0.68∗∗∗
0.59∗∗∗
0.62∗∗∗
0.79∗∗∗
0.11
0.74∗∗∗
0.88∗∗∗
0.82∗∗∗
0.78∗∗∗
0.89∗∗∗
0.87∗∗∗
0.82∗∗∗
0.77∗∗∗
0.82∗∗∗
0.76∗∗∗
0.79∗∗∗
0.75∗∗∗
0.75∗∗∗
0.75∗∗∗
0.96∗∗∗
0.85∗∗∗
0.68∗∗∗
0.64∗∗∗
0.91∗∗∗
0.78
0.71∗∗∗
0.68∗∗∗
0.72∗∗∗

a

Variables randomly assigned to groups for loading onto latent factors.
∗∗∗
p < .001.

onto the latent variable (p < .10). Table 2 presents all of
the factor loadings for the seven latent variables. With this
factor structure of the seven latent factors confirmed, correlations among the latent and manifest variables were
computed (see Table 3).
Structural Equation Models

As shown in Figure 1, our first model predicted that stressful life events would be positively associated with depressed mood, substance use, and risky behavior. On the
basis of the existing literature, we also predicted that female participants would report greater depressed mood
and that stronger AI or white identity would be associated
with less substance use and less risky behavior. On the
basis of the correlations obtained in the CFA, we added a
path between age and substance use.

1387

ROLE OF CULTURE AND CONTEXT IN BUFFERING STRESS

TABLE 3. Correlations among manifest and latent variables (N = 221)
1.

Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.


−0.18
−0.03
0.15
−.14
0.02
0.01
−0.07
0.14
0.07

2.

0.10
0.26
0.26
0.12
0.20
0.27
−0.00
−0.09

3.


−0.14
−0.16
−0.07
−0.03
0.49
0.38
0.49

4.


0.59
0.24
0.21
−0.32
0.02
−0.21

5.


0.17
0.08
0.10
−0.06
−0.28

6.

7.

8.

9.


−0.11
0.07
0.05
−0.11


0.02
−0.12
−0.06


0.19
0.35


0.45

Note. 1 = age; 2 = sex; 3 = stress; 4 = social support; 5 = protective family and peer influence; 6 = American Indian identity;
7 = White identity; 8 = depressed mood; 9 = substance abuse; 10 = risky behavior.
Coefficients in bold indicate p < .05 or less.

As shown in Figure 2, similar to the first model,
age was expected to be related to substance use, female
participants were expected to report greater depressed
mood, and stressful events were expected to be positively related to depressed mood, substance use, and
risky behavior. Model 2 differs from Model 1 in
that it hypothesizes that females and those who are
higher in AI and/or white identities will report greater
social support and protective family and peer influence. In turn, social support and protective family
and peer influence are expected to affect the outcome
variables. Specifically, greater social support was expected to be associated with less depressed mood and
less risky behavior, while protective family and peer
influence were expected to be associated with less
substance use or other risky behavior. On the basis of
the correlations obtained in the CFA, we also included a

path between age and social support. Given that neither
AI identity nor white identity was significantly correlated
with any of the outcome variables, social support and protective family and peer influence were not expected to mediate the relationships between cultural identity and the
outcome variables.
Our first goal was to assess whether both models
provided a good account of the data and to observe if the
addition of social support and protective family and peer
influence appeared to add predictive variance in the outcome variables. In order to make these comparisons, we
first tested Model 1 and then Model 2 using EQS, a structural modeling program (Bentler, 1995). For each model,
the variance of the seven latent factors was fixed at 1.00
(for identification purposes) and their respective indicator variables were allowed to load freely onto the latent
factor. The first model fit adequately (χ 2 (487) = 934.59,

FIGURE 1. Hypothesized direct effects model. Circles denote hypothesized latent constructs and rectangles denote measured variables. Paths
represent hypothesized relationships among the variables. All paths represent positive associations between the variables.

Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

1388

J. A. BALDWIN ET AL.

FIGURE 2. Hypothesized mediational effects model. Circles denote hypothesized latent constructs and rectangles denote measured variables.
Paths represent hypothesized relationships among the variables. All paths represent positive associations between the variables.

p < .001, χ 2 : df = 1.92; CFI = 0.88, NNFI = 0.87, SRMR
= 0.16, RMSEA = 0.07). Inspection of LaGrange multiplier tests suggested that adding three pairs of correlated
error terms would improve the fit of the model, which they
did (χ 2 (484) = 849.51, p < .001, χ 2 : df = 1.76; CFI =
0.90, NNFI = 0.89, SRMR = 0.17, RMSEA = 0.06; significance test χ 2 (3) = 85.08, p < .0001). Although not
depicted in the figure, age and sex were negatively correlated (−0.18, p < .05), indicating that female respondents were younger than male respondents. Path coeffi-

cients, similar to partial regression weights, are presented
in Figure 3. As indicated in Figure 3, stressful life events
were positively associated with depressed mood (0.49,
p < .001), substance use (0.32, p < .001), and risky behavior (0.49, p < .001). In addition, older participants
were more likely to use substances (0.13, p < .05), and
female participants were more likely than male respondents to report depressed mood (0.22, p < .001). Neither
AI identity nor white identity was significantly associated with substance use or risky behavior. The amount of

FIGURE 3. Final SEM direct effects model (N = 221). Rectangles represent measured variables and circles denote latent constructs.

p < .05, ∗∗∗ p < .001.

Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

ROLE OF CULTURE AND CONTEXT IN BUFFERING STRESS

variance in each of the outcome variables accounted for
by this model was as follows: depressed mood (27%), substance use (13%), and risky behavior (25%).
The first examination of Model 2 also revealed a satisfactory fitting model (χ 2 (479) = 837.05, p < .001, χ 2 :
df = 1.75; CFI = 0.90, NNFI = 0.89, SRMR = 0.12, RMSEA = 0.06). Inspection of LaGrange multiplier tests suggested that adding three pairs of correlated error terms and
one nonstandard effect would improve the fit of the model,
which they did (χ 2 (475) = 690.07, p < .001, χ 2 : df =
1.45; CFI = 0.94, NNFI = 0.94, SRMR = 0.12, RMSEA
= 0.05; significance test χ 2 (4) = 146.98, p < .0001). A
nonstandard effect is a relationship between the portion
of variance in an indicator variable that is not captured by
the latent variable and a latent variable (Newcomb, 1990).
In this case, a relationship between the error term (e.g.,
the unique variance) for the lowest loading item (the item
described earlier that only loaded 0.11 onto the factor)
loading onto “Protective family and peer influence” was
correlated with the latent variable “risky behavior.” Age
and sex were negatively correlated (−0.18, p < .05), as
in Model 1. Path coefficients, similar to partial regression
weights, are presented in Figure 4. Stressful life events
were positively associated with depressed mood (0.44,
p < .001), substance use (0.32, p < .001), and risky behavior (0.38, p < .001). In addition, older participants were
more likely to use substances (0.14, p < .05) and female
participants were more likely than male respondents to report depressed mood (0.31, p < .001). These patterns of
relationships were very similar to those found in Model 1.
The unique aspects of Model 2 were that female respondents reported greater social support (0.35, p < .01) and
protective and family and peer influence (0.22, p < .01),
older respondents reported more social support (0.21,
p < .01), and AI identity and white identity were related
to social support (0.20, p < .05 and 0.17, p < .05, respectively), and AI identity was positively related to protective
family and peer influence (0.13, p < .05). In turn, social
support was negatively related to depressed mood (−0.29,
p < .001) and protective family and peer influence was
related to less risky behavior (−0.18, p < .05). The nonstandard effect suggests that the unique portion of the item
“friends not trying to get participants to use” (reversecoded) that was not captured by the latent variable
protective family and peer influence, was negatively correlated with risky behavior (−0.38, p < .001). The amount
of variance in each of the outcome variables accounted
for by this model was as follows: depressed mood (33%),
substance use (13%), and risky behavior (33%).
One of the advantages of examining our two models
is that we can determine that Model 2 actually provided
a significantly better fit to the data (χ 2 (8) = 159.44, p <
.0001). Inspection of the amount of variance accounted for
by the models also suggests Model 2 accounted for more
of the predicted variance in the outcome variables.
DISCUSSION

This study examined the relationships between the challenges faced by AI youth and moderators of stressors

