Differential Acculturation and Adolescent Health Outcomes 1
RUNNING HEAD: DIFFERENTIAL ACCULTURATION AND ADOLESCENT HEALTH OUTCOMES
In press, Drug and Alcohol Dependence.
Substance Use and Sexual Behavior among Recent Hispanic Immigrant Adolescents:
Effects of Parent‐Adolescent Differential Acculturation and Communication
Seth J. Schwartz1, Jennifer B. Unger2, Sabrina E. Des Rosiers1, Shi Huang1, Lourdes Baezconde‐Garbanati2,
Elma Lorenzo‐Blanco3, Juan A. Villamar1, Daniel W. Soto2, Monica Pattarroyo2, and José Szapocznik1
1University of Miami, 2University of Southern California, 3University of Michigan
Preparation of this article was supported by Grant DA025694, co‐funded by the National Institute of
Drug Abuse and the National Institute on Alcohol Abuse and Alcoholism. Please address correspondence
to Seth J. Schwartz, Department of Epidemiology and Public Health, Leonard M. Miller School of
Medicine, University of Miami, 1425 N.W. 10th Avenue, Miami, FL33136, USA (email
Differential Acculturation and Adolescent Health Outcomes 2
Objectives: To ascertain the effects of parent‐adolescent acculturation gaps, perceived
discrimination, and perceived negative context of reception on adolescent cigarette smoking, alcohol
use, sexual activity, and sexual risk taking. We used an expanded, multidimensional model of
Method: A sample of 302 recently immigrated parent‐adolescent dyads (152 from Miami and 150
from Los Angeles) completed measures of acculturation (Hispanic and American practices and
identifications, and individualist and collectivist values) and parent‐adolescent communication.
Adolescents completed measures of recent cigarette smoking, alcohol use, sexual behavior, and sexual
Results: Parent‐adolescent gaps in American practices and ethnic identity, and perceptions of a
negative context of reception, predicted compromised parent‐adolescent communication. In Miami
only, adolescent‐reported communication negatively predicted odds of cigarette smoking, occasions of
drunkenness, and number of sexual partners. Also in Miami only, parent‐reported communication
positively predicted these outcomes, as well as occasions of adolescent binge drinking, drunkenness,
number of sexual partners, and odds of unprotected sex. The only significant findings in Los Angeles
were protective effects of parent‐reported communication on frequency of alcohol use and of binge
drinking. Mediational effects emerged only in the Miami sample.
Conclusions: Effects of parent‐adolescent acculturation gaps vary across Hispanic groups and
receiving contexts. The especially strong parental control in many Mexican families may account for
these differences. However, other important differences between Hispanic subgroups and communities
of reception could also account for these differences. Prevention efforts might encourage Hispanic
youth both to retain their culture of origin and to acquire American culture.
KEY WORDS: Hispanic, acculturation, discrimination, alcohol use, recent immigrants.
Differential Acculturation and Adolescent Health Outcomes 3
Important health disparities exist between Hispanics and non‐Hispanic White early and middle
adolescents, including cigarette and alcohol use (Johnston et al., 2011) and sexual risk taking (CDC,
2011). Specifically, in 2010, 59% of Hispanic 10th graders, compared to 53% of White 10th graders, had
used alcohol in the past year (Johnston et al., 2011); and 55% of Hispanic high school students,
compared to 63% of Whites, reported using a condom at last sexual intercourse (CDC, 2011). Identifying
cultural predictors of these disparities, and of the behaviors associated with them, could guide the
development of prevention programs to reduce health disparities (Krieger, 2012).
1.1. Acculturation and Health Outcomes in Hispanics
Although the construct of acculturation has a long history, the majority of public health studies have
used measures and models that do not reflect the lived reality of immigrants (Thomson and Hoffman‐
Goetz, 2009). As defined within cultural studies, acculturation is a bidimensional process, in which
heritage‐culture retention and American‐culture acquisition represent separate dimensions (e.g., Berry,
1997). Previous studies vary in their conceptualization of the domains of acculturation, including
attitudes (Berry, 1980), cultural practices such as language use, culinary preferences, media, and choice
of friends (Szapocznik et al., 1980); and ethnic identity (Phinney, 2003). Cultural values (e.g.,
individualism, collectivism) are seldom included as indices of acculturation.
