University of Southern California Dissertations and Theses

The role of acculturation in the sun-safe behaviors of US Latino adults in the United States

(USC Thesis Other)

The role of acculturation in the sun-safe behaviors of US Latino adults in the United States

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

THE ROLE OF ACCULTURATION IN THE SUN-SAFE BEHAVIORS OF LATINO
ADULTS IN THE UNITED STATES

by

Valentina A. Andreeva

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
PREVENTIVE MEDICINE

December 2008

Copyright 2008 Valentina A. Andreeva

ii
Acknowledgements

I am very thankful for the support, patience, and guidance from each member of my
dissertation committee, my family and friends throughout the entire process.

iii
Table of Contents
Acknowledgements ii
List of Tables v
List of Figures vi
Abbreviations vii
Abstract viii
Chapter 1: Introduction 1
Overview 1
Ultraviolet radiation and skin cancer 3
Sun-safe behaviors 7
Acculturation and health practices 12
Limitations of sun safety and acculturation research 19
Theoretical foundation of the present research 20
Research hypotheses 26

Chapter 2: Methods 29
Data source 29
Measures 31
Statistical analysis 33

Chapter 3: Study 1 37
Abstract 38
Introduction 39
Methods 42
Results 44
Discussion 52

Chapter 4: Study 2 58
Abstract 59
Introduction 60
Methods 63
Results 67
Discussion 73

Chapter 5: Study 3 80
Abstract 81
Introduction 82
Methods 87
Results 91
Discussion 94

iv
Chapter 6: Discussion 101
Summary of findings with respect to each hypothesis 101
Limitations and strengths of the present research 109
Overall conclusion and implications for public health research and education 115

References 119

v
List of Tables

Table 1. Demographic and acculturation characteristics of Latino 45
respondents in HINTS 2005

Table 2. Prevalence of sun-safe behaviors among Latino respondents 48
in HINTS 2005

Table 3. Bivariate linear regression models of acculturation and sun-safe 50
behaviors among US Latinos

Table 4. Multivariable linear regression models of sun-safe behaviors among 51
US Latinos

Table 5. Prevalence of the hypothesized mediating variables among Latino 68
respondents in HINTS 2005

Table 6. Univariate and bivariate analyses of the hypothesized mediating and 91
outcome variables

Table 7. Analyses of moderated mediation between acculturation and sunscreen 93
use among US Latinos

Table 8. Analyses of moderated mediation between acculturation and wearing 94
protective clothing among US Latinos

vi
List of Figures

Figure 1. General theoretical model for the present research 23

Figure 2. Mediated effects of acculturation on sunscreen use: structural 71
equation analysis

Figure 3. Mediated effects of acculturation on seeking shade: structural 72
equation analysis

Figure 4. Mediated effects of acculturation on wearing protective clothing: 73
structural equation analysis

Figure 5. Hypothetical model of moderated mediation of sun-safe behaviors 87
among US Latinos

vii
Abbreviations

BRFSS Behavioral Risk Factor Surveillance System
CFI Comparative fit index
EFA Exploratory factor analyses
HINTS Health Information National Trends Survey
NHIS National Health Interview Survey
NMSC Non-melanoma skin cancer
RMSEA Root mean-square error of approximation
SCT Social Cognitive Theory
SEM Structural equation model(ing)
SES Socioeconomic status
SPF Sun protective factor
US United States
UVA Ultraviolet A radiation
UVB Ultraviolet B radiation
UVR Ultraviolet radiation
WHO World Health Organization

viii
Abstract

Acculturation has emerged as an independent risk factor for different health practices
across immigrant groups. Prompted by evidence of increasing melanoma incidence among
Latinos, the present research investigated the impact of acculturation on Latinos’ sun safety. Data
from 496 Latino respondents to the 2005 Health Information National Trends Survey were
analyzed with multiple linear regression and structural equation models. Sunscreen use, seeking
shade and wearing protective clothing were the primary outcomes, assessed by frequency scales.
Acculturation was the primary predictor, assessed by a composite index. The direct, mediated,
and moderated effects of acculturation on sun-safe behaviors were investigated in three studies.
Hypothesized mediators included health status, education, and social networks; hypothesized
moderators included gender, age, region, and health insurance status.
In adjusted models in Study 1, increased acculturation was associated with decreased use
of shade or protective clothing when outdoors (all p<0.05), across gender and region of residence.
In the mediation models in Study 2, education level emerged as a potential mediator for sunscreen
use and wearing protective clothing (both p<0.05). Perceived physical health status and social
networks involvement showed mediating effects only for sunscreen use (p<0.05). The results of
the moderated mediation analyses in Study 3 revealed that the effect of education as a mediator
between acculturation and sunscreen use might be moderated by gender, with the association
between education and sunscreen use being stronger among women than men (p<0.08). Also,
there was evidence that the association between acculturation and social networks was stronger
among insured than uninsured Latinos (p<0.08).
Despite data limitations, the findings demonstrated that acculturation has adverse effects
on skin cancer risk-related practices and suggested mechanisms through which acculturation

ix
might impact different sun-safe behaviors. Regarding interventions, initiatives for low-
acculturated Latinos could reinforce existing sun-safe behaviors via informal, Spanish-language
strategies; initiatives for high-acculturated Latinos might require more resources due to the need
for behavior modification. The mediation and moderated mediation analyses have descriptive
importance and could inform intervention development as well as studies on causality. The
results of this research should be replicated with large Latino samples and longitudinal datasets,
encompassing a variety of psychosocial measures.

1
Chapter 1: Introduction

Overview
Latinos represent the most rapidly growing segment of the United States (US) population
(U.S. Census Bureau, 2006). Over 45 million Latinos live in the US (15% of the population),
mainly in areas with high ultraviolet radiation (UVR), such as California and Texas (U.S. Census
Bureau, 2008). UVR exposure, however, might predispose individuals towards melanoma and is
clearly implicated in non-melanoma skin cancer (NMSC) development (Gandini et al., 2005;
Rubin et al., 2005; American Cancer Society, 2008). While skin pigmentation is protective
(Gloster Jr. et al., 2006; Byrd-Miles et al., 2007), findings have linked UVR exposure with DNA
damage and skin cancer across skin types (Tadokoro et al., 2003; Hu et al., 2004; Gloster Jr. et
al., 2006). Thus, skin cancer, particularly melanoma, is considered a public health concern for all
ethnic groups in the US (Cormier et al., 2006).
Validated data on melanoma incidence among US Latinos is available from the extended
Surveillance, Epidemiology, and End Results program, which covers 11 cancer registries
nationwide. For the period 1992-1996, the annual age-adjusted melanoma incidence among
Latinos was 3.7 per 100,000, and during 2001-2005 it reached 4.9 (32.4% change) (Eide et al.,
2005; Surveillance Epidemiology and End Results Program, 2008). Melanoma incidence
increased among male Latinos in California during 1988-2001 at 1.8% annually (Cockburn et al.,
2006). Latinos generally experience more advanced melanoma tumors with a poorer prognosis
than non-Latino whites (Cress et al., 1997; Cormier et al., 2006; Cockburn et al., 2006; Hu et al.,
2006), possibly due to lower awareness about risk factors or symptoms (Olsen et al., 1993;
Pipitone et al., 2002; Byrd-Miles et al., 2007), lack of health insurance (Roetzheim et al., 1999;

2
De Navas-Walt et al., 2007), insufficient efforts for melanoma screening (Cockburn et al., 2006),
and delay in seeking treatment (Olsen et al., 1993; Friedman et al., 1994; Cormier et al., 2006).
A practical estimate of UVR exposure, dose-adjusted for skin type, is sunburn (Cockburn
et al., 2006). Analyses of the 1999, 2003, and 2004 Behavioral Risk Factor Surveillance System
(BRFSS) revealed that Latinos were indeed at greater risk than non-Latino whites of experiencing
sunburn (Saraiya et al., 2007). Specifically, for all three years combined, sunburn prevalence
(e.g., reporting ≥1 sunburn episode during the preceding year) was 35.6% among non-Latino
white adults versus 45.6% among Latino adults. Moreover, both groups were almost equally
likely to report ≥4 sunburn episodes during the preceding year. Such findings among Latinos
might be partially explained by high UVR exposure due to low perceived skin cancer risk
(Pipitone et al., 2002; Ma et al., 2007).
Sun safety constitutes actively taking precautions against UVR overexposure and is a key
behavioral component of primary prevention of skin cancer regardless of skin type (U.S.
Department of Health and Human Services, 2000; American Cancer Society, 2008). Sun-safe
behaviors/strategies include applying sunscreen, avoiding midday sun exposure, seeking shade,
and wearing protective clothing and sunglasses. Due to the cumulative effect of incidental sun
exposure, scientists advocate practicing sun-safe behaviors daily (Nole et al., 2004). Several
terms regarding protection from excessive UVR have been employed in the literature, including
sun protection, UVR protection, sun safety, sun-safe behaviors, and sun-safe practices. In the
present research, the term sun-safe behaviors is used.

3
Ultraviolet radiation and skin cancer
Ultraviolet radiation (UVR)
Ultraviolet light is the source of UVR and is the most important spectrum of sunlight
with respect to skin cancer (Saladi et al., 2005). UVR is more intense in the summer, at higher
altitudes, lower latitudes, and during midday (between 10 a.m. and 3 p.m.) (Mack et al., 1991;
Urbach, 1997; Stiller, 2003; Nole et al., 2004; Eide et al., 2005; Fears et al., 2006; Sinclair, 2007).
In general, small amounts of ultraviolet light are known to be beneficial to human health due to
the synthesis of vitamin D in the skin (Diffey, 1998). The World Health Organization (WHO)
has advised that for the majority of people, minimal levels of exposure to ultraviolet light,
combined with normal dietary vitamin D intake, are sufficient to maintain vitamin D at adequate
levels that prevent health problems (World Health Organization, 2006).
The three primary components of ultraviolet light are: ultraviolet A (UVA, 320-400nm,
shortest wavelength range), ultraviolet B (UVB, 290-320nm), and ultraviolet C (UVC, 200-
290nm, longest wavelength range) (Nole et al., 2004; Saladi et al., 2005). Most of the health
compromising effects of UVR, such as tanning, burning, and skin cancer, have been attributed to
UVB (Urbach, 1997; Stiller, 2003; Nole et al., 2004), however, UVA is known to contribute to
the carcinogenic effect of UVB (Kelfkens et al., 1991). The actual mechanisms of skin
carcinogenesis pertain to genetic mutations caused by UVB and to flawed repair of the induced
damage (Halpern et al., 1999). In 1992, a Working Group convened by the International Agency
for Research on Cancer (a WHO affiliate) performed an extensive review of the scientific
evidence and concluded that there is sufficient evidence to consider solar UVR carcinogenic to
humans (International Agency for Research on Cancer, 1992).
Generally, one’s sensitivity to UVR depends on the amount of melanin (e.g., brown
pigment) in the skin, which acts as a density filter absorbing ultraviolet light and preventing its

4
penetration deeper into the skin (Kaidbey et al., 1979; Stiller, 2003). A system of classifying skin
types based on sun-reactivity has been developed, with skin type I representing white skin
without constitutive melanin pigmentation, and skin type VI representing black skin (Fitzpatrick,
1988). However, a Brazilian study utilizing a multiracial sample documented that the classic skin
types do not fully predict sun sensitivity (Dornelles et al., 2004). Further, like melanin in the
skin, ozone (an oxygen-type gas) in the stratosphere can absorb UVB (Godlee, 1991). Due to the
diminishing ozone layer, however, surface UVB levels have increased substantially at all latitudes
except the tropics (Madronich et al., 1994). It has been suggested that a 10% increase in
ultraviolet light is associated with 13.7% increase in melanoma incidence (Aase et al., 1994).