1389

in their lives. We hypothesized that greater exposure to
stressful life events would be positively associated with
depressed mood, substance use, and risky behavior, but
that social support, greater cultural identity, and the positive influences of family and peers would reduce their impact. Model 1 demonstrated (as hypothesized) the main
effect relationships between stressful life events and depressed mood, substance use, and other risky behavior.
Model 2 contributed to the literature by exploring such
factors as social support, cultural beliefs, and protective
family and peer influences, which might improve our
understanding of the relationship between stressful life
events and negative outcomes such as depression, substance abuse, and risk-taking. Our use of structural equation models to test our hypotheses had the advantage of
reducing measurement error that is so typical in self-report
studies dealing with psychosocial data (Newcomb, 1990),
resulting in more confident estimates of the relationships
among those latent variables. In addition, by including latent variables measuring key contextual variables of cultural identity, perceived social support, family and peer
protective influences, we were actually able to improve
our ability to explain the relationships between stressful
events and risky outcomes for AI adolescents.
Our findings were also consistent with the literature in
that most studies have not found a direct relationship between cultural identify and deleterious outcomes such as
substance abuse. While Rieckmann and colleagues (2004)
found “cultural attitude” to be weakly protective against
depression in Navajo adolescents, other studies found that
AI identity was unrelated to alcohol involvement (Bates
et al., 1997; Trimble, 1995). In these studies, peer alcohol associations and, to a lesser measures extent, family
alcohol associations or family sanctions predicted alcohol
use. According to Trimble (1995), ethnic identity is a distinct psychological variable, but widely discrepant definitions and measures of ethnic identity make generalizations
and comparisons across studies difficult and ambiguous.
Ethnic identity has also been considered to be contextual
(Rieckmann et al., 2004); therefore, deciding how much a
person identifies with a particular culture can become very
arbitrary. A focus on AI and Anglo cultures alone ignores
the relevance of subcultures, such as youth or drug subcultures, which may lead to spurious conclusions about the
impact of cultural identification. Speaking of an “American Indian culture” can also lead one to overlook the fact
that AIs and ANs in the United States account for over 500
separate tribal groups with unique cultural backgrounds
(Roubideaux, 2002). Despite these limitations as to how
cultural identity is measured and operationalized, however, our analysis suggests that there is merit to continuing
to explore how cultural identity might be related to other
intervening or moderating variables that, in turn, have an
impact on negative health outcomes for youth.
Model 2 sheds light on the possible mechanisms by
which AI identification and white identification may help
us understand these negative outcomes. Specifically, there
is a strong association between both AI identification and
white identification with social support, and a strong association of AI identification with protective family and

Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

1390

J. A. BALDWIN ET AL.

FIGURE 4. Final SEM mediational effects model (N = 221). Rectangles represent measured variables and circles denote latent constructs.

p < .05, ∗∗ p < .01, ∗∗∗ p < .001.

peer influences. The positive relationship between cultural identification and social support lends credence to
the idea that cultural identification is a protective factor
for AIs. Early theories of “acculturation” posited that AIs
(and other minority groups) gradually take on the cultural
characteristics of the dominant Anglo culture and thereby
achieve better health and higher economic standards of
living (Oetting & Beauvais, 1991). An opposing theory
that is implicit or explicit in many recent substance user
programs and interventions posits that for Indigenous people, cultural identity is a protective factor for health. The
process by which ethnic minorities reaffirm their own cultures has been called “enculturation.” Maintaining AI cultural heritage, norms, and traditional values is considered
to be important for general psychological wellness, for
reduction of distress in particular, and for social cohesiveness (Walters et al., 2002). Incorporating local cultural traditions into AI/AN substance user programs has become
increasingly popular since the early 1990s (Noe, Fleming, & Manson, 2003). Spiritual identity is also considered
by many AIs and some researchers to be very important
for both health and healing (Lewton & Bydone, 2000).
AIs base their belief system on balance and demonstrate
a wealth of knowledge and experience with many healing
practices (O’Brien, Anslow, Begay, Pereira, & Sullivan,
2002).
The strong relationship between AI identification and
protective family and peer influences also supports the
need for programs that involve the whole family in culturally appropriate ways. Other researchers and practitioners
have agreed with this view. According to Beauvais (2001),
AI families play a significant role in their children’s lives,

and unlike non-Indian families, their parental influence
extends throughout adolescence. Furthermore, Swaim and
colleagues (1993) have found that Indian youth are more
responsive to family influences and less responsive to peer
influences in their decisions to use or not to use drugs and
alcohol when compared to non-Indian youth. The centrality of the family in AI culture and the living arrangements
that support higher levels of interaction with same-aged
relatives might partly account for this difference. Building
prevention programs on family and cultural values may
hold promise as successful strategies in reducing risky behavior among youth (Beauvais & Oetting, 1999; Hurdle,
Okamata, & Miles, 2003). Petoskey, Van Stelle, and De
Jong (1998) assert that personal successes rely on “early
grounding in values of one’s Native culture and a return
in adulthood to these spiritual roots” (p. 149).
Also interesting is the relationship between the error
variance of the one item in the latent construct of “Protective Family and Peer Influence” about “friends not trying
to get participants to use alcohol and drugs.” One study
of 8th, 10th, and 12th graders found that protective school
norms against substance use helped limit students’ use of
alcohol, cigarettes, and marijuana. “School norms,” which
are norms held by the student body in general, are different from “peer norms,” which generally refer to the norms
held by a smaller and more specific student peer group.
Kumar, O’Malley, and Johnston (2002) found that both
the norm of disapproval itself and its relationship with less
substance use diminished at each higher grade level. In
another study on adolescent students, Robin and Johnson
(1996) explored the concept of “cross pressure” to show
that the perception of peer disapproval of substance use

Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

ROLE OF CULTURE AND CONTEXT IN BUFFERING STRESS

was relevant. According to Beal, Ausiellio, and Perrin’s
(2001) study of urban minority students, peer disapproval
was significantly associated with decreased tobacco, alcohol, and marijuana use, as well as sexual activity, although
the association with alcohol was only significant at the
bivariate level. The absence of peer sanctions altogether
has also been noted. For example, Beauvais (1992) states
that AI adolescent heavy drug users are less likely to have
friends who would try to stop them from using drugs.
Thus, this study contributes to the existing knowledge base by exploring the potential contribution of
variables such as cultural identity, perceived social support, and family and peer protective influences to a
well-established relationship between stressors and their
negative sequelae—alcohol/drug use, depression, and
risk-taking behavior. There are, however, some limitations of this study that should be noted. First, only 36%
of the variance for the outcome variables, however, is
explained. Several other possible social and contextual
variables found to be important in the literature were
not measured in our study. These include coping skills,
perceived discrimination, poverty, historical trauma, and
community-level factors, such as neighborhood characteristics and social cohesiveness. Future studies should examine the influence of these and other social, cultural, and
political determinants of health.
Furthermore, the cross-sectional data used in these
analyses preclude the ability to make causal inferences.
Also, despite our efforts to encourage participation in the
study, only 42% of the students in the dormitories agreed
to participate. Therefore, it is likely that our study suffers from some self-selection bias, and we are unable
to adequately characterize those students who chose not
to participate. When asked why they refused to participate, some students replied that the interviews took too
much time and interfered with work or sports activities,
or that they were not interested. Future studies should explore methods of encouraging higher participation or seek
larger sample sizes to reduce systematic differences in
populations studied.
Another potential limitation was related to our interviewers. Interviewers who recognized potential participants recused themselves from those interviews. Even
with this precaution, however, our interviewers were demographically diverse, and their age or ethnicity may have
influenced the amount and type of information that participants were willing to share. Future research should
explore different methods of obtaining sensitive information, such as computer-assisted self-interviewing (CASI)
systems, that have proved successful in increasing reporting on these issues (Cooley et al., 2001; Turner et al.,
1998).
Despite these limitations, our study points to the need
to examine protective as well as risk factors among
AI youth. Much of the research on adolescent development has focused on problem behaviors, risk factors, and
deficits in youths’ lives, but a growing body of work has
begun to examine factors that may protect youth from debilitating consequences of potentially harmful behaviors.

1391

Reaching children and adolescents before they begin to
experiment with substances or engage in other unhealthy
behaviors is essential (Schinke, 1996). Furthermore, AI
youths’ identification with and participation in cultural
traditions are expected to have a positive influence on
their self-esteem because self-acceptance and self-worth
are manifestations of adolescents’ affinity to their cultural
background. Guyette (1982) suggests that enculturation
is especially significant for AI adolescent development
because it may moderate other factors associated with alcohol and substance use. Future studies examining the
protective processes that are culturally specific to AI youth
are needed and have the potential to teach us more about
the role of resilience in overcoming life challenges.
Declaration of Interest

The authors report no conflicts of interest. The authors
alone are responsible for the content and writing of the
article.
THE AUTHORS
Julie A. Baldwin (Cherokee),
Ph.D., Professor and Chair,
Department of Community
and Family Health, University
of South Florida, earned her
doctorate in Behavioral Sciences
and Health Education from
the Johns Hopkins University
School of Hygiene and Public
Health. Her research over the
years has focused on both
infectious and chronic disease
prevention. Cross-cutting themes
which have characterized her work include: utilizing communitybased participatory research approaches, working with AI
populations, and addressing health disparities by developing and
implementing culturally competent public health interventions.

Betty G. Brown, Ph.D.,
M.P.H, Assistant Professor,
Department of Health Sciences,
Northern Arizona University,
earned her Ph.D. in Medical
Sociology at Arizona State
University and a Master of
Public Health in Epidemiology
from Tulane University, School
of Public Health and Tropical
Medicine. Her research interests
include both interpersonal- and
community-level work in HIV,
mental health, brain injury, and disabilities, with a focus on rural
health.

1392

J. A. BALDWIN ET AL.

Heidi A. Wayment, Ph.D.,
Professor of Psychology,
Northern Arizona University,
is a health psychologist whose
current research examines
“quiet ego processes” or
characteristics related to a less
defensive stance toward the self
and others and how quiet ego
processes facilitate well-being.
She is especially interested in
how quiet ego characteristics
are affected by, and in turn
influence, individuals’ reactions to stressful life events.

Root Mean Square Error of Approximation (RMSEA): An
index of model fit, particularly useful in the comparison
of non-nested models (Kline, 1998).
Standardized Root Mean Squared Residual (SRMR): An
index of model fit based on average standardized covariance residuals, the difference between what is observed and what is implied by the model (Kline, 1998).
Structural equation modeling (SEM): This modeling approach is applicable in both exploratory and confirmatory analyses and allows for broad hypothesis testing
of relationships between observed and latent variables
(Kline, 1998).