Schwartz et al. (2010) proposed an integrative, multidimensional perspective, positing cultural
practices, values, and identifications as domains of acculturation; and heritage and receiving cultural
streams as operating within each of these domains. For example, for U.S. Hispanics, acculturation
includes Hispanic practices, American practices, collectivist values, individualist values, Hispanic identity,
and American identity. Each process may be differently linked with substance use and sexual risk
behavior, although retention of Hispanic cultural practices and values typically is considered protective
(Schwartz et al., 2011).
Differential Acculturation and Adolescent Health Outcomes 4
Acculturation is especially challenging for recent immigrant families (Smokowski and Bacallao,
2011). Adolescents and parents may face discrimination and a hostile context of reception in their new
homelands (Portes and Rumbaut, 2006). Moreover, recent‐immigrant adolescents acculturate to the
receiving society quickly, whereas parents do not (Schwartz et al., 2006b), leading to differential
1.1.a. Differential Acculturation. Because Hispanic children and adolescents attend school in the
U.S., they typically gravitate toward U.S. culture, learning to function both in the receiving society and in
their often traditionally‐oriented families and communities (Padilla, 2006). Conversely, adults –
especially those living in ethnic enclaves – may function well using their native languages and customs
and may not “acculturate” much (e.g., Schwartz et al., 2006b). Consequently, as hypothesized by
Szapocznik and colleagues (1978), parent‐child acculturation gaps emerge. These gaps may be a function
of both greater exposure to U.S. culture and greater cultural plasticity among children and adolescents
than among adults (Cheung et al., 2011). Regardless of their source, cultural gaps can be measured as
the parent‐adolescent difference in a given acculturation‐related variable (Telzer, 2010). Such an
approach uses data from multiple reporters to create a gap score, as opposed to asking parents or
adolescents to report on the “acculturation gap” in their families.
The effects of acculturation gaps on adolescent outcomes have received some empirical attention.
Szapocznik and Kurtines (1980) found that Cuban families with troubled and substance abusing
adolescents had large parent‐adolescent acculturation gaps. These gaps were believed to exacerbate
problems with parent‐adolescent communication and other aspects of family functioning, which in turn
would predict adolescent substance use and sexual risks (Szapocznik and Kurtines, 1993).
Studies have examined parts of this “differential acculturation hypothesis” (Telzer, 2011).
Smokowski et al. (2008) found associations of acculturation gaps with compromised family adaptability
and cohesion, but they did not investigate links with adolescent outcomes. Martinez (2006) found that
Differential Acculturation and Adolescent Health Outcomes 5
differential acculturation predicted behavior problems and substance use, whereas Lau et al. (2005) did
not. However, neither of these studies examined family processes as mediators. Unger et al. (2009)
found that parent‐adolescent discrepancies in American cultural practices predicted low family
cohesion, which predicted adolescent substance use. However, parents’ acculturation was assessed via
adolescents’ reports. Telzer (2011) concluded that families where adolescents are more acculturated
toward the U.S. than their parents may be less problematic than families where adolescents lose their
cultural heritage while parents retain it. Although most acculturation gap research has focused on
cultural practices, we expected the pattern identified by Telzer (2011) to apply to cultural values and
identifications as well.