Skin cancer
The WHO has identified nine adverse health outcomes resulting from excessive sun
exposure: cutaneous malignant melanoma, squamous cell carcinoma of the skin and cornea, basal
cell carcinoma, photoaging, erythema (sunburn), cortical cataract, pterygium, and herpes of the
lips reactivation (World Health Organization, 2006). The first three of these outcomes have very
long latent periods (Urbach, 1997) and represent the most common types of cancer in the US
(American Cancer Society, 2008). However, there is indication that their prevention is of low
priority among US physicians (Christos et al., 2004). Meanwhile, NMSC is among the five most
costly cancers to Medicare (Housman et al., 2003). Squamous cell carcinoma is a type of NMSC
that accounts for approximately 20% of all NMSC cases, however, accurate incidence is difficult
to calculate because of limited registration (Marks, 1996; Jerant et al., 2000). Approximately 50-
60% of these tumors occur on the head or neck and the case-fatality rate has been estimated to be
around 1% (Marks, 1996; Jerant et al., 2000). The prevalence of this type of cancer has been
strongly linked with chronic long-term UVR (particularly UVB) exposure (Strickland et al., 1989;

5
MacKie, 2006). Basal cell carcinoma encompasses six types of tumors and is the most common
and least aggressive type of skin cancer (Jerant et al., 2000). It constitutes approximately 80% of
all NMSC cases, however, its absolute incidence is also difficult to determine because the disease
is excluded from cancer registries (Rubin et al., 2005).
Cutaneous malignant melanoma (or melanoma) represents a class of malignant and often
lethal tumors with different histological and clinical features. About 30% of melanomas develop
in preexisting benign nevi (moles) (Marks et al., 1990; MacKie, 2006). Melanoma originates in
the cells producing melanin (melanocytes), frequently entails aggressive local growth, and is
likely to metastasize (Jerant et al., 2000; Atillasoy, 2003; American Cancer Society, 2008).
Melanoma among Latinos differs not only in incidence, but also in site distribution, histological
type, and stage at diagnosis from that among non-Latino whites (Black et al., 1987; Cress et al.,
1997), which suggests the possibility of differences in behavioral patterns with respect to sun
exposure and sun safety among Latinos and non-Latino whites. There is evidence that skin
tumors among Latinos were more frequently located on the extremities than among non-Latino
whites (Vázquez-Botet et al., 1990 ). Further, a randomized retrospective study found that
pigment could be identified twice as frequently in basal cell carcinoma in patients with Latino
surnames than in basal cell carcinoma in patients with northern European surnames (Bigler et al.,
1996).
An analysis of the California Cancer Registry data from the period 1988-1993 revealed
that metastasized melanoma was more commonly diagnosed among Latino than non-Latino white
adults (15% versus 6% among males; 7% versus 4% among females) (Cress et al., 1997). A more
extensive review of the same registry, covering the period 1988-2001, corroborated the prior
findings and highlighted the alarming increase in presentation with thick melanomas at diagnosis
among Latinos (annual increase of 12% and 9% among males and females, respectively)

6
(Cockburn et al., 2006). Findings from Latin America are fairly consistent with those observed
among US Latinos in terms of histological type and delayed diagnosis (Gonzalez Cervantes et al.,
1990; Cormier et al., 2006). However, the etiology of skin cancer and especially melanoma in
populations other than non-Latino whites is not well understood and merits further investigation
(Gloster Jr. et al., 2006).
Overall, skin cancer incidence increases with age, possibly due to a decreased capacity to
repair cell damage (Saladi et al., 2005). However, age is difficult to separate from sun exposure
since both are time-dependent (Marks, 1996; Saladi et al., 2005). One Italian study estimated the
risk for melanoma attributable to sun exposure to lie between 17 and 38% (Zanetti et al., 1999).
Research suggests that intermittent, intense sun exposure among susceptible individuals is a risk
factor for melanoma; chronic (such as occupational) sun exposure is a protective factor for
melanoma, but a risk factor for squamous cell carcinoma, and has an equivocal effect on basal
cell carcinoma (Marks, 1996; Gandini et al., 2005; Berwick et al., 2006). According to the
American Cancer Society, risk factors for all types of skin cancer include sun sensitivity (i.e.,
sunburning easily, natural blond/red hair), a history of excessive sun exposure (including
sunburn), use of artificial tanning, presence of an immunosuppressive disease, personal skin
cancer history, and certain occupational chemical exposures (American Cancer Society, 2008).
Across different populations including Latinos, males have been identified as being at
higher skin cancer risk than females, possibly due to their lower engagement in sun-safe
behaviors and higher rates of sunburn (Robinson et al., 1997; Hoegh et al., 1999; Scerri et al.,
2002; Hall et al., 2003; Stanton et al., 2004; Saraiya et al., 2007). For example, among
participants in the 2003 BRFSS who had identified themselves as Hispanic white, the prevalence
of ≥1 sunburn during the preceding year was 24.8% among males and 19.7% among females
(Saraiya et al., 2007).

7
Sun-safe behaviors
Assessment of sun-safe behaviors in different populations
The American Academy of Dermatology recognizes that sun-safe behaviors at any age
are important in order to prevent skin damage and reduce the risk of cancer. It recommends
wearing wide-brim hats, sunglasses, and protective clothing (such as long-sleeve shirts and long
pants), applying broad-spectrum sunscreen with sun protective factor (SPF) ≥15 even on cloudy
days, using sunscreen in conjunction with other methods, and avoiding deliberate sun exposure
whenever possible (American Academy of Dermatology, 2008). Whereas some studies have
assessed the sun-safe behaviors together as a scale (Weinstock et al., 2002; Cafri et al., 2006;
Manne et al., 2006) or summary score (Glanz et al., 1999; Coups et al., 2005; Azzarello et al.,
2006), it appears that the behaviors are most frequently studied as individual outcomes (Hall et
al., 1997; Hoegh et al., 1999; Shoveller et al., 2000; Santmyire et al., 2001; Bergenmar et al.,
2001; Boggild et al., 2003; Salas et al., 2005; Pichon et al., 2005).
Some of the sun-safe behaviors (e.g., sunscreen application) have received considerably
more attention in the research literature than others (e.g., avoiding midday sun exposure).
Sunscreen has the capacity to scatter/reflect or absorb ultraviolet light, and many of these
products work through both mechanisms (Naylor et al., 2005). The SPF of each sunscreen
product is defined as the ratio of ultraviolet light energy required to produce one minimal
erythema dose on protected skin to the ultraviolet light energy required to produce one minimal
erythema dose on unprotected skin, and is thus inversely related to the amount of ultraviolet light
reaching the skin (Nole et al., 2004). There is some evidence that use of sunscreen with high SPF
prolongs recreational sun exposure and is positively associated with sunburn (Garbe et al., 2000;
Hall et al., 2003; Thieden et al., 2005; Autier et al., 2007). Research suggests that even sunscreen
products with low SPF (4 to10) offer significant benefits (Nole et al., 2004).

8
Use of protective clothing (especially long-sleeve shirts and wide-brim hats) are
frequently measured, yet infrequently reported by research participants (Hall et al., 1997; Glanz et
al., 1999; Hoegh et al., 1999; Shoveller et al., 2000; Santmyire et al., 2001; Bergenmar et al.,
2001; Boggild et al., 2003; Pichon et al., 2005; Cafri et al., 2006; Manne et al., 2006; Azzarello et
al., 2006). Seeking shade on sunny days has also been measured and the reported frequency of
this behavior is again low (Hall et al., 1997; Glanz et al., 1999; Shoveller et al., 2000; Santmyire
et al., 2001; Bergenmar et al., 2001; Boggild et al., 2003; Cafri et al., 2006; Manne et al., 2006;
Azzarello et al., 2006). Similar to the prevalence of wearing protective clothing and use of
sunscreen, in 2003 about a third of the general US adult population reported being very likely to
seek shade when exposed to sunlight (National Cancer Institute, 2005). Unlike use of sunscreen
and protective clothing, however, there was no indication of an increasing trend in use of shade.
Generally, sun safety barriers include protective clothing being too hot, forgetting about
protection, inconvenience, desire for a suntan, perceived advantages of sunbathing, sunscreen
lotion being too sticky/greasy, poor labeling of sunscreen products, and cost (especially in regards
to sunscreen use as often as recommended) (Boggild et al., 2003; Manne et al., 2006; Nicol et al.,
2007). The last one could be a particularly salient barrier among Latinos whose median income
is estimated to be about 30% lower than that of non-Latino whites (De Navas-Walt et al., 2006).

Theoretical foundations of sun safety research
A number of theoretical constructs and models have been tested in the sun safety
literature, most of which concerns the experience of non-Latino whites. Frequently utilized
frameworks include Becker’s (1974) and Rosenstock’s (1974) Health Belief Model (Douglass et
al., 1997; Stone et al., 1999; Greene et al., 2003; Bränström et al., 2004), Bandura’s (1986) Social
Cognitive Theory (SCT) (Stone et al., 1999; Glanz et al., 1999; Robinson et al., 2004), Ajzen and

9
Fishbein’s (1980) Theory of Reasoned Action (Bränström et al., 2004; Manne et al., 2004) and
Theory of Planned Behavior (Hillhouse et al., 2000; Myers et al., 2006). Nonetheless, some
authors have criticized sun safety research for being atheoretical (Glanz et al., 1999; Danoff-Burg
et al., 2006) and have argued that existing models do not address the impact of short-term sensual
and sociocultural influences on sun-safe behaviors (Baum et al., 1998). Indeed, the frequently
utilized cost-benefit Health Belief Model does not adequately capture socio-cultural phenomena;
constructs such as perceived susceptibility to and severity of skin cancer have been shown to
exert only minimal influence on UVR-related behaviors (Beasley et al., 1997; Coups et al., 2005;
Azzarello et al., 2006; Bränström et al., 2006).
Researchers have also noted that most of these theories, particularly the Health Belief
Model, the Theory of Reasoned Action, and the Theory of Planned Behavior, have limited
inferential potential with respect to Latinos and generally are not well-suited for this population
because of their lack of cultural relevance (Flores et al., 1995). As a dynamic and culturally-
applicable theoretical framework, however, Bandura’s SCT appears appropriate for the proposed
sun safety research. It views health practices (e.g., sun-safe behaviors) as being reciprocally
influenced by personal factors (e.g., acculturation) and the social and cultural environment (e.g.,
social networks). The theory’s utility for studies on sun-safe behaviors is enhanced by the
entailed importance of observing and modeling the behavior of similar others. SCT has already
been successfully employed in sun safety research with Caucasian populations (Stone et al., 1999;
Glanz et al., 1999; Robinson et al., 2004).

Major predictors of sun-safe behaviors
Whereas most of the findings on the predictors of sun-safe behaviors pertain to non-
Latino whites, some also apply to Latinos. Across age and racial/ethnic groups, as well as across

10
time, gender invariably emerges as one of the strongest predictors (especially for sunscreen and
shade use), with women reporting higher rates of use than men (Keesling et al., 1987; Hall et al.,
1997; Garbe et al., 2000; Santmyire et al., 2001; Scerri et al., 2002; Stanton et al., 2004;
Bränström et al., 2004; Pichon et al., 2005; Lewis et al., 2006). Women also appear more
knowledgeable about skin cancer and sun safety than men (Vail-Smith et al., 1993; Hillhouse et
al., 1996; Stanton et al., 2004). Another consistent determinant of sun-safe behaviors across
different populations is age, with older individuals reporting higher rates of protection and lower
rates of sunburn than younger ones (Hall et al., 1997; Glanz et al., 1999; Baum et al., 1998;
Hoegh et al., 1999; Saraiya et al., 2002; Sciamanna et al., 2002; Hall et al., 2003; Stanton et al.,
2004; Manne et al., 2006). Such findings might be due to changes in leisure-time behaviors,
increasing importance of cognitive factors, decreasing sensual and social influences, and/or
decreasing salience of self-presentation motives with age (Leary et al., 1993; Baum et al., 1998).
In developed countries and particularly among non-Latino whites, having a suntan is
generally perceived as attractive and appearance enhancing (Leary et al., 1993; Beasley et al.,
1997; Robinson et al., 1997; Knight et al., 2002; Hymowitz et al., 2006). Across time, the most
common motives for tanning have been those related to self-presentation, such as perceived
improvement in physical attractiveness and appearance, including the appearance of health
(Keesling et al., 1987; Vail-Smith et al., 1993; Leary et al., 1993; Beasley et al., 1997; Hillhouse
et al., 2000; Bergenmar et al., 2001; Knight et al., 2002). Interestingly, findings reveal that white
Hispanic students were more likely to tan deeply and were 2.5 times more likely to report using a
tanning bed during the preceding year than were white non-Hispanics (Ma et al., 2007).
Nonetheless, research on the psychosocial and sociocultural determinants of sun-safe behaviors
among Hispanic/Latinos is currently scare and merits increased attention.