REFERENCES
Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

Ramona Antone Nez
(Navajo/Iroquois Oneida),
M.P.H., B.S.N., Principal
Planner, Navajo Nation Division
of Health, earned her Master
of Public Health from the
University of Arizona and
BSN from Northern Arizona
University. She works on
planning, research, evaluation
and policy development with
Tribal Health Systems.

Kathleen M. Brelsford,
M.A., is a doctoral candidate
of medical anthropology
at the University of South
Florida, with a dual master’s
degree in public health. Her
main areas of research are
immunization decisions, STIs,
and reproductive health issues.

GLOSSARY

Bentler-Bonett Non-Normed Fit Index (NNFI): Another
relative index of model fit similar to CFI, but whose
value scale can be lower than that of other fit indexes
(Bentler, 1995; Kline, 1998).
Bentler Comparative Fit Index (CFI): A relative index of
model fit that compares the proportional improvement
of the model to the null. An advantage of CFI over other
model indexes is that it is less affected by the sample
size (Bentler, 1995; Kline, 1998).
Cultural identification: Cultural identification describes
the degree to which an individual feels connected to a
particular cultural way of life. According to the orthogonal model, strong identification with one culture does
not preclude strong identification with another (Oetting
and Beauvais, 1991).

Alexander, C. S., Somerfield, M. R., Ensminger, M. E., Kim, Y. J.,
& Johnson, K. E. (1995). Gender differences in injuries among
rural youth. Injury Prevention, 1(1), 15–20.
Bates, S. C., Beauvais, F., & Trimble, J. E. (1997). American Indian adolescent alcohol involvement and ethnic identification.
Substance Use and Misuse, 32(14), 2013–2031.
Beal, A. C., Ausiello, J., & Perrin, J. M. (2001). Social influences
on high-risk behaviors among minority middle school students.
Journal of Adolescent Health, 28(6), 474–480.
Beauvais, F. (1992). Drug use of friends: A comparison of reservation and non-reservation Indian youth. American Indian and
Alaska Native Mental Health Research, 5(1), 43–50.
Beauvais, F. (2000). Indian adolescence: Opportunity and challenge. Adolescent Diversity in Ethnic, Economic, and Cultural
Contexts, 10, 110–140.
Beauvais, F. (2001). Do school-based drug and alcohol abuse
prevention programs work in American Indian communities? In J. E. Trimble & F. Beauvais (Eds.), Health promotion and substance abuse prevention among American
Indian and Alaska Native Communities: Issues in cultural
competence (pp. 203–214). Rockville, MD: U.S. Department
of Health and Human Services, National Institute on Drug
Abuse.
Beauvais, F., & Oetting, E. R. (1991). The American drug and alcohol survey. Fort Collins, CO: CSU, Department of Psychology,
Tri-Ethnic Center for Prevention Research.
Beauvais, F., & Oetting, E. R. (1999). Drug use, resilience, and
the myth of the golden child. In M. D. Glantz & J. L. Johnson
(Eds.), Resilience and development: Positive life adaptation (pp.
101–107). New York: Kluwer.
Beebe, L. A., Vesely, S. K., Oman, R. F., Tolma, E., Aspy, C. B.,
& Rodine, S. (2008). Protective assets for non-use of alcohol,
tobacco, and other drugs among urban American Indian youth
in Oklahoma. Maternal and Child Health Journal, 12, S82–
S90.
Benally, C. J., Werito, M., Begay, D., Jones, T., & Yabeny,
V. (2000). Navajo high school youth risk behavior
surveillance system report, Retrieved June 1, 2011, from
http://www.yrbs.navajo/org/
Bentler, P. M. (1995). EQS: Structural equations program manual.
Los Angeles, CA: BMDP Statistical Software.
Brave Heart, M. Y. H. (1998). The American Indian Holocaust:
Healing historical unresolved guilt. American Indian and Alaska
Native Mental Health Research, 8(2), 56–78.
Brave Heart, M. Y. H. (1999). Gender differences in the historical
trauma response among the Lakota. Journal of Health and Social
Policy, 10(4), 1–21.

Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

ROLE OF CULTURE AND CONTEXT IN BUFFERING STRESS

Cameron, L. A. (1999). Understanding alcohol abuse in American Indian/Alaskan Native youth. Pediatric Nursing, 25(3),
297–302.
Christensen, M., & Manson, S. M. (2001). Adult attachment as a
framework for understanding mental health and American Indian families: A study of three family cases. American Behavioral Scientist, 44(9), 1447–1465.
Cooley, P. C., Rogers, S. M., Turner, C. F., Al-Tayyib, A. A., Willis,
G., & Ganapathi, L. (2001). Using touch-screen audio-CASI to
obtain data on sensitive topics. Computers in Human Behavior,
17(3), 285–293.
Dinges, N. G., & Duong-Tran, Q. (1993). Stressful life events
and co-occurring depression, substance abuse and suicidality
among American Indian and Alaska Native adolescents. Culture, Medicine and Psychiatry, 16, 487–502.
Federal Register. (2008, April 4). Department of the Interior, Bureau of Indian Affairs: Indian entities recognized and eligible to
receive services from the United States Bureau of Indian Affairs.
Federal Register Notices, 73(66), 18553.
Frank, M. L., & Lester, D. (2002). Self-destructive behaviors in
American Indian and Alaska Native high school youth. American Indian and Alaska Native Mental Health Research, 10(3),
24–32.
Grandbois, D. (2005). Stigma of mental Illness among American Indian and Alaska Native nations: Historical and contemporary perspectives. Issues in Mental Health Nursing, 26(10),
1001–1024.
Green, B. E., Sack, W. H., & Pambrum, A. (1981). A review
of child psychiatric epidemiology with special reference to
American Indian and Alaska Native children. White Cloud Journal, 2(2), 22–36.
Greene, J. P., & Forster, G. (2003). Public high school graduation
and college readiness rate in the United States. Education Working Paper 3. New York, NY: Center for Civic Innovation at the
Manhattan Institute for Policy Research.
Guyette, S. (1982). Selected characteristics of American Indian
substance users. International Journal of the Addictions, 17(6),
1001–1014.
Hamill, S. K., Scott, W. D., Dearing, E., & Pepper, C. M. (2009).
Affective style and depressive symptoms in youth of a North
American Plains tribe: The moderating roles of cultural identity,
grade level, and behavioral inhibition. Personality and Individual Differences, 47(2), 110–115.
Hill, D. L. (2009). Relationship between sense of belonging as connectedness and suicide in American Indians. Archives of Psychiatric Nursing, 23(1), 65–74.
Hofboll, S., Jackson, A., Hofboll, I., Pierce, C. A., & Young, S.
(2002). The impact of communal-mastery versus self-mastery
on emotional outcomes during stressful conditions: A prospective study of Native American women. American Journal of
Community Psychology, 30, 853–871.
Hurdle, D. E., Okamoto, S. K., & Miles, B. (2003). Family influences on alcohol and drug use by American Indian youth: Implications for prevention. Journal of Family Social Work, 7(1),
53–68.
King, J., Beals, J., & Manson, S. M. (1992). A structural equation
model of factors related to substance use among American Indian adolescents. Drugs & Society, 3–4, 253–268.
Kline, R. B. (1998). Principles and practice of structural equation
modeling. New York: Guilford.
Kumar, R., O’Malley, P. M., & Johnston, L. D. (2002). Effects of
school-level norms on student substance use. Prevention Science, 3(2), 105–124.

1393

Kunitz, S. J. (2006). Life-course observations of alcohol use among
Navajo Indians: Natural history of careers? Medical Anthropology Quarterly, 20(3), 279–296.
LeMaster, P. L., Beals, J., Novins, D. K., & Manson, S. M.
(2004). The prevalence of suicidal behaviors among Northern
Plains American Indians. Suicide and Life-Threatening Behaviors, 34(3), 242–254.
Lewinsohn, P. M., Roberts, R. E., Seeley, J. R., Rhode, P., Gotlib,
I. H., & Hops, H. (1994). Adolescent psychopathology: II. Psychosocial risk factors for depression. Journal of Abnormal Psychology, 103, 302–315.
Lewton, E. L., & Bydone, V. (2000). Identity and healing in three
Navajo religious traditions: Sa’ah Naaghai Bik’eh Hozh. Medical Anthropology Quarterly, 14(4), 476–97.
Manson, S. (1997). Ethnographic methods, cultural context, and
mental illness: Bridging different ways of knowing and experience. Ethos, 25(2), 249–258.
May, P. A. (1994). The epidemiology of alcohol abuse among
American Indians: The mythical and real properties. American
Indian Culture and Research Journal, 18(2), 121–143.
Mehrabadi, A., Patterson, K., Pearce, M., Patel, S., Craib, K. J.
P., Moniruzzaman, A., et al. (2008). Gender differences in HIV
and Hepatitis C related vulnerabilities among aboriginal young
people who use street drugs in two Canadian cities. Women &
Health, 48(3), 235–260.
Miller, K. A., Beauvais, F., Burnside, M., & Jumper-Thurman,
P. (2008). A comparison of American Indian and Non-Indian
fourth to sixth graders’ rates of drug use. Journal of Ethnicity in
Substance Abuse, 7(3), 258–267.
Mitchell, C. M., O’Neill, T. D., & Beals, J. (1996). Dimensionality of alcohol use among American Indian adolescents: Latent structure, construct validity, and implications for developmental research. Journal of Research on Adolescence, 6(2),
151–180.
Mohatt, G. V., Rasmus, M. S., Thomas, L., Allen, J., Hazel, K.,
& Hensel, C. (2004). Tied together like a woven hat: Protective
pathways to Alaska native sobriety. Harm Reduction Journal, 1,
1–12.
Newcomb, M. D. (1990). What structural equation modeling can
tell us about social support. In B. R. Sarason, I. G. Sarason, &
G. R. Pierce (Eds.), Social support: An interactional view (pp.
26–63). New York: John Wiley.
Noe, T., Fleming, C., & Manson, S. (2003). Healthy nations:
Reducing substance abuse in American Indian and Alaska
Native communities. Journal of Psychoactive Drugs, 35(1),
15–25.
Novins, D. K., Beals, J., & Roberts, R. E. (1999). Factors associated
with suicide ideation among American Indian adolescents: Does
culture matter? Suicide and Life-Threatening Behavior, 29(4),
332–346.
O’Brien, B. L., Anslow, R. M., Begay, W., Pereira, S. B. A., & Sullivan, M. P. (2002). 21st century rural nursing: Navajo traditional
and western medicine. Nursing Administration Quarterly, 26(5),
47–57.
Oetting, E. R., & Beauvais, F. (1991). Orthogonal cultural identification theory: The cultural identification of minority adolescents. International Journal of the Addictions, 24(5A–6A), 655–
685.
Pearce, M. E., Christian, W. M., Patterson, K., Norris, K., Moniruzzaman, A., Craib, K. J. P., et al. (2008). The Cedar Project: Historical trauma, sexual abuse and HIV risk among young aboriginal people who use injection and non-injection drugs in two
Canadian cities. Social Science & Medicine, 66(11), 2185–2194.