1.2. Perceived Discrimination and Context of Reception
Perceived discrimination and negative perceived context of reception are additional cultural
variables associated with health outcomes. Discrimination includes others’ actions that cause one to feel
unwanted, stereotyped, or demeaned (Lee, 2005). Perceived discrimination may have long‐term health
consequences for Hispanics, including hypertension, depression, diabetes, cardiovascular illness, and
other health problems (Finch and Vega, 2003; Todorova et al., 2010). Context of reception refers to
immigrants’ opportunities in the U.S. A negative context of reception may be discouraging to immigrants
(Portes and Rumbaut, 2006), especially if the receiving society systematically denies immigrants
opportunities available to members of the dominant group (Leong, 2008; Steiner, 2009). Among
Hispanics, Mexicans and Puerto Ricans are often marginalized, whereas Cubans generally fare well,
especially in Miami (Stepick and Stepick, 2002). Unlike Mexicans, many of whom are undocumented and
seek “under‐the‐table” positions (Henderson, 2011), and Puerto Ricans, many of whom migrate to the
Northeast and South to escape poverty (Acosta‐Belen and Santiago, 2006), many Cubans arrive in the
U.S. as political refugees – though some do immigrate to escape poverty. The original cohort of Cubans
settled in Miami and claimed positions of political and economic power (Stepick et al., 2003). Although
Differential Acculturation and Adolescent Health Outcomes 6
Miami is also home to many Central and South Americans, Cubans remain the dominant Hispanic group.
Thus, the context of reception in Miami differs from that in other parts of the U.S. Multi‐city
comparisons between Miami and other U.S. cities – including Cubans as well as other Hispanic groups –
may be useful in examining the effects of context of reception (e.g., Schwartz, Unger, et al., 2012).
One such city is Los Angeles, which is home to more than 2 million individuals of Mexican ancestry
(Hayes‐Bautista, 2004). Some Mexican‐descent individuals in Los Angeles can trace their lineage to the
Mexican territories that were annexed by the United States after the Mexican‐American War, whereas
others are recent or second‐generation immigrants. Although the Los Angeles area already ranked
second in Hispanic population size in 2000, this population grew by more than 20% between 2000 and
2010 (Ennis et al., 2011), due in part to immigration (Walters and Trevelyan, 2011). Some Mexican
Americans are gaining power in Los Angeles, though the majority still live at or near the poverty level.
Miami and Los Angeles, as two very different receiving communities for Hispanic immigrants, served
as contexts for the present study. Multisite studies of acculturation and health outcomes are important
because acculturation may take different forms depending on the context to which individuals are
acculturating (Alba and Nee, 2006). We examined multiple domains of acculturation (heritage and U.S.
practices, values, and identifications) in relation to substance use, sexual activity, and unprotected sex in
recently arrived Hispanic immigrant adolescents in these two cities. Given the importance of
discrimination and context of reception for immigrant and minority health (Krieger, 2012), we also
examined these cultural processes as predictors of substance use and sexual outcomes. Finally, in light
of the role of family dynamics in the effects of cultural processes on health outcomes (Smokowski et al.,
2008; Unger et al., 2009), we posited parent‐adolescent communication as a mediating mechanism.
1.3. The Present Study
In the present longitudinal study, we sampled recent‐immigrant parent‐adolescent dyads from
Miami and Los Angeles. Each parent‐adolescent dyad completed measures of Hispanic and American
Differential Acculturation and Adolescent Health Outcomes 7
cultural practices, values, and identifications; perceived discrimination; perceived negative context of
reception; and parent‐adolescent communication. We used parent‐adolescent communication because
it captures the affective valence of family interactions (Guilamo‐Ramos et al., 2006). Adolescents
completed measures of smoking, alcohol use, and sexual behavior and risk taking.
We hypothesized that both adolescent and parent reports of communication would negatively
mediate the effects of acculturation gaps on adolescent‐reported substance use, sexual activity, and
unprotected sex (Figure 1). Specifically, Following Telzer (2011), we anticipated that (1) heritage‐cultural
gaps – those involving Hispanic practices, collectivist values, and ethnic identity – would be most
strongly and negatively linked with parent‐adolescent communication; and (2) adolescent and parent
reports of communication would negatively predict substance use, sexual activity, and sexual risk taking.
Because adolescent and parent reports of family processes tend to be only modestly intercorrelated
(Schwartz et al., 2005), we included both adolescent and parent reports of communication as mediators.