11
The prevalence of sun-safe behaviors among Latinos
Research from Puerto Rico documented that half of the residents reported rarely or never
using sunscreen while sunbathing (Ross et al., 1990). One Chilean study on the prevalence of
sun-safe behaviors (N>1,100) showed that slightly over half of the participants used sun
protection of some type (although the period of assessment was not specified) (Molgo et al.,
2005). Only one study on sun-safe behaviors from Mexico was identified (Castanedo-Cazares et
al., 2006). It sampled 964 individuals ≥15 years of age and found that 70% of the sample took
some sun safety measures when exposed to sunlight, 41% reported wearing hats, 11.8% reported
using sunscreen, and 26.5% believed that a suntan was healthy. In comparison, age-adjusted data
from the 2003 National Health Interview Survey (NHIS) indicate that 33.2% of the general US
adult population reported being very likely to use sunscreen when exposed to sunlight (National
Cancer Institute, 2005). US studies have also documented that the perception of a tan as healthy
is higher than that observed in Mexico (Robinson et al., 1997).
Research with over 8,000 US Latino men found that sun-safe behaviors varied by country
of origin and US region of residence (Ramirez et al., 2004). For example, among men <40 years
of age, the percentage of those who did not practice any sun-safe behaviors (i.e., did not report
using sunscreen, protective clothing, seeking shade, or avoiding midday sun exposure) ranged
from 22% among Puerto Ricans in New York to 40% among Mexican Americans in San
Francisco. Recent data showed that US Latinos (aged 18-39) were about half as likely as non-
Latino whites to report multiple skin cancer risk behaviors. However, no racial/ethnic group
appeared to meet existing sun safety recommendations (Coups et al., 2008). A recent pilot study
in Florida revealed no differences in time spent in the sun on weekends and weekdays, or in
seeking shade by Hispanic ethnicity. However, white Hispanic students were more likely to tan
deeply, and less likely to use protective clothing or sunscreen than white non-Hispanics, after

12
adjustment for relevant factors (Ma et al., 2007). Further, the authors observed that among white
Hispanics and white non-Hispanics with similar skin type and family history, the Hispanic
students were less knowledgeable about skin cancer. Some of the results of this study are not
consistent with findings among adults and might indicate an alarming trend of decreasing sun-
safe behaviors in younger populations of Latinos.
Overall, the limited data reveal that sun-safe behaviors among indigenous populations in
Latin America were not optimal, yet might be better with respect to skin cancer prevention than
those observed in the US (Robinson et al., 1997; Molgo et al., 2005; Pichon et al., 2005;
Castanedo-Cazares et al., 2006). Therefore, a plausible explanation for the recent increase in
melanoma rates among Latinos pertains to worsening of these behaviors with increasing
acculturation (i.e., adoption of non-Latino pro-tanning attitudes and practices regarding sun/UVR
exposure). Acculturation has already been established as a independent risk factor for a number
of health practices, including increased substance use and cancer incidence, and worsening of
dietary practices (Kaplan et al., 1990; Vega et al., 1994; Lara et al., 2005; Abraido-Lanza et al.,
2006; Barcenas et al., 2007).

Acculturation and health practices
Acculturation: definition and measurement
Acculturation refers to the process of interchange by which immigrants and their
offspring accommodate the language, behaviors, and norms predominant in the host country
(Unger et al., 2000). It is a complex, selective, and dynamic process, influenced by personal,
social, environmental, and political factors, and reflecting changes that occur as a function of
time, degree of exposure, and context (Berry, 1980; Mendoza, 1984; LaFromboise et al., 1993;

13
Monzo et al., 2006). Overall, researchers note that acculturation is not an inherently stressful
process, but it might be stressful to be unacculturated (Neff et al., 1993).
The literature is replete with models and explanations of the processes by which people
from one culture develop competence in another culture (LaFromboise et al., 1993). Two of the
most popular acculturation paradigms include the unidirectional and the bidirectional/
bidimensional models. The former model represents the traditional view of acculturation and
describes a simple, ongoing assimilation process where increased exposure and involvement with
the host culture (i.e., the desirable outcome of assimilation) corresponds to decreased
involvement and association with the heritage culture (Gordon, 1964). In essence, it views the
acculturation process as a gradual and complete adoption of and assimilation into the host,
mainstream culture (i.e., “melting pot”). The bidirectional model, which is relatively more
recent, presupposes the existence of two independent cultural orientations (e.g., heritage and host)
and distinguishes between degrees and types of acculturation (Berry, 1980; Mendoza et al., 1981;
Flannery et al., 2001). Research has demonstrated that both the unidirectional and bidirectional
models acknowledge many factors involved in the acculturation process, such as linguistic,
economic, and social; both models have good reliability, perform well as predictors, and correlate
well with external validity criteria (Flannery et al., 2001).
Because of its elusive nature, it is not surprising that acculturation is usually assessed by
proxy variables (Mendoza, 1984; Sue, 2003). The literature contains numerous measures and
scales designed to investigate acculturation to the mainstream, white American culture, most of
which have been developed specifically for Latinos (Zane et al., 2003). Virtually all measures
rely on self-reports, assess behavior and/or attitudes, and include language (preference, use,
skills), social networks, habits, cultural knowledge, values, and identification (Mendoza et al.,
1981; Unger et al., 2002; Zane et al., 2003). Reviews of the literature reveal that language

14
proficiency and use are the most frequently employed, robust (Arcia et al., 2001; Zea et al.,
2003), and practical (Lara et al., 2005) indicators of acculturation, and are considered useful in
laying the groundwork for new areas of research (Abraido-Lanza et al., 2006). Psychometric
analyses have demonstrated that language items explain most of the variance in acculturation
scales (Marin, 1992; Epstein et al., 1996).
It is important to distinguish language issues per se from acculturation issues. While the
level of acculturation is a function of English language proficiency (Clement, 1986), it is
conceivable for an individual to become more “proficient” as an English learner, but not more
“acculturated.” For example, the level of education received in the native country and the level of
one’s confidence about English have been shown to predict ability, willingness, and frequency of
use of the English language (Clement, 1986; Espenshade et al., 1997; Clement et al., 2003).
However, these constructs are not regarded as predictors of acculturation.
There is some disagreement in the literature regarding the function of socioeconomic
status (SES) in acculturation research. SES factors have been considered confounders or
predictors of acculturation (Negy et al., 1992a; Borrayo et al., 2003) and acculturation has been
considered a predictor of SES (McManus et al., 1983), which is also the view in the present
research. Acculturation has been positively associated with cultural capital (e.g., acquisition of
new information, second language fluency, formal education) (Rueda et al., 2003; Franzini et al.,
2004; Berrigan et al., 2006; Mainous et al., 2006; Pichon et al., 2007). While sufficient research
evidence attests to the independent role of acculturation on health behaviors and outcomes (Wells
et al., 1989; Kaplan et al., 1990; O'Malley et al., 1999; Derose et al., 2000; Crespo et al., 2001;
Jurkowski et al., 2005; Abraido-Lanza et al., 2005), acculturation research is strengthened when it
accounts for the potential impact of SES.

15
Acculturation of US Latinos
It has been suggested that any research endeavor with Latino populations is liable to be
incomplete if it does not account for the impact of acculturation (Flores et al., 1995).
Acculturation is known to have an important role in defining the values and behavior of Latinos
(Marin, 1992). Studies have found that acculturation is negatively related to traditionalism and
sex-role differentiation, thus more acculturated Latinos become less likely to endorse such norms
and values than their less acculturated counterparts (Kranau et al., 1982; Negy et al., 1992b;
Cuellar et al., 1995). Highly acculturated Latinos have also been shown to exhibit less fatalism
and less endorsement of folk beliefs than those with lower acculturation (Neff et al., 1993;
Cuellar et al., 1995). Such findings have direct implications for health care utilization in general,
and for melanoma prevention in particular, since less fatalism translates into more proactive
health behaviors and possibly earlier diagnoses. Further, studies note that as Latinos acculturate,
they become increasingly involved with a wider social network (Griffith et al., 1985;
Rueschenberg et al., 1989; Franzini et al., 2004). The composition of one’s social network has
also been suggested as a mediator between acculturation and certain risk behaviors (Unger et al.,
2000).

Acculturation and health behaviors and outcomes among Latinos
The most frequently studied health behaviors and outcomes in acculturating populations
are participation in cancer screening, dietary practices, body mass index, obesity, and substance
use and abuse. Findings with Latinos are fairly consistent and indicate that acculturation is
positively associated with receipt of cancer screening (O'Malley et al., 1999; Jacobs et al., 2005;
Jurkowski et al., 2005; Lopez et al., 2006), such as clinical breast examinations (Lopez et al.,
2006), mammography (Palmer et al., 2005), and cervical screening (i.e., Pap smear) (Wu et al.,

16
2001; Moreland et al., 2006; Shah et al., 2006). In turn, higher levels of acculturation appear to
be linked to less fruit and vegetable consumption (Neuhouser et al., 2004; Gregory-Mercado et
al., 2006) and decreased consumption of basic nutrients (Guendelman et al., 1995). A greater
likelihood of higher body mass index has also been associated with increasing acculturation, after
controlling for age and SES (Himmelgreen et al., 2004; Abraido-Lanza et al., 2005). Further,
controlling for SES, age, and lifestyle factors, higher levels of language acculturation and longer
duration of US residence have been related to increased relative risk of obesity in Latinos (Khan
et al., 1997; Goel et al., 2004; Slattery et al., 2006; Barcenas et al., 2007).
The process of acculturation appears to have a negative (i.e., health compromising)
impact on substance use and abuse, although existing findings are sometimes conditional.
Several recent studies have found that higher acculturation was associated with a greater
likelihood of high alcohol intake and with being a former or current smoker (versus a non-
smoker), after controlling for age and SES (Abraido-Lanza et al., 2005; Masel et al., 2006;
Trinidad et al., 2006). These results partially corroborate previous studies that have found a
positive association between acculturation and drinking (Alaniz et al., 1999) and smoking (Maher
et al., 2005) among Latino women but not men. Some authors suggest that gender differences in
drinking decline with increasing acculturation, thus leading to the more pronounced relationship
between acculturation and drinking among women than men (Markides et al., 1990; Polednak,
1997; Zemore, 2007). Drug abuse and dependence have also been positively associated with
acculturation after controlling for age and gender (Burnam et al., 1987; Farabee et al., 1995).
It is important to note that positive effects of acculturation on health practices have also
been reported. Research fairly consistently indicates that higher levels of acculturation are
associated with increased levels of physical activity (Perez-Stable et al., 1994; Crespo et al.,
2001; Evenson et al., 2004; Slattery et al., 2006; Pichon et al., 2007). For example, a recent study

17
(N >700 Latinos) indicated that participants with more language acculturation reported more
involvement in moderate-intensity sports, and in regular and vigorous activity, less job activity,
and were more likely to meet the national physical activity recommendations than participants
with lower English proficiency levels (Slattery et al., 2006). Overall, such findings appear at
odds with those noted above regarding the positive association between acculturation and obesity
and underlie the intricate nature of the relationship between acculturation and health. In
particular, the available findings imply that acculturation-related diet changes surpass the physical
activity-level changes to produce a positive energy balance.