Subst Use Misuse Downloaded from informahealthcare.com by University of Montreal on 09/24/11
For personal use only.

1394

J. A. BALDWIN ET AL.

Petoskey, E. L., Van Stelle, K. R., & De Jong, J. A. (1998). Prevention through empowerment in a Native American community.
Drugs & Society, 12(1–2), 147–162.
Procidano, M. E., & Heller, K. (1983). Measures of perceived social support from friends and family: Three validation studies.
American Journal of Community Psychology, 11, 1–23.
Reynolds, W. (1988). Suicidal ideation questionnaire: Professional
manual. Odessa, FL: Psychological Assessment Resources.
Richmond, C., Ross, N. A., & Egeland, G. M. (2007). Societal resources and thriving health: A new approach for understanding
the health of indigenous Canadians. American Journal of Public
Health, 97(10), 1827–1833.
Richmond, C. A. M., & Ross, N. A. (2008). Social support, material circumstance and health behaviour: Influences on health in
First Nation and Inuit communities in Canada. Social Science &
Medicine, 67(9), 1423–1433.
Rieckmann, T. R., Wadsworth, M. E., & Deyhle, D. (2004). Cultural identity, explanatory style, and depression in Navajo adolescents. Cultural Diversity and Ethnic Minority Psychology,
10(4), 365–382.
Robbins, R., Colmant, S., Dorton, J., Schultz, Y. L. C., & Ciali,
P. (2006). Colonial instillations in American Indian boarding
school students. Educational Foundations, 20, 3–4.
Robin, S. S., & Johnson, E. O. (1996). Attitude and peer cross pressure: Adolescent drug and alcohol use. Journal of Drug Education, 26(1), 69–99.
Roubideaux, Y. (2002). Current issues in Indian health
policy (Rev. ed.). Udall Center for Studies in Public
Policy, University of Arizona. Retrieved September 7,
2005, from http://udallcenter.arizona.edu/publications/pdf/
indian health policy.pdf
Schinke, S. (1996). Behavioral approaches to illness prevention for
Native Americans. In P. M. Kato & T. Mann (Eds.), Handbook of
diversity issues in health psychology (pp. 367–386). New York:
Plenum Press.
Scott, W. D., Dearing, E., Reynolds, W. R., Lindsay, J. E., Baird, G.
L., & Hamill, S. K. (2008). Cognitive self-regulation and depression: Examining academic self-efficacy and goal characteristics
in youth of a northern plains tribe. Journal of Research on Adolescence, 18(2), 379–394.
Shaughness, L., & Jones, S. E. (2003). Tobacco, alcohol, and
other drug use among high school students in Bureau of Indian Affairs-funded schools—United States, 2001. Morbidity
and Mortality Weekly Report, 52(44), 1070–1072.
Shaughnessy, L., Doshi, S. R., & Jones, S. E. (2004). Attempted suicide and associated health risk behaviors among Native American high school students. Journal of School Health, 74(5),
177–182.
Spicer, P., Novins, D. K., & Mitchell, C. M. (2003). Aboriginal social organization, contemporary experience and American Indian adolescent alcohol use. Quarterly Journal of Studies on Alcohol, 64(4), 450–457.
Strickland, C., Walsh, E., & Cooper, M. (2006). Healing fractured
families: Parents’ and elders’ perspectives on the impact of colonization and youth suicide prevention in a Pacific Northwest
American Indian tribe. Journal of Transcultural Nursing, 17(1),
5–12.
Swaim, R. C., Nemeth, J., & Oetting, E. R. (1995). Alcohol use and
socialization characteristics among Hungarian adolescents: Path
models. Drugs & Society, 8(3–4), 47–63.
Swaim, R. C., Oetting, E. R., Edwards, R. W., & Beauvais, F.
(1989). Links from emotional distress to adolescent drug use:

A path model. Journal of Consulting and Clinical Psychology,
57(2), 227–231.
Swaim, R. C., Oetting, E. R., & Thurman, P. J. (1993). American
Indian adolescent drug use and socialization characteristics: A
cross-cultural comparison. Journal of Cross-Cultural Psychology, 24(1), 53–70.
Trimble, J. E. (1995). Toward an understanding of ethnicity, ethnic identity, and their relationship with drug use research. In G.
Botvin, S. Schinke, & M. Orlandi (Eds.), Drug abuse prevention
with multi-ethnic youth (pp. 3–27). Newbury Park, CA: Sage.
Turner, C. F., Ku, L., Rogers, S. M., Lindberg, L. D., Pleck, J. H., &
Sonenstein, F. L. (1998). Adolescent sexual behavior, drug use,
and violence: Increased reporting with computer survey technology. Science, 280, 867–873.
Vernon, I. S., & Jumper-Thurman, P. (2002). Prevention of
HIV/AIDS in Native American communities: Promising interventions. Public Health Reports, 117(1), S96–S103.
Waldvogel, J. L., Rueter, M., & Oberg, C. N. (2008). Adolescent
suicide: Risk factors and prevention strategies. Current Problems in Pediatric and Adolescent Health Care, 38(4), 110–
125.
Wallace, J. M., Backman, J. G., O’Malley, P. M., Johnston, L. D.,
Schulenberg, J. E., & Cooper, S. M. (2002). Tobacco, alcohol, and illicit drug use: Racial and ethnic differences among
U.S. high school seniors, 1976–2000. Public Health Reports,
117(S1): S67–S75.
Wallace, J. M., Backman, J. G., O’Malley, P. M., Schulenberg,
J. E., Cooper, S. M., & Johnston, L. D. (2003). Gender and ethnic differences in smoking, drinking, and illicit drug use among
American 8th, 10th, and 12th grade students, 1976–2000. Addiction, 98(2), 225–239.
Walters, K. L., Simoni, J. M., & Evans-Campbell, T. (2002). Substance use among American Indians and Alaska Natives: Incorporating culture in an “Indigenist” stress-coping paradigm. Public Health Reports, 117(1), S104–S117.
Westermeyer, J. (2001). Alcoholism and co-morbid psychiatric disorders among American Indians. American Indian and Alaska
Native Mental Health Research, 10(2), 27–51.
Wiegman Dick, R., Manson, S. M., & Beals, J. (1993). Alcohol use
among male and female Native American adolescents: Patterns
and correlates of student drinking in a boarding school. Journal
of Studies on Alcohol, 54, 172–177.
Whitbeck, L. B., Adams, G. W., Hoyt, D. R., & Chen, X. (2004).
Conceptualizing and measuring historical trauma among American Indian people. American Journal of Community Psychology,
33(3–4), 119–130.
Whitbeck, L. B., Hoyt, D. R., McMorris, B. J., Xiaojin, C.,
& Stubben, J. D. (2001). Perceived discrimination and early
substance abuse among American Indian children. Journal of
Health and Social Behavior, 42, 405–424.
Whitbeck, L. B., McMorris, B. J., Hoyt, D. R., Stubben, J. D., &
LaFromboise, T. (2002). Perceived discrimination, traditional
practices, and depressive symptoms among American Indians
in the Upper Midwest. Journal of Health and Social Behavior,
43, 400–418.
Whitbeck, L. B., Yu, M., McChargue, D. E., & Crawford, D. M.
(2009). Depressive symptoms, gender and growth in cigarette
smoking indigenous adolescents. Addictive Behaviors, 34(5),
421–426.
Yu, M., Stiffman, A. R., & Freedenthal, S. (2005). Factors affecting
American Indian adolescent tobacco use. Addictive Behaviors,
30(5), 889–904.



Documents similaires


10826084 2011 592432
adolescent mental health
the role of media in childhood obesity
h7uzrat
motivational speakers for youth
sexualviolence


Sur le même sujet..