We analyzed data separately for the Miami and Los Angeles samples, because these Hispanic
subpopulations differ in terms of social position within their receiving community (i.e., positions of
power in Miami versus marginalized positions in Los Angeles; Hayes‐Bautista, 2004; Stepick and Stepick,
2002). Indeed, as reported in Schwartz et al. (2012), in our sample, Los Angeles parents reported a
significantly more negative context of reception – as well as lower levels of education – compared to
Participants were 302 parent‐adolescent dyads from Miami (N=152) and Los Angeles (N=150). Each
adolescent participated with a primary parent/caregiver. Although not all of the caregivers were the
adolescents’ biological parents, we use the term “parent” for simplicity. Participating parents were
mothers (70%), fathers (25%), stepparents (3%), and grandparents/other relatives (2%). These data
Differential Acculturation and Adolescent Health Outcomes 8
represent the first two assessment timepoints. Miami families were primarily Cuban (61%); Los Angeles
families were primarily Mexican (70%). Adolescents were in or entering 9th grade (mean age=14.51
years, SD=0.88, range 14‐17; 53% boys). Parents’ mean age was 41.09 years (SD = 7.13, range 22‐64);
77% were married or cohabiting with a partner. Miami families were more recent immigrants (Mdn = 1
year in the U.S.), and Miami parents were more likely (69.8%) to have graduated high school, compared
to Los Angeles families (Mdn = 3 years in the U.S. and 40.0% high school graduates). Table 1 provides
significant demographic differences between the Miami and Los Angeles samples.
2.2.a. School Selection and Participant Recruitment. Families were recruited from randomly
selected schools in Miami‐Dade and Los Angeles Counties. Because many Hispanic recent immigrants
live in heavily Hispanic areas (Kasinitz et al., 2008; Stepick et al., 2003), we selected schools that were
≥75% Hispanic (10 schools in Miami and 13 in Los Angeles). The study was approved by the IRBs at the
University of Miami and the University of Southern California, and by the participating school districts.
We recruited students from English for Speakers of Other Languages (ESOL) classes and from the overall
student body. Interested students provided their parent/guardian’s phone numbers.
Study staff contacted parents to screen them for eligibility: adolescent lived in the US ≤5 years,
adolescent in or entering 9th grade, and family planned to remain in South Florida or Southern California
during the four years of the study. Families who met these inclusion criteria were consented and
scheduled for evening or weekend assessment at convenient locations. Of the 632 families who
provided phone numbers, 197 were unreachable (incorrect or disconnected numbers). Of the remaining
435 families, 302 (69%) participated. This participation rate does not include families whom we
attempted to contact but were unable to reach, because we do not know how many of them would
have been eligible to participate. Time 2 assessments occurred 6‐8 months following the baseline
assessments. Of the original sample of 302 families, 278 were reassessed. Although families were
Differential Acculturation and Adolescent Health Outcomes 9
recruited from schools, we initiated follow‐up contacts directly with parents. We did this so that
adolescents who dropped out of school, or who switched schools, could still be retained in the sample.
2.2.b. Assessment Procedures. At baseline, due to closer proximity of families’ homes to the
research center and schools in Miami compared to Los Angeles, all Miami families were assessed at the
research center (66%) or at their adolescent’s school (34%). Most Los Angeles families were assessed in
their homes (46%) or convenient community locations (36%). At Time 2, the number of home and
community assessments increased at both sites, to 27% in Miami and 91% in Los Angeles. Each parent
received $40 at baseline and $45 at Time 2. Adolescents received movie tickets at each timepoint.
Before the baseline assessment, parents provided informed consent for themselves and their
adolescents, and adolescents provided informed assent, in separate rooms. Assessments were
completed on laptop computers (for adolescents) or on touch‐screen tablet PCs (for parents). An audio
computer‐assisted interviewing (A‐CASI) system (Turner et al., 1998; Cooley et al., 2003) was used to
administer surveys. Most parents (98%) and adolescents (84%) completed their baseline assessments in
Spanish; 13% of adolescents (but no parents) switched languages at Time 2.