Acculturation and sun-safe behaviors
While studies on the link between acculturation and sun-safe behaviors are scarce, recent
research with a multiethnic sample (including 301 Latinos) in California shed some light on the
effect of acculturation on leisure-time UVR exposure (Wolch et al., 2004). Specifically, the study
documented that beach use was higher among US-born Latinos than among those born in
Mexico, after controlling for SES. The authors also indicated that based on language use, recent
immigrants were less likely to go to the beach than longer-term immigrants or the native-born.
Furthermore, findings in the business and economics literatures have revealed that the perception
of having more social/leisure hours per day increased with acculturation and approached that
among the mainstream western culture (Manrai et al., 1995).

The Hispanic paradox
A popular and debated research area with Latinos is the so-called “Hispanic paradox.”
During the last two decades, studies have been accumulating in support of the finding that
Hispanic/Latino populations have better health and lower mortality than non-Latino whites

18
despite their less favorable profiles with respect to SES (i.e., higher rates of being uninsured,
lower income, education) (Markides et al., 1986; Sorlie et al., 1993; Franzini et al., 2001; Cho et
al., 2004; Elo et al., 2004; Markides et al., 2005; Turra et al., 2007). Mortality and survival have
been the most frequently studied outcomes in this literature (Turra et al., 2007). Overall,
significant variations by age, gender, subgroup, and acculturation have been identified (Franzini
et al., 2001; Markides et al., 2005) and it now appears that the greatest mortality advantage is
present among Mexican American older men (Markides et al., 2005).
Explanations for the Hispanic paradox include data artifacts, specific characteristics of
migrants and migration, and a cultural buffering effect (Palloni et al., 2004). Two of the most
popular hypotheses pertain to the healthy migrant effect (or selection bias) and the “salmon bias.”
The former hypothesis holds that immigrants generally have good health that allows them to
undertake the transition between countries/cultures (Cho et al., 2004; Crimmins et al., 2005). In
turn, the “salmon bias” hypothesis pertains to the return of migrants to their native countries
following a diagnosis with a terminal illness, thus contributing to an under-ascertainment of
Latino mortality in US death registries (Patel et al., 2004; Palloni et al., 2004). Finally,
depending on the classification method (i.e., nativity versus surname), authors have observed
inconsistencies in death counts and an omission of 15-20% of US mortality data (Smith et al.,
2006).
Because the Hispanic paradox does not seem to apply to all Hispanics, researchers have
considered it “spurious and misleading” (Borrell, 2005). Some studies have in fact documented
findings contrary to the Hispanic paradox (Hunt et al., 2003; Crimmins et al., 2007). Regarding
the questions in the present research, the presence of a Hispanic paradox was expected to be
conditional due to several considerations. Findings among Caucasians indicate that low SES
might be a predictor of less sunscreen use (Hoegh et al., 1999). Also, analyses of the 2003

19
BRFSS indicate that increased income and education were predictors of sunburn (Brown et al.,
2006). If that is the case specifically with Latinos, then less sunscreen use/more sunburn would
translate into higher rather than lower skin cancer risk (i.e., no paradox). Information on the sun-
safe behaviors of Latino outdoor workers, however, suggests that increasing acculturation and
increasing SES might result in a deterioration of these practices (e.g., less use of protective
clothing, increasing intermittent sun exposure) which would be in line with the paradox.

Limitations of sun safety and acculturation research
Limitations of sun safety research
First, much of the knowledge about sun-safe behaviors is derived from studies with
predominantly non-Latino white adolescent and college samples. Thus, generalizability of the
findings is limited and information about the prevalence, risk factors, and predictors of sun-safe
behaviors cannot always be easily extrapolated to other populations with different socio-cultural
circumstances. Second, reviews of the literature have revealed substantial variation in study
methods (for example, different sun exposure assessment methods, differences in type of outdoor
activities that influence sun exposure, different ambient UVR levels), which hinders comparisons
among findings (Wright et al., 2005). Third, much of the research with Latinos (studied
individually or as part of multiethnic samples) has been concerned with occupational sun safety.
Hence, data on these behaviors during leisure-time in the general Latino population are scarce.
Fourth, researchers have pointed out that there is no gold standard for quantifying personal UVR
exposure (Wright et al., 2005). Because of the growing melanoma incidence and high prevalence
of sunburn among US Latinos, as well as the size of this population, it is critical to assess and
target their UVR exposure, skin cancer and sun safety knowledge and behaviors.

20
Limitations of acculturation research
Whereas some authors criticize the acculturation literature for being atheoretical
(Landrine et al., 2004), others view it as being replete with theoretical models (Abraido-Lanza et
al., 2006). The controversy appears to lie in the disparity among the acculturation models, the
relative diversity in acculturation conceptualization, methodology and measures, the reliance on
proxy measures and self-reports, and the paucity of theories and tests that account for the
underlying causal paths and specific mechanisms responsible for the effects of acculturation on
health (Negy et al., 1992b; Myers et al., 2003; Hunt et al., 2004; Abraido-Lanza et al., 2005).
Some authors argue that culture is a complex and context specific phenomenon and as such
cannot be reduced to a measurable variable (Hunt et al., 2004). Others, however, note that
cultural attitudes, values and behaviors associated with the acculturation process can be reliably
measured and have empirical construct validity (Cuellar et al., 1995). Strategies for minimizing
the conceptual and methodological deficiencies in acculturation research include providing a
clear rationale for the selected acculturation model, delineating the cultural differences (Hunt et
al., 2004), and performing mediation analyses (Unger et al., 2000). In addition, the acculturation
literature covers a somewhat limited range of health practices, which so far have not included
sun-safe behaviors.

Theoretical foundation of the present research
The literature has highlighted several common psychosocial characteristics across
national origins, which make Latinos a markedly identifiable group (Sabogal et al., 1987; Marin
et al., 1991) despite variability in cultural characteristics (Kranau et al., 1982). Such
characteristics include familialism (or familism) (Buriel, 1984; Hutchison, 1987; Marin et al.,
1991), allocentrism (or collectivism), simpatia (i.e., interest in promoting pleasant social

21
relationships), gender roles (with males being culturally expected to be in charge, to be strong,
and to provide for their families), and flexible attitudes toward time (Hutchison, 1987; Marin et
al., 1991). The cultural theme of fatalismo (e.g., the perception of life’s outcomes as being
outside of one’s control) has also been consistently reported (Perez-Stable et al., 1992; Olsen et
al., 1993; Anez et al., 2005). Partially attributed to a fatalistic outlook on illness is the delay in
seeking treatment for symptoms (Olsen et al., 1993) which translates into late-stage diagnosis
(Cockburn et al., 2006).
The overall goal of the present research, consisting of three studies, was to expand
knowledge about the role of Latinos’ acculturation in health practices related to skin cancer risk
and thus inform public health research and education initiatives for this population. Bandura’s
SCT was chosen to guide this research because it offers a dynamic, culturally-relevant paradigm,
explicitly accommodates the impact of the sociocultural environment, and because it has
successfully guided prior sun safety research (Stone et al., 1999; Glanz et al., 1999; Robinson et
al., 2004). The primary predictor - acculturation - was regarded as a SCT-based personal factor.
In turn, the primary outcomes – sun-safe behaviors - were regarded as being environmentally and
culturally influenced, also in line with SCT.
The present research also employed Myers and Rodriguez’s model (2003), which
highlights the indirect effects of acculturation on health. However, many of the mediated paths in
that model have not been empirically examined. Mediating variables are constructs that transmit
the effect of the predictor to the outcome variable and contribute to understanding of the process
or mechanism underlying that relationship (MacKinnon et al., 2007). For the present research,
the selection of the sociocultural mediators of the relationship between acculturation and sun-safe
behaviors was informed by SCT and findings in the literature, yet was restricted by data
availability.

22
The inclusion of acculturation has been considered critical for any studies with Latinos
(Flores et al., 1995). Since the present research was concerned with the impact of acculturation
on health behaviors in the sun, rather than with the actual process of acculturation, the construct
was treated as a fixed, antecedent variable (Flores et al., 1995). Also, the available data allowed
for the assessment of acculturation on a unidirectional scale without rejecting its bidirectionality.
Specifically, the main interest was on the impact of adopted mainstream practices, and not on the
degree of separation from the heritage culture. The full theoretical model for the present research
in presented in Figure 1.

Model rationale
Since the US mainstream norm is more pro-sunscreen use than the norm among Latinos
(Hall et al., 1997; Pichon et al., 2005), it was expected that sunscreen use would increase with
acculturation. Meanwhile, the US cultural norm is more pro-suntan (Robinson et al., 1997) than
the norm in Latin America (Castanedo-Cazares et al., 2006), hence it was also expected that
acculturation would be negatively associated with use of shade or protective clothing. While the
causal mechanisms of the effects of acculturation on health are not well established (Hunt et al.,
2004; Abraido-Lanza et al., 2006), acculturation has been positively associated with healthcare
access (Wells et al., 1989; Abraido-Lanza et al., 2004) which, in turn, has been associated with
sunscreen use, skin cancer examinations (Santmyire et al., 2001) and sunburn (Hall et al., 2003).
The potential mediating effect of healthcare access on health practices might be due to having a
regular healthcare provider (Moreland et al., 2006; Cairns et al., 2006) thus a regular source of
health information (Gombeski Jr et al., 1982; Wathen, 2006), a favorable attitude towards
provider communication behavior (Rutten et al., 2006), and a source of recommendation for
screening (Cairns et al., 2006).

23
Figure 1. General theoretical model for the present research
Sun-safe behaviors
Healthcare access
Social networks
Health status
Education
Acculturation
Gender
Age
SES
Region

Theoretically, healthcare access has been proposed as causal agent (e.g., mediator) in
individuals’ health practices (Myers et al., 2003) and its mediating effects were assessed in Study
2. However, it might be a more plausible or proximal mediator for types of health/risk behaviors
for which provider contact is critical (e.g., cancer screening, smoking cessation) (Abraido-Lanza
et al., 2004; Ranney et al., 2006) than for sun-safe behaviors, which are dictated by a complex set
of personal, sociocultural, and environmental factors (Leary et al., 1993; Clarke et al., 1997;
Jackson et al., 2000; Garbe et al., 2000; Bränström et al., 2004; Thieden et al., 2005). Some
researchers have suggested SES variables as potential moderators of health behaviors and
outcomes (Kraemer et al., 2006). SES markers have typically included income/poverty measures
and education (Chesney et al., 1982; Khan et al., 1997; Borrayo et al., 2003; Franzini et al.,
2004), however, access to services has also been proposed as an important SES indicator (Hunt et
al., 2004). Further, researchers have advocated the need for a better understanding of the

24
mechanisms through which SES impacts health behaviors (Borrayo et al., 2003). A review of the
acculturation literature noted that it is important to investigate how SES factors might be
associated with cultural factors (Hunt et al., 2004). Hence, it is possible that health insurance
status (and not access in general) might act as a moderator in the mediated relationship between
acculturation and sun-safe behaviors. Insured individuals might have more opportunities to join
provider-based health promotion groups/networks, and to use preventive services (Wells et al.,
1989; Abraido-Lanza et al., 2004), thus improving self-perceptions of health, compared to their
uninsured counterparts.
Research suggests that self-perceptions of health status improve with acculturation
(Markides et al., 1991; Shetterly et al., 1996; Franzini et al., 2004; Lara et al., 2005; Kandula et
al., 2007). Perceived health status, in turn, might influence the amount of time spent outdoors.
For example, sunburn frequency - an estimate of UVR overexposure, dose-adjusted for skin type
(Cockburn et al., 2006) - was noted to be higher among individuals reporting excellent/good
health than among others (Hall et al., 2003).
Individuals’ social networks have been shown to undergo changes in size and
composition with acculturation (Griffith et al., 1985; Rueschenberg et al., 1989; Unger et al.,
2000; Franzini et al., 2004). Features of the social network (e.g., networks regarding
occupation/recreational pursuits, belonging to a health club, having friends who sunbathe) have
also been correlated with individual sun-safe behaviors (Keesling et al., 1987; Robinson et al.,
2004). Having a trusted source of health information (Gombeski Jr et al., 1982) and the
opportunity to encounter health/risk behavior influences from one’s peers and other contacts
(Unger et al., 2000) have been suggested as reasons for the potential mediating effects of the
social networks construct.

25
In turn, there is evidence of a positive association between acculturation and cultural
capital (e.g., acquisition of new information, second language fluency, formal education) (Rueda
et al., 2003; Franzini et al., 2004; Berrigan et al., 2006; Mainous et al., 2006; Pichon et al., 2007),
and between education level and skin cancer prevention/early detection behaviors (e.g., sunscreen
use, seeking shade, skin cancer examination). The latter might be due to increased awareness
about risk factors and/or symptoms (Michielutte et al., 1996; Hoegh et al., 1999; Santmyire et al.,
2001; Azzarello et al., 2006). Research has also suggested that adults with a university degree
might be more likely to vacation at seaside resorts (thus have increased UVR exposure) than
those who had completed only elementary school (Bränström et al., 2004).
Regarding potential moderators, evidence suggests that women experience more
unfavorable acculturation effects than men (Bethel et al., 2005; Zemore, 2007; Barcenas et al.,
2007), which has been attributed to baseline differences in certain behaviors (e.g., smoking) and
gender-based social norms in the countries of origin and in the US (Alaniz et al., 1999; Caetano et
al., 2008). Research also indicates that vulnerability to the adverse effects of acculturation
decreases with age due to weakening of distress factors related to mainstream integration (Kaplan
et al., 1990). Thus, the potential moderating effects of gender and age on sun-safe behaviors were
investigated.
It is possible that gender might moderate the main effects of acculturation on sun-safe
behaviors, or the relationship between acculturation and social networks, since social network
dynamics have been shown to differ for men and women during the mainstream integration
process (Hagan, 1998). Further, the relationship between education and sun-safe behaviors might
be moderated by gender. This conjecture is based on two premises. First, findings among
college students (i.e., the same education level across gender) have revealed that women were
more likely to use sunscreen (Jerkegren et al., 1999) and had more positive attitudes and

26
intentions towards sun safety than men (Hillhouse et al., 1996). Second, women in general have
been shown to be more knowledgeable about skin cancer and sun-safe behaviors (Vail-Smith et
al., 1993; Hillhouse et al., 1996; Stanton et al., 2004).
Finally, due to evidence that UVR exposure and sun-safe behaviors might vary by region
of residence and by SES (Robinson et al., 1997; Hoegh et al., 1999; Santmyire et al., 2001;
Brown et al., 2006), in the present research it was expected that the observed effects would be
more pronounced in areas with higher versus lower UVR levels, and among Latinos of increased
SES.

Research hypotheses
The direct, mediated, and moderated relationships between acculturation and each of five
sun-safe behaviors (e.g., wearing hats, long-sleeve shirts, and long pants, using sunscreen, and
seeking shade) were investigated. In all three studies the primary predictor was acculturation,
assessed by an index including a cognitive component (perceived comfort with the English
language), a behavioral component (choice of interview language), and demographic components
(place of birth, age at US arrival, and duration of US residence). Study 1 assessed the main
effects of acculturation on sun-safe behaviors. Study 2 assessed the effects of education level,
perceived physical health status, healthcare access, and involvement with social networks
regarding health matters as potential mediators of the relationship between acculturation and sun-
safe behaviors. Finally, Study 3 employed a moderated mediation model and assessed how
different mediated paths varied as a function of gender and health insurance. The specific
hypotheses in this research were as follows:

27
Study 1
Hypothesis 1: High-acculturated Latinos would report more sunscreen use when
outdoors in the sun than their low-acculturated counterparts.
Hypothesis 2: Acculturation would be negatively associated with use of shade or
protective clothing when outdoors in the sun.
Hypothesis 3: The observed effects of acculturation on Latinos’ sun-safe behaviors
would be stronger among women than men, among younger than older Latinos, among
individuals of higher versus lower SES, and more pronounced in areas with higher versus lower
UVR levels.

Study 2
Hypothesis 4: Perceived health status, level of formal education, healthcare access, and
involvement with social networks regarding health matters would display statistically significant
mediation effects between acculturation and sun-safe behaviors.
Hypothesis 5: Each sun-safe behavior would be mediated by a unique set of potential
mediating constructs.

Study 3
Hypothesis 6: The association between acculturation and social networks would be
stronger among men than women, while the association between social networks and sun-safe
behaviors would remain constant across gender.

28
Hypothesis 7: The association between acculturation and education level would be
constant across gender, while the association between education level and sun-safe behaviors
would be stronger among women than men.
Hypothesis 8: The association between acculturation and perceived health status and
social networks, respectively, would be stronger among insured compared to uninsured Latinos,
while the association between each of these mediators and sun-safe behaviors would be constant
across health insurance status.

29
Chapter 2: Methods
Data source
The present research utilized cross-sectional data from the 2005 Health Information
National Trends Survey (HINTS). It was developed and implemented by the National Cancer
Institute for the purpose of monitoring national health communication and behavior trends. The
survey was grounded in a rich theoretical framework encompassing theories of health
communication, media usage, behavior change, and diffusion of innovations (Rutten et al., 2007).
It incorporates constructs from the BRFSS and NHIS, thus complementing these large, national
surveillance tools (Rutten et al., 2007). The main contributions of HINTS, however, are entailed
in the opportunity to examine behaviors relevant to health communication in the current
environment of rapidly changing communication options, and in its inclusion of interdisciplinary
measures that could be used to assess relationships across different domains (Rutten et al., 2007).
This is particularly relevant for the present research, which aimed to investigate previously
unexplored relationships (e.g., the link between acculturation and sun-safe behaviors) utilizing
health communication in social networks.
The instrument had undergone extensive pre-testing and expert review before field-
testing and subsequent implementation. Psychometric analyses and documentation of the
reliability and validity of the HINTS items are forthcoming (Rutten et al., 2007). The survey
conception, design, testing, and implementation are described in more detail in other publications
(Nelson et al., 2004; Davis et al., 2005). The questionnaire is available at:
http://hints.cancer.gov/instrument.jsp.
Data for HINTS 2005 were gathered between February and August 2005 via geographic
stratification and list-assisted random-digit dialing, using a probability-based sample of the US
civilian, non-institutionalized, adult population. The survey employed a computer-assisted

30
telephone interviewing system, which automated the processes of call scheduling, interviewing,
and data collection for quality control purposes (Rutten et al., 2007). Using skip logic and
question paths depending on age and gender, survey items progressed from questions about use of
communications media to more specific questions about health knowledge and behaviors. Thus,
no individual was asked all questions. Several non-response reduction methods were employed,
such as matching telephone numbers to mailing addresses, offering small financial incentives to
select sub-samples, and making up to 30 call attempts. One adult (aged 18+) was selected at
random within each household and was given the option of responding in English or Spanish.
5,586 individuals completed the full interview, of whom 496 answered affirmatively to the
question, “Are you Hispanic or Latino?” and only their data were used in this research. Latinos
were not oversampled in HINTS 2005. The Spanish language version of the questionnaire was
developed by a team of bilingual translators who had initially translated the instrument from
English to Spanish, and then back-translated it from Spanish to English as a quality control check
(Rutten et al., 2007).
According to the US Federal Register, the term ‘Hispanic’ is commonly used in the
eastern portion of the US, while the term ‘Latino’ is commonly used in the western portion of the
country (Federal Register, 1997). The same source also indicates that Hispanic/Latino is
considered an ethnicity and refers to persons of Mexican, Puerto Rican, Cuban, Central/South
American, or other Spanish heritage, regardless of race. Nonetheless, researchers have called the
term ‘Hispanic,’ “an example of identity imperialism” and have considered it inaccurate and
offensive (Comas-Diaz, 2001). In order to reflect the Latin American origin or descent of the
sample irrespective of other characteristics, the participants in the present research were referred
to as Latino. Unlike the term ‘Hispanic,’ Latino is a nationality-neutral and sex-neutral term
which most people of Latin American origin or descent find least objectionable (Hayes Bautista

31
et al., 1987; Comas-Diaz, 2001). However, in order to maintain accuracy when citing prior
research, the terms of choice in those studies were retained.

Measures
Sun safety: This construct was assessed by five primary prevention behaviors, each with a
five-point frequency scale (1=always; 5=never). Respondents were told that the questions
pertained to sun protection practices and were asked how often they used sunscreen, wore hats
that shade the face, neck, and ears, wore long-sleeve shirts and long pants, and stayed in the shade
when outside for >1 hour on a warm, sunny day. The items were comparable to the standardized
set of survey measures of sun exposure and sun protection habits, and are considered applicable
across different populations (Glanz et al., 2008). Each item was reverse-coded so that higher
scores reflected higher endorsement of the behaviors. Correlation and exploratory factor analyses
(EFA) were employed in order to determine whether to model the sun-safe behaviors as a sun
safety index or as separate outcome variables.
Acculturation: Five items were available for the assessment of the primary predictor:
nativity (US- or foreign-born), interview language (English or Spanish), a six-level item for
perceived comfort with the English language (1=completely comfortable; 6=do not speak
English), a continuous measure for age at US arrival (assessed among the foreign-born and
calculated by subtracting the year of birth from the year of arrival), and a continuous measure of
duration of US residence (assessed among the foreign-born and calculated by subtracting the year
of arrival from the year of data collection, 2005). As in prior acculturation research (Cantero et
al., 1999), each item was standardized (mean=0, standard deviation=1) and the Cronbach’s
reliability score was calculated (α=0.75). The acculturation index was obtained by computing the
mean of all items that loaded on a single factor in EFA.

32
Mediators: Six variables were available for the assessment of social networks. One was a
frequency item about the number of community organizations where the participant has
membership; another item assessed whether any of these organizations provided health
information (yes/no). These two items were combined into a single measure with the following
response categories: 0=no membership, 1=membership in a non-health organization, and
2=membership in a health-related organization. The next two items assessed whether the
participant had any family and/or friends with whom he/she talked about health (yes/no) and the
frequency of talking with them (1=not very frequently to 3=very frequently). These two items
were also combined into a single measure with the following response categories: 0=no
family/friends, 1=have family/friends but do not talk to them frequently; 2=have family/friends
and talk to them somewhat frequently; 3=have family/friends and talk to them very frequently.
The next social network item was a continuous measure for the number of neighbors on whom
the participant could rely for transportation for healthcare. The last item measured the frequency
(0=never, 4=every week) with which the participant attended religious services excluding
weddings and funerals. All social network variables were evaluated with EFA in order to explore
their factor clustering and to select the most reliable measures.
Healthcare access was assessed with two items: availability of health insurance (yes/no)
and one frequency item measuring use of services by a provider during the past 12 months
(excluding emergency room visits). These two variables were also evaluated with EFA for
potential modeling as a single access factor. Perceived physical health status was assessed with a
single item whose response options ranged from 1=excellent to 5=poor. This item was reverse-
coded so that higher scores corresponded to better health. Education level was assessed with a
single item with responses ranging from 1=never attended school to 11=professional/graduate
degree.

33
Moderators and covariates: Moderators in Study 1 included gender, age, SES (education
level measured on a continuous scale), and Census region of residence. The division of regions
into areas with high (South, West) and low (Midwest, Northeast) UVR relied on Saraiya and
colleagues’ (2002) methodology and evidence that latitude and mean annual UVR index are
strongly correlated (Eide et al., 2005). Nonetheless, these categories represent a crude proxy for
UVR exposure. Moderating variables in Study 3 included gender and having health insurance
(yes/no). Covariates included recoded dichotomous measures of marital status (married/living
with partner versus other), employment (employed/self-employed versus other), and cancer
history (yes/no). Finally, all “refused” and “don’t know/other” responses were coded as missing.