2.3. Measures – Predictors at Baseline
2.3.a. Acculturation. Hispanic and American cultural practices were measured using the Bicultural
Involvement Questionnaire (Szapocznik et al., 1980), which includes 24 items, 12 assessing American
practices and 12 assessing Hispanic practices. Cronbach’s alphas for adolescents and parents,
respectively, were .91 and .91 for American practices and .89 and .86 for Hispanic practices. Sample
items include “I speak English at home” and “I enjoy Hispanic‐oriented places”.
Cultural values were assessed using a 16‐item measure of individualism and collectivism (Triandis,
1995; Triandis and Gelfand, 1998). Sample items include “I’d rather depend on myself than on others”
(individualism); and “Family members should stick together, no matter what sacrifices are required”
(collectivism). Items are rated on a five‐point Likert Scale ranging from 1 (Strongly Disagree) to 5
Differential Acculturation and Adolescent Health Outcomes 10
(Strongly Agree). Alpha coefficients for adolescents and parents, respectively, were: individualism, .73
and .74; and collectivism, .79 and .70.
Hispanic/ethnic and American identifications were assessed using parallel versions of the Multi‐
Group Ethnic Identity Measure (Roberts et al., 1999). In the American‐identity version, we used “the
United States” instead of “my ethnic group”. Sample items include “I am proud to be a member of my
ethnic group.” Cronbach’s alphas for adolescents and parents, respectively, were .88 and .88 for
American identity and .91 and .89 for ethnic identity.
2.3.b. Perceived discrimination was assessed with seven items (Phinney et al., 1998) tapping into
perceptions of unfair treatment based on ethnicity (e.g., “How often do teachers or employers treat you
unfairly or negatively because of your ethnic background?”). The 5‐point Likert scale ranges from 0 (Not
at all) to 4 (Almost always). Cronbach’s alphas were .89 for adolescents and .87 for parents.
2.3.c. Perceived Negative Context of Reception was assessed using a 6‐item scale (Schwartz, Unger,
et al. 2012). Items assess the extent to which the opportunity structure (e.g., employment or grades) do
not favor one’s ethnic group. A sample item is “People from my country are not welcome here.”
Cronbach’s alphas were .83 for adolescents and .88 for parents.
2.3.d. Parent‐adolescent communication was assessed using the Parent‐Adolescent Communication
Scale (Barnes and Olson, 1982). The adolescent and parent versions each contain 20 items (α=.94 for
adolescents and .85 for parents) measuring listening and trust. Adolescents were asked to report on
their relationship with the parent figure in the study with them.
2.4. Measures – Outcomes at Time 2
2.4.a. Cigarette, Alcohol, and Illicit Drug Use was assessed using an adaptation of the Monitoring
the Future survey (Johnston et al., 2011). We asked about frequency of cigarette use, alcohol use, heavy
drinking (binge drinking and largest number of drinks consumed in a day), drunkenness, and illicit drug
Differential Acculturation and Adolescent Health Outcomes 11
use in the previous 90 days. Because illicit drug use was reported only by 8 adolescents at Time 2, we did
not analyze this outcome.
2.4.b. Sexual Risk Taking. Using items from the Sexual Behavior Instrument (Jemmott et al., 1998),
participants self‐reported how many times in the previous 90 days they had engaged in oral, vaginal,
and anal sex; unprotected oral, anal, or vaginal sex; and number of sexual partners. Because the mean
age of the sample was 14.86 years at Time 2, sexual activity at this age can be considered precocious
and risky (Dillon et al., 2010).
2.5. Analytic Plan
The analytic plan included three steps. First, we reported descriptive statistics for substance use and
sexual behavior/risk taking. Second, we created parent‐adolescent difference scores for each
acculturation variable and modeled these difference scores, along with perceived discrimination and
negative context of reception, as predictors of baseline parent‐adolescent communication – which was
then modeled as a predictor of substance use and sexual behavior/risks. Because we had observed
measures for each of the acculturation dimensions, difference scores for each American acculturation
dimension were created by subtracting the parent’s score from the adolescent’s score, and difference
scores for each Hispanic dimension were created by subtracting the adolescent’s score from the
parent’s score (Thomas and Zumbo, 2012). Therefore, the gap score would be positive (i.e., parents
more highly endorsing Hispanic culture, and adolescents more highly endorsing American culture) for
most families. Third, using structural equation modeling and mediation tests (MacKinnon, 2008), we
examined the extent to which parent‐adolescent communication mediated the effects of the differential
acculturation variables and of perceived discrimination and context of reception on the outcomes. All of
the predictors and outcomes were included in one structural equation model to avoid inflated Type I
error. Each step was conducted separately for the Miami and Los Angeles sub‐samples, because of
important differences between these two receiving contexts (Rumbaut and Portes 2001; Schwartz et al.,
Differential Acculturation and Adolescent Health Outcomes 12
2012). For all analyses other than descriptive statistics, the sandwich estimator (Kauermann and Carroll,
2001) was used to adjust standard errors for the effects of multilevel nesting (families within schools).
Number of occasions of alcohol use, largest number of drinks consumed in a single day, number of
binge‐drinking days, and occasions of drunkenness were entered into the model as count variables.
Cigarette smoking was dichotomized because only 15 adolescents (6%) reported smoking in the previous
3.1. Descriptive Statistics
Table 2 displays baseline descriptive statistics for predictor variables. Miami adolescents (primarily
of Cuban origin) scored higher than their Los Angeles (primarily of Mexican origin) counterparts on
Hispanic practices, collectivist and individualist values, and ethnic and American identity. For parents,
collectivist values and American identity were higher in Miami than Los Angeles.
Table 3 presents frequencies of outcome variables by site. Alcohol was the most frequently used
substance at Time 2 (13.0% of Miami adolescents and 8.3% of Los Angeles adolescents reported drinking
alcohol in the 90 days prior to the Time 2 assessment). By Time 2, 32% of Miami adolescents had
engaged in vaginal, oral, or anal sex at least once, compared to 13% of Los Angeles adolescents.
Nineteen percent of Miami adolescents, compared to 6% of Los Angeles adolescents, had been sexually
active in the 90 days prior to Time 2. Follow‐up analyses indicated that, in Miami, Cubans were more
likely (37%) to report having initiated sex compared to adolescents from other countries (24%); whereas
in Los Angeles, Mexicans were less likely (10%) to report having initiated sex compared to adolescents
from other countries (20%).
3.2. Hypothesis Tests: Differential Acculturation Predicting Health Behaviors, as Mediated by Parent‐
Differential Acculturation and Adolescent Health Outcomes 13
3.2.a. Estimating the Structural Equation Model. The structural equation model (Figure 1) included
baseline acculturation gaps and parent‐adolescent communication, and Time 2 health outcomes. For
each city, we estimated (a) effects of acculturation discrepancies on parent‐adolescent communication
at baseline; (b) effects of baseline parent‐adolescent communication on Time 2 adolescent outcomes;
and (c) mediated effects of acculturation discrepancies on adolescent outcomes through parent‐
adolescent communication. Attempts to add direct effects of acculturation discrepancies on adolescent
outcomes to model (c) created multicolinearity (i.e., unusually large standard errors prevented sizeable
regression coefficients from reaching statistical significance; Tabachnick and Fidell, 2007). We therefore
examined only indirect effects of acculturation discrepancies on adolescent outcomes.
Standard structural equation modeling fit indices (e.g., χ2, CFI, RMSEA) are not available for models
with count variables (Muthén and Muthén, 2010). The results of the structural equation analyses are
described below and in Figure 2.
3.2.c. Differential Acculturation Predicting Parent‐Adolescent Communication. Table 4 presents
associations of differential acculturation variables with adolescent and parent reports of
communication. In Miami, for both parents and adolescents, differential collectivism was significantly
related to adolescent‐reported communication (negatively for adolescents and positively for parents);
and at both sites, parents’ perceptions of negative context of reception were negatively related to their
reports of communication with their adolescents. For adolescent reports, negative context of reception
was a negative predictor of parent‐adolescent communication in Miami, but perceived discrimination
was a negative predictor of parent‐adolescent communication in Los Angeles.