Statistical analysis
Preliminary data analyses in all three studies consisted of descriptive univariate and
bivariate analyses, including t-tests and 2x2 chi-square tests. Because each sun-safe behavior was
a five-level dependent variable and because of the fairly large sample size (which arguably
justifies the use of parametric tests despite some non-normality of the underlying distributions),
multiple linear regression analyses were performed in Studies 1 and 3. Also, in these two studies
the magnitude of the regression slopes (betas) was considered, statistical significance was
established by the F test, and potential moderating effects were assessed with tests for interaction.
The adjusted models included only variables with statistical significance in the preliminary
analyses in order to minimize experimentwise Type I error and to investigate only highly targeted
hypotheses (Cohen, 1990). All statistical procedures for Studies 1 and 3 were performed with
SAS, version 9.1 (SAS Institute Inc., 2004).
In Study 2, the preliminary data analyses included principal component EFA with oblique
and orthogonal rotations for examining the factor loadings of the sun safety, social network and

34
healthcare access items, respectively, and for selecting the most reliable measures. Then a
measurement model specifying the relationships between the measured and latent variables was
created. Next, a structural model establishing the relationships among the latent factors was
specified (MacKinnon et al., 2003). Since the HINTS data are cross-sectional, the explicit
assumption that the primary predictor (acculturation) preceded each mediator, which preceded the
outcome (i.e., each sun-safe behavior) was made. The complete model for Study 2 included
unidirectional paths from acculturation to each hypothesized mediator, as well as unidirectional
paths from each hypothesized mediator to each sun-safe behavior. It also included correlations
between each two mediating constructs.
The main analyses for Study 2 included structural equation models (SEM) with
standardized covariance matrices as input, the maximum likelihood function with robust
estimation, and the Lagrange multiplier tests. Model fit was evaluated with the chi-square
goodness-of-fit statistic (which could be inflated by large sample sizes), the comparative fit index
(CFI) and the root mean-square error of approximation (RMSEA), both of which are robust to
sample size biases (Hu et al., 1995; Fan et al., 1997). Conventionally, CFI values ≥0.95 and/or
RMSEA values ≤0.06 signify appropriate fit (Hu et al., 1999; Buhi et al., 2007). It has been
shown that mild to moderate data non-normality has negligible effects on SEM fit indices and
parameter estimates (Fan et al., 1997). Also, the sample size available for this research (N=496)
was unlikely to cause improper SEM solutions (Fan et al., 1997). All SEM analyses in Study 2
were performed with EQS version 6.1 for Windows (Bentler, 2004).
For the moderated mediation analyses in Study 3, the procedures outlined by Muller,
Judd and colleagues (2005) were followed. Specifically, moderated mediation effects were
assessed by creating interaction terms between each hypothesized moderator and acculturation in
three equations:

35
(1) Y= α
10
+ b
11
X + b
12
Z + b
13
X*Z + e
1

(2) M= α
20
+ b
21
X + b
22
Z + b
23
X*Z + e
2

(3) Y= α
30
+ b
31
X + b
32
Z + b
33
X*Z + b
34
M + b
35
M*Z + e
3

In these equations, Y represents the primary outcome (sun-safe behaviors), X represents the
primary predictor (acculturation), M represents a mediator, Z represents a moderator, α represents
an intercept, and e represents an error term. The first equation assesses the overall effect of
acculturation on sun-safe behaviors, as well as whether this effect is moderated by Z. The second
equation assesses whether the effect of acculturation on the mediator (M) differs across levels of
the moderator (Z). The third equation assesses the moderated effect of the mediator on sun-safe
behaviors, and the residual moderated effect of acculturation on sun-safe behaviors, after
controlling for the mediator (Muller et al., 2005).
According to Muller and colleagues, in order to establish that moderated mediation
exists, the following three criteria must be met: 1) in equation (1), acculturation should have a
significant effect on the sun-safe behavior but this overall effect should not be moderated (e.g.,
b
11
≠0; b
13
=0); 2) in equations (2) and (3) the effect of acculturation on the mediator must depend
on the moderator, and/or the partial effect of the mediator on the sun-safe behavior must depend
on the moderator (e.g., both b
23
and b
34
≠0 or both b
21
and b
35
≠0); and 3) the residual direct effect
of acculturation on the sun-safe behavior after controlling for the mediator must be moderated
(e.g., b
33
), however, statistical significance is not necessary. Analyses estimating equations (1),
(2) and (3) were conducted separately for each combination of the posited mediators and
moderators. To avoid redundancy, only equation (3) was adjusted for covariates. Moderated
mediation analyses entail the assumptions that the predictor and moderator are independent, and
the residuals are independent and normally distributed (Muller et al., 2005).

36
Overall, the primary interest in all three studies was in establishing associations among
the constructs, and not in making predictions or obtaining population estimates. Also, this
research utilized roughly 9% of the HINTS 2005 sample, thus sampling weights were not
modeled. The level of significance for all tests was set to alpha=0.05.

37
Chapter 3: Study 1

Acculturation and sun-safe behaviors among US Latinos: Findings from the 2005 Health
Information National Trends Survey

Valentina A. Andreeva,
1
Jennifer B. Unger,
1
Amy L. Yaroch,
2
Myles G. Cockburn,
1

Lourdes Baezconde-Garbanati,
1
Kim D. Reynolds
1

1
Department of Preventive Medicine, Keck School of Medicine, University of Southern
California, Los Angeles, CA
2
Health Promotion Research Branch, Division of Cancer Control and Population Sciences,
National Cancer Institute, Bethesda, MD

Contact: Valentina A. Andreeva
Institute for Prevention Research
Keck School of Medicine, University of Southern California
1000 South Fremont Ave., Unit 8
Alhambra CA 91803
Tel: 626-457-6613/ Fax: 626-457-4012
Email: andreeva@usc.edu

Status: In review at the American Journal of Public Health

38
Chapter 3 Abstract
Objectives. Prompted by evidence of increasing melanoma incidence among Latinos, the
relationship between acculturation and sun safety was examined in this population.
Methods. Data from 496 Latino respondents to the 2005 Health Information National
Trends Survey (HINTS) were analyzed with linear regression models. Sunscreen use, staying in
the shade and wearing protective clothing were the primary outcomes, assessed by frequency
scales, and acculturation was assessed by a composite index.
Results. In bivariate models, acculturation was negatively associated with use of shade
and protective clothing, and positively associated with sunscreen use (all p<0.004). In adjusted
models, acculturation was negatively associated with staying in the shade or wearing protective
clothing (all p<0.05) across gender and region of residence.
Conclusions. Despite HINTS data limitations, our results highlight the need for sun safety
education for Latinos, as well as support existing outdoor worker protection policy
recommendations. Initiatives for low-acculturated Latinos could reinforce existing sun-safe
behaviors via simple, Spanish-language, informal strategies, whereas initiatives for high-
acculturated Latinos might require more resources due to the need for behavior modification.
Future research could augment the present findings by utilizing large, longitudinal datasets.

39
Introduction
Latinos represent the most rapidly growing segment of the US population and live
primarily in areas with high UVR, such as California and Texas (U.S. Census Bureau, 2008).
UVR exposure, however, might predispose individuals towards melanoma and is clearly
implicated in non-melanoma skin cancer development (Gandini et al., 2005; Rubin et al., 2005;
American Cancer Society, 2008). While skin pigmentation is protective (Gloster Jr. et al., 2006;
Byrd-Miles et al., 2007), findings have linked UVR exposure with DNA damage and skin cancer
across skin types (Tadokoro et al., 2003; Hu et al., 2004; Gloster Jr. et al., 2006). The annual
age-adjusted melanoma incidence among US Latinos was 3.7 per 100,000 during 1992-1996, and
during 2001-2005 it reached 4.9 (32.4% change) (Eide et al., 2005; Surveillance Epidemiology
and End Results Program, 2008). Melanoma incidence increased among male Latinos in
California during 1988-2001 at 1.8% annually (Cockburn et al., 2006). Latinos experience more
advanced tumors with a poorer prognosis than non-Latino whites (Cress et al., 1997; Cockburn et
al., 2006; Hu et al., 2006), possibly due to lower awareness about risk factors or symptoms
(Pipitone et al., 2002; Byrd-Miles et al., 2007), lack of health insurance (Roetzheim et al., 1999;
De Navas-Walt et al., 2007), insufficient efforts for melanoma screening (Cockburn et al., 2006),
and delay in seeking treatment (Friedman et al., 1994; Cormier et al., 2006).
Recent research documented that Latinos were at greater risk of sunburn, which is an
estimate of UVR overexposure dose-adjusted for skin type (Cockburn et al., 2006), than non-
Latino whites (Saraiya et al., 2007). Specifically, adult sunburn prevalence (e.g., reporting ≥1
sunburn during the preceding year) was 35.6% among non-Latino whites and 45.6% among
Latinos. Moreover, both groups were almost equally likely to report ≥4 sunburn episodes during
the preceding year (Saraiya et al., 2007). Regarding Latinos, such findings might be partially

40
explained by high UVR exposure due to low perceived skin cancer risk (Pipitone et al., 2002; Ma
et al., 2007).

Sun-safe behaviors among Latinos
Sun safety constitutes avoiding UVR overexposure by means of applying sunscreen,
avoiding midday sun exposure, and wearing protective clothing, and is important for the primary
prevention of skin cancer (American Cancer Society, 2008).

Scarce findings on sun-safe
behaviors and attitudes among adults in Latin America reveal lower perceptions of suntans as
healthy than in the US (Robinson et al., 1997; Castanedo-Cazares et al., 2006) and evidence that
sun safety is not optimal, but might be better than in the US (Robinson et al., 1997; Pichon et al.,
2005; Molgo et al., 2005; Castanedo-Cazares et al., 2006). Recent data showed that US Latinos
(aged 18-39) were about half as likely as non-Latino whites to report multiple skin cancer risk
behaviors (Coups et al., 2008). However, no racial/ethnic group appeared to meet existing sun
safety recommendations.

Acculturation and health
The growing melanoma incidence among US Latinos could be partially due to worsening
of sun safety with increasing acculturation, which has been established as an independent risk
factor for a number of health practices (Kaplan et al., 1990; O'Malley et al., 1999; Crespo et al.,
2001; Barcenas et al., 2007). Acculturation describes the process of interchange by which
immigrants and their offspring accommodate the language, behaviors, and norms predominant in
the host country (Unger et al., 2000). Reviews of research with Latinos indicate that higher
acculturation is associated with increased substance use and cancer incidence, and worsening of
dietary practices (Vega et al., 1994; Lara et al., 2005; Abraido-Lanza et al., 2006). Evidence

41
suggests that women experience more unfavorable acculturation effects than men (Bethel et al.,
2005; Zemore, 2007; Barcenas et al., 2007), which has been attributed to baseline differences in
certain behaviors (e.g., smoking) and gender-based social norms (Alaniz et al., 1999; Caetano et
al., 2008). Research also indicates that vulnerability to the adverse effects of acculturation
decreases with age due to weakening of distress factors related to mainstream integration (Kaplan
et al., 1990).

The present study
We investigated the influence of acculturation on Latinos’ sun safety, which could impact
their skin cancer risk. Since the US mainstream norm is more pro-sunscreen use than the norm
among Latinos (Hall et al., 1997; Pichon et al., 2005), we expected that sunscreen use would
increase with acculturation. Meanwhile, the US cultural norm is more pro-suntan (Robinson et
al., 1997) than the norm in Latin America (Castanedo-Cazares et al., 2006), hence we expected
that acculturation would be negatively associated with use of shade or protective clothing.
Consistent with the literature (Santmyire et al., 2001; Stanton et al., 2004; Barcenas et al., 2007),
we also hypothesized that the observed effects would be stronger among women than men,
among younger than older Latinos, and more pronounced in areas with higher versus lower UVR
levels. Finally, led by findings among non-Latino whites (Hoegh et al., 1999), we hypothesized
that the role of acculturation in sun safety would be greater among those of increased SES.
Bandura’s Social Cognitive Theory (1986) which has been employed in prior sun safety research
(Glanz et al., 1999; Stone et al., 1999; Robinson et al., 2004) and which accommodates the
influence of the sociocultural environment on health behaviors, guided this study.

42
Methods
Data source
We used cross-sectional data from the 2005 Health Information National Trends Survey
(HINTS). Data were gathered between February and August 2005 via geographic stratification
and list-assisted random-digit dialing, using a probability-based sample of the US civilian, non-
institutionalized, adult population. One adult (aged 18+) was selected at random within each
household and was given the option of responding in English or Spanish. 5,586 individuals
completed the full interview, of whom 496 answered affirmatively to the question, “Are you
Hispanic or Latino?” and only their data were used in this study. Details on the survey’s
conception, testing, and implementation are published elsewhere (Nelson et al., 2004; Davis et
al., 2005). The questionnaire is available at: http://hints.cancer.gov/instrument.jsp.

Measures
Sun safety: This construct was assessed by five primary prevention behaviors. On a five-
point frequency scale (1=always; 5=never) respondents were asked how often they used
sunscreen, wore hats that shade the face, neck, and ears, wore long-sleeve shirts and long pants,
and stayed in the shade when outside for >1 hour on a warm, sunny day. The items were
comparable to the standardized set of survey measures of sun exposure and sun protection habits,
and are considered applicable across different populations (Glanz et al., 2008). Each item was
reverse-coded so that higher scores reflected higher endorsement of the behaviors. Since
correlation and exploratory factor analyses (EFA) showed that these variables were not highly
inter-correlated (all correlations <0.37), and consistent with our hypotheses, each item was treated
as separate outcome variable.

43
Acculturation: Five items were available for the assessment of the primary predictor:
nativity (US- or foreign-born), interview language (English or Spanish), a six-level item for
perceived comfort with the English language (1=completely comfortable; 6=do not speak
English), a continuous measure for age at US arrival (assessed among the foreign-born and
calculated by subtracting the year of birth from the year of arrival), and a continuous measure of
duration of US residence (assessed among the foreign-born and calculated by subtracting the year
of arrival from the year of data collection, 2005). As in prior acculturation research (Cantero et
al., 1999), each item was standardized (mean=0, standard deviation=1) and the Cronbach’s
reliability score was calculated (α=0.75). The acculturation index was obtained by computing the
mean of all items that loaded on a single factor in EFA.
Moderators and covariates: Moderators included gender, age, socioeconomic status
(SES) (education level measured on a continuous scale), and Census region of residence. The
division of regions into areas with high (South, West) and low (Midwest, Northeast) UVR relied
on Saraiya and colleagues’ (2002) methodology and evidence that latitude and mean annual UVR
index are strongly correlated (Eide et al., 2005). Nonetheless, these categories represent a crude
proxy for UVR exposure. Covariates included recoded dichotomous measures of marital status
(married/living with partner versus other), employment (employed/self-employed versus other),
availability of healthcare coverage (yes/no), current health status (1=excellent; 5=poor), and
cancer history (yes/no). All “refused” and “don’t know/other” responses were coded as missing.

Statistical analyses
We started with descriptive univariate and bivariate analyses, including stratified 2x2 chi-
square tests to assess the distinctiveness of the sun-safe behaviors. Because each sun-safe
behavior was a five-level dependent variable and because of the fairly large sample size (which

44
arguably justifies the use of parametric tests despite some non-normality of the underlying
distributions), multiple linear regression analyses were performed. In the adjusted models we
included only variables with statistical significance in the preliminary analyses in order to
minimize experimentwise Type I error and to investigate only highly targeted hypotheses (Cohen,
1990).
The differential effect of acculturation on sun safety based on gender, age, SES, and
region was assessed with tests for interaction. Because the primary interest in this study was in
establishing associations among the constructs, and not in making predictions or obtaining
population estimates, the magnitude of the regression slopes (betas) was considered and statistical
significance was established by the F test, without including sampling weights (especially since
only a small subset of the HINTS 2005 data was used). The level of significance for all tests was
set to alpha=0.05. All statistical procedures were performed with SAS, version 9.1 (SAS Institute
Inc., 2004).

Results
Sample characteristics
As summarized in Table 1, the mean age in the sample was 41.3 (SD=15.5; range: 18-95
years) and there were more women (61.5%) than men. About a third (35.7%) of the participants
were foreign-born and slightly over half (54.6%) were interviewed in Spanish. EFA of the
acculturation items revealed one factor (eigenvalue=2.31). Factor loadings for all items except
nativity ranged between 0.72 and 0.80. The nativity item did not load into the factor and was
dropped from further modeling.

45
Table 1. Demographic and acculturation characteristics of Latino
respondents in HINTS 2005 (N=496)
n %
Demographic variables
Gender
Male 191 38.5
Female 305 61.5
Age
18-34 188 37.9
35-49 180 36.3
50-64 75 15.1
65+ 52 10.5
Other/missing 1 0.2
Mean (SD) = 41.3 (15.5)
Marital status
Married 259 52.2
Divorced/widowed/separated 96 19.4
Never married 80 16.1
Living with partner 60 12.1
Other/missing 1 0.2
Healthcare coverage
Yes 299 60.3
No 196 39.5
Other/missing 1 0.2
Personal cancer history
Yes 33 6.7
No 459 92.5
Other/missing 4 0.8
Current health status
Excellent 38 7.7
Very good 82 16.5
Good 162 32.7
Fair 179 36.0
Poor 33 6.7
Other/missing 2 0.4

46
Table 1. Continued
n %
Education
Less than high school diploma 208 41.9
High school graduate/GED 111 22.4
Vocational/some college/Associate’s degree 100 20.2
College graduate/Bachelor’s degree 48 9.7
Master’s/Doctorate/Professional degree 27 5.4
Other/missing 2 0.4
Employment
Employed/self-employed 290 58.5
Homemaker/retired/student 153 30.9
Unemployed 26 5.2
Other/missing 27 5.4
Census region
Northeast 54 10.9
Midwest 50 10.1
South 179 36.1
West 213 42.9
Acculturation variables
US-born
Yes 177 35.7
No 319 64.3
Interview language
English 225 45.4
Spanish 271 54.6
Comfort with English
Completely comfortable 138 27.8
Very comfortable 103 20.8
Somewhat comfortable 48 9.7
A little comfortable 90 18.1
Not at all comfortable 47 9.5
Does not speak English 62 12.5
Other/missing 8 1.6

47
Table 1. Continued
n %
Duration of US residence
(among the foreign-born)

< 5 years 42 8.5
5 - 9 years 76 15.3
10 - 19 years 83 16.7
≥ 20 years 107 21.6
Don’t know/other/missing 188 37.9
Mean (SD) = 17.43 (14.47)
Age at arrival to the US
(among the foreign-born)

0-12 37 7.5
13-18 60 12.1
19-35 169 34.1
36-49 30 6.0
50+ 12 2.4
Don’t know/other/missing 188 37.9
Mean (SD) =23.95 (11.68)

The EFA results supported the inclusion of English comfort, interview language, age at US
arrival, and duration of US residence in the acculturation index, which had a mean of 0.15
(SD=0.83; range: -1.82 to 2.29). The index had relatively high reliability (Cronbach’s
alpha=0.75).

Sun safety prevalence and associations with demographic variables
As summarized in Table 2, wearing long pants and staying in the shade had the highest
“always use” endorsement (approximately 43% and 26%, respectively). Using sunscreen and
hats had the highest “never use” endorsement (approximately 39% and 33%, respectively).

48
Table 2. Prevalence of sun-safe behaviors among
Latino respondents in HINTS 2005 (N=496)
n %
Sunscreen
Always 76 15.3
Often 45 9.1
Sometimes 101 20.4
Rarely 65 13.1
Never 191 38.5
Other/missing 18 3.6
Mean (SD) = 1.48(1.48)
Shade
Always 128 25.8
Often 140 28.2
Sometimes 123 24.8
Rarely 59 12.0
Never 25 5.0
Other/missing 21 4.2
Mean (SD) = 2.60(1.16)
Hat
Always 110 22.1
Often 57 11.5
Sometimes 86 17.3
Rarely 59 12.0
Never 165 33.3
Other/missing 19 3.8
Mean (SD) = 1.77(1.58)
Long-sleeve shirt
Always 66 13.3
Often 53 10.7
Sometimes 103 20.8
Rarely 95 19.2
Never 161 32.4
Other/missing 18 3.6
Mean (SD) = 1.52(1.41)

49
Table 2. Continued
n %
Long pants
Always 213 42.9
Often 76 15.3
Sometimes 101 20.4
Rarely 40 8.1
Never 48 9.7
Other/missing 18 3.6
Mean (SD) = 2.77(1.36)

Data from the full HINTS 2005 sample showed that compared with other racial/ethnic groups,
wearing long-sleeve shirts and long pants were most frequently reported by Latinos (National
Cancer Institute, 2007).
Bivariate regression analyses revealed significant differences by gender, with women
being more likely to report using sunscreen and shade (both p<0.0001), and men being more
likely to report wearing protective clothing (all p<0.02). There was also a significant positive
association between age and wearing protective clothing (all p<0.05). Respondents from the
South and West were more likely than those elsewhere to report using sunscreen and protective
clothing (all p<0.05). Perceived health status was negatively associated with sunscreen use and
positively associated with wearing long pants (both p<0.05).
Education level was positively associated with sunscreen use (p<0.0001), but negatively
associated with staying in the shade, or wearing protective shirts or pants (all p<0.03).
Participants with healthcare coverage were less likely to report wearing long-sleeve shirts
(p=0.05) or long pants (p<0.0001), but more likely to report using sunscreen (p<0.02). No
statistically significant differences in sun safety by marital status, employment status, or personal
cancer history were found. Finally, more acculturated participants were more likely to be older,

50
to have more formal education, and to have healthcare coverage than their less acculturated
counterparts (all p<0.02). Because of the statistically significant correlation between education
and availability of healthcare coverage and for reasons of parsimony, only education was retained
as an SES indicator for further modeling.

Associations between acculturation and sun-safe behaviors
We performed several stratified 2x2 chi-square tests to assess whether sun safety varied
by acculturation when holding one of the sun-safe behaviors constant. Only for these
supplementary analyses we dichotomized acculturation into high/low (split by the mean) and each
sun-safe behavior into always/other. For example, we tested whether sun safety varied by
acculturation among all participants who reported always staying in the shade (N=128) and did
not find any statistically significant associations (all p>0.06). Among participants who reported
always wearing long pants when outdoors in the sun (N=213), the other sun-safe behaviors did
not vary by acculturation (all p>0.08). Thus, we confirmed the distinctiveness of the sun-safe
behaviors explored in the study.
As shown in Table 3, each bivariate relationship was in the hypothesized direction.

Table 3. Bivariate linear regression models of acculturation and sun-
safe behaviors among US Latinos (N=496)
F t p
Sunscreen 8.88 2.98 0.003
Shade 10.05 -3.17 0.002
Hat 0.00 0.06 0.953
Long-sleeve shirt 14.60 -3.82 0.000
Long pants 37.50 -6.12 <0.0001
Each sun-safe behavior coded on a 5-point scale: 0=never use to 4=always use;
acculturation assessed by an index including interview language, comfort with
English, age at US arrival, and duration of US residence

51
More acculturated participants were more likely to report using sunscreen (p<0.01) but less likely
to report staying in the shade (p<0.002), wearing long-sleeve shirts (p<0.001), or long pants
(p<0.0001) when in the sun than less acculturated participants. The association between
acculturation and wearing hats was not statistically significant (p>0.9) and was not explored
further. Following these analyses, we tested multiple linear regression models, the results of
which are summarized in Table 4.

Table 4. Multivariable linear regression models of sun-safe behaviors among US
Latinos (N=496)

Sunscreen Shade
Long-sleeve
shirt
Long pants

b SE b SE b SE b SE
Gender -0.53^ 0.13 -0.44^ 0.11 0.34
#
0.13 0.55^ 0.12
Age 0.00 0.00 0.01 0.00 0.01
#
0.00 0.01 0.00
SES/education 0.10
#
0.03 -0.01 0.03 -0.01 0.03 -0.08
#
0.03
Health status -0.13 0.07 0.04 0.05 -0.01 0.06 0.01 0.06
Region 0.58
#
0.16 0.09 0.13 0.56
#
0.16 0.24 0.15
Acculturation 0.06 0.09 -0.18
*
0.07 -0.29
#
0.09 -0.32
#
0.08
*
p<0.05;
#
p<0.01; ^p<0.0001; each sun-safe behavior coded on a 5-point scale: 0=never use to 4=always
use; acculturation assessed by an index including interview language, comfort with English, age at US
arrival, and duration of US residence; region= South/West versus Northeast/Midwest

All models were adjusted for gender, age, SES/education, health status, and region. There was no
statistically significant interaction between acculturation and gender or region, respectively,
regarding sun safety (results not shown). However, only among individuals with higher SES,
higher acculturation was associated with less use of long pants when outdoors in the sun
(p<0.0001). In turn, only among younger participants, higher acculturation was associated with
more sunscreen use (p<0.001) and less use of long-sleeve shirts (p<0.0001) (results not shown).
As hypothesized, acculturation was strongly and negatively associated with wearing long pants
when outside on sunny days (F=12.53, p<0.0001). Significant associations were also found for

52
staying in the shade and wearing long-sleeve shirts (F=5.58 and F=7.31, respectively, both
p<0.0001), with higher acculturation being associated with lower endorsement of these behaviors.
The positive association between acculturation and sunscreen use was attenuated following
adjustment for covariates (p>0.48).
Due to the lack of consensus about the assessment of acculturation (Zane et al., 2003;
Abraido-Lanza et al., 2006) and for purposes of increasing our confidence in the findings, we re-
estimated the adjusted models after replacing the acculturation index with the English comfort
variable, used as a proxy for acculturation. The results regarding use of sunscreen, long-sleeve
shirts and long-pants were practically identical across the two acculturation measures. The
findings regarding staying in the shade were similar across the two acculturation measures,
although the statistical significance of the link between English comfort and staying in the shade
was attenuated (results not shown).

Discussion
In our sample, 15% of the participants reported always using sunscreen when outdoors in
the sun for >1 hour, compared to 23% among Latino postal workers (Pichon et al., 2005) and
1.5% among Latino farmworkers (Salas et al., 2005). The postal workers study, however,
assessed sun safety during the previous five workdays; in the farmworkers study, sun safety was
assessed when in the sun for >15 minutes. Thus, discrepancies in reported sunscreen use might
be due to differences in measures and the demographic characteristics of the samples, especially
occupation. Among the farmworkers, for example, 83% had reported always/often wearing long-
sleeve shirts (Salas et al., 2005), compared to 24% in our study. Generally, about a third of US
adults report using sunscreen, shade, or protective clothing when outdoors in the sun (National
Cancer Institute, 2005).

53
Our findings support evidence about the adverse effect of acculturation on different
health practices (Vega et al., 1994; Lara et al., 2005; Abraido-Lanza et al., 2006) and are also
consistent with the Hispanic paradox (Cho et al., 2004). To our knowledge, this study is the first
to document that acculturation influences various skin cancer risk-related behaviors among
Latinos. Specifically, R
2
values in the full model for each sun-safe behavior except sunscreen use
increased compared to the R
2
values obtained when only the demographic variables were
modeled as predictors. This finding attested to the importance of acculturation in accounting for
variance in sun safety above and beyond demographic factors. Supporting some of the
hypotheses, acculturation was negatively associated with staying in the shade or wearing
protective clothing when outdoors on sunny days. Since dress customs are generally incorporated
faster with increasing acculturation than are less tangible qualities (Mendoza et al., 1981),
acculturation might lead to a more salient acceptance of US clothing and sun exposure practices
while displaying a weaker effect on sunscreen use. Regarding skin cancer prevention, we found
that less acculturated Latinos have better sun safety profiles than their more acculturated
counterparts, who are more similar to non-Latino whites.
No evidence of gender-specific acculturation effects in sun safety was observed, possibly
due to minor importance of social norms for these behaviors. However, the statistically
significant effects of age and SES as moderators are consistent with existing knowledge.
Specifically, education level has been shown to have a significant impact on sun safety of non-
Latino whites, possibly due to increased awareness about skin cancer risk factors and symptoms
(Hoegh et al., 1999; Azzarello et al., 2006)

and changes in leisure-time behavior patterns
(Bränström et al., 2004; Brown et al., 2006). In turn, research with diverse populations has
revealed that sun safety is positively associated with age (Stanton et al., 2004), conceivably due to
decreased importance of sensual and sociocultural factors (Leary et al., 1993; Baum et al., 1998).

54
Our acculturation index included items that are considered proxies of acculturation and
are often employed in acculturation research (Crespo et al., 2001; Abraido-Lanza et al., 2005;
Berrigan et al., 2006; Moran et al., 2007). Utilizing an index instead of a dichotomous measure is
consistent with modern conceptualizations of acculturation as a process and not a dichotomy.

Limitations
We utilized roughly 9% of the HINTS 2005 sample, which was predominantly (>73%)
non-Latino whites. In order to improve the representativeness of Latinos relative to national
estimates (U.S. Census Bureau, 2008), HINTS 2007 (data collected in 2008) will oversample
Latinos. Representativeness concerns are somewhat mitigated because many of the demographic
characteristics of our sample parallel those in nationally-representative health behavior research.
For example, weighted prevalence estimates from the 2000 National Health Interview Survey
(NHIS) (N=4,558 Latino adults) showed that 45% of the participants had less than high school
education (versus 42% in our study) and 38% were US-born (versus 36% in our study) (Berrigan
et al., 2006). In addition, current census information reveals that 33% of US Latinos do not have
health insurance (De Navas-Walt et al., 2007), which parallels our prevalence estimate.
Some selection bias - related to sampling Latinos by landline telephones, social
desirability effects, and non-response - might have occurred. As in other telephone surveys, the
HINTS response rate was low (Davis et al., 2005; Rutten et al., 2007). Regarding non-coverage,
estimates suggest that during the period January 2004 - June 2005, 7.2% of the Latino households
had only wireless telephones and 3.2% had no telephone service (Blumberg et al., 2006).
Research indicates that individuals without landline telephones might be more likely to smoke
and to be uninsured, and that wireless telephone substitution was greater among males, among
those <24 years of age, and those living in poverty (Blumberg et al., 2006). Thus, our findings

55
might be more generalizable to relatively more acculturated Latinos who have been in the US
longer and have greater resources. A selection bias in the direction of non-coverage of the least
acculturated Latinos might have occurred and the results might underestimate true effects.
Finally, psychometric and non-response analyses for HINTS are not currently available (Rutten et
al., 2007).
While using cross-sectional data prevented the establishment of causality, the
acculturation measure implied a linear time sequence. Regarding that measure, HINTS data
allowed for the assessment of acculturation on a unidirectional scale, but a bidirectional measure -
with Latino and US orientations as separate scores - would have been the measure of choice.
Also, for the acculturation index we utilized the mean of all available acculturation-related
variables for all available participants. In EFA, dichotomous variables do not perform as well as
multiple-category or continuous variables, and this might explain why nativity did not load onto
the acculturation factor and was excluded from the acculturation index. For the foreign-born
participants the mean acculturation score included interview language, comfort with English,
duration of US residence, and age at US arrival; for their US-born counterparts, the corresponding
score included language of interview and comfort with English. Language items, however,
generally explain most of the variance in acculturation scales (Marin, 1992; Epstein et al., 1996).
Another data limitation pertains to the inability to distinguish among Latino subgroups.
Yet, out of various risk factors explored among Latinos (e.g., lack of sun safety, smoking,
drinking, poor diet, insufficient exercise), the lack of sun safety has shown the least amount of
variation across countries of origin (Ramirez et al., 2004). This might be due to Latinos’
decreased awareness of sun safety compared to other health practices. Further, research has
suggested a potential confound (which could not be assessed with HINTS data) between
acculturation and skin color, such that the latter might impact one’s motivation or ability to

56
acculturate due to perceived socioeconomic discrimination against dark-skin individuals (Gomez,
2000; Murguia et al., 2002). However, this does not indicate that acculturation causes people to
discontinue engaging in sun safety.
Finally, research has documented a negative association between acculturation and
occupational physical activity, indicating that access to non-manual labor opportunities improves
with acculturation (Marquez et al., 2006). Since HINTS 2005 did not include occupation data,
we considered modeling income; however, the continuous measure of combined household
annual income had 185 (37%) missing responses. Due to evidence from NHIS 2000 that income
and education were strongly associated among Latinos (Berrigan et al., 2006), we were reassured
that our findings would not be seriously compromised by the exclusion of income. Nonetheless,
we acknowledge this as a limitation and suggest it for future research.

Future directions
Sun safety education initiatives often produce only modest behavioral effects among non-
Latino whites (Mahler et al., 1997; Baum et al., 1998; Weinstock et al., 2002) and it might be
even harder to reach Latinos. Our findings suggest that potentially successful efforts need to
target low- and high-acculturated Latinos separately and be age- and SES-tailored. Considering
factors associated with low acculturation (e.g., being monolingual, lack of healthcare coverage,
limited exposure to media-based health information, familism, manual occupations) (Ruiz et al.,
1992; Cuellar et al., 1995; Marquez et al., 2006; Mainous et al., 2006), sun safety messages for
low-acculturated Latinos could utilize informal, Spanish-language strategies, such as posting or
distributing leaflets advocating use of shading structures, sunscreen, and protective clothing in
bus stations, churches and community centers. Such messages could also emphasize the
importance of having a healthy family and could be incorporated in interactive activities at Latino

57
cultural events. Our findings also echo recommendations for outdoor worker protection (e.g.,
discussions about UVR, eye-catching sun safety posters, supervisor training) (International
Commission on Non-Ionizing Radiation Protection, 2007) since many outdoor Latino workers
exhibit low acculturation (Salas et al., 2005).
Overall, education efforts with low-acculturated Latinos could be resource-efficient since
the aim is to reinforce existing sun-safe behaviors. In contrast, our findings indicate that
education efforts with high-acculturated Latinos might require more resources due to the need for
behavior modification. Such initiatives could use the English language, emphasize modifying
individual sun safety, and could be delivered in health clinics and through the media, similar to
initiatives targeted at the whole US population.
The results reinforce existing services and policy recommendations (e.g., authorizing
periodic health surveillance for early identification of skin cancer, providing shading structures
and sunscreen dispensers at outdoor worksites, and requiring personal protection) (International
Commission on Non-Ionizing Radiation Protection, 2007). Finally, future research could
augment our findings by utilizing large, longitudinal datasets. Despite data limitations, however,
we demonstrated that acculturation has an adverse effect on Latinos’ skin cancer risk-related
behaviors. Appropriate interventions with this population could improve engagement in sun
safety as well as knowledge about skin cancer signs and symptoms, and could result in decreased
rates of advanced melanoma.

58
Chapter 4: Study 2

Acculturation and sun-safe behaviors among Latinos: Preliminary evidence for mediation by
health status, education level, and involvement with social networks

Valentina A. Andreeva,
1
Myles G. Cockburn,
1
Amy L. Yaroch,
2
Jennifer B. Unger,
1
Robert Rueda,
3
Kim D. Reynolds
1

1
Department of Preventive Medicine, Keck School of Medicine, University of Southern
California, Los Angeles, CA
2
Health Promotion Research Branch, Division of Cancer Control and Population Sciences,
National Cancer Institute, Bethesda, MD
3
Department of Psychology in Education, Rossier School of Education, University of Southern
California, Los Angeles, CA

Contact: Valentina A. Andreeva
Institute for Prevention Research
Keck School of Medicine, University of Southern California
1000 South Fremont Ave., Unit 8
Alhambra CA 91803
Tel: 626-457-6613/ Fax: 626-457-4012
Email: andreeva@usc.edu

59
Chapter 4 Abstract
According to recent research, Latinos’ acculturation might be negatively associated with
their sun-safe behaviors, which has implications for their skin cancer risk. We sought to elucidate
the relationship between acculturation and sun safety by exploring the potential mediating effects
of healthcare access, perceived health status, education level and involvement with social
networks. We tested structural equation models using data from 496 adult Latino respondents to
the 2005 Health Information National Trends Survey. Acculturation was assessed by a four-item
index. Applying sunscreen