3.2.d. Parent‐Adolescent Communication Predicting Outcomes. Table 5 presents patterns of
longitudinal associations between parent‐adolescent communication and outcomes. In Miami,
adolescent‐reported communication was protective against cigarette use, occasions of drunkenness,
number of sexual partners, and number of oral sex partners. The effect for binge drinking days
Differential Acculturation and Adolescent Health Outcomes 14
approached significance. In Los Angeles, adolescent‐reported communication did not reach significance
as a predictor of any of the outcomes, although the protective effect for cigarette smoking approached
In Miami, parents’ reports of communication with their adolescents were positively predictive of
number of binge drinking days, occasions of drunkenness, number of sexual partners, number of oral
sex partners, inconsistent condom use, and unprotected oral sex. In Los Angeles, the only significant
finding for parent‐reported communication was a negative effect on alcohol use occasions (the effect
for binge drinking days approached significance).
3.2.e. Mediation Tests. We tested whether parent‐adolescent communication mediated the effects
of differential acculturation on health outcomes, using the asymmetric distribution of products test
(MacKinnon, 2008) and PRODCLIN software (MacKinnon et al., 2007). All but one of the significant
mediated effects emerged in Miami. Four acculturation gap variables – differential American practices,
differential ethnic identity, and both adolescent and parent reports of negative context of reception –
each indirectly predicted drunkenness, number of sexual partners, and number of oral sex partners (see
Table 6 and Figure 2). In addition, cigarette use was indirectly predicted by differential American
practices and by adolescent‐reported negative context of reception; binge drinking days were predicted
by differential ethnic identity and by parent‐reported negative context of reception; and number of
alcohol use occasions was predicted by parent‐reported negative context of reception (in Los Angeles).
This study examined the differential acculturation hypothesis (Szapocznik and Kurtines, 1980, 1993)
– that parent‐adolescent acculturation gaps negatively influence family functioning (operationalized as
parent‐adolescent communication), which in turn predisposes Hispanic adolescents toward problematic
behavior. Perceived discrimination and negative context of reception–two stressors confronting
Differential Acculturation and Adolescent Health Outcomes 15
immigrant and minority individuals (e.g., Lee, 2005; Portes and Rumbaut, 2006) were also examined as
predictors of adolescent outcomes through parent‐adolescent communication.
This study represents a clear advance over past research. We included gaps in Hispanic and
American practices, values, and identifications. Differential American practices, which was explicitly
referenced by Szapocznik and Kurtines (1980), indirectly predicted cigarette‐ and alcohol‐related
outcomes through adolescent reports of communication with parents, but indirectly predicted sexual
risk outcomes through parent reports of communication with adolescents. This pattern suggests that, at
least in recently immigrated Hispanic adolescents in Miami, the effects of differential rates of
Americanization on adolescent sexual risks may be more related to parents’ communication styles than
to adolescents’ own perceptions of their relationships with their parents.
Gaps in ethnic identity predicted adolescent sexual outcomes through parent‐reported
communication in Miami, but this effect was in an unexpected direction. Differential ethnic identity
positively predicted parent‐reported communication, which in turn positively predicted adolescent
sexual risk behaviors. Despite their apparent advantages, Miami adolescents were more likely (32%
versus 13%) to have initiated sex by Time 2. There are many reasons that could account for these
differences, including different peer and parent norms across sites, different nationalities, and potential
differences in religiosity. Mexican families may be especially family‐oriented and high on parental
control (Halgunseth et al., 2006). Because we did not collect data on these variables (with the exception
of nationality), it is not possible for us to support one interpretation over another. Perhaps the more
positive context of reception in Miami permitted adolescents in our Miami sample to more quickly
affiliate with U.S.‐born peers. Specific correlates of sexual activity (e.g., early entry into romantic
relationships) were not measured in the current study but may have accounted for the site differences
that we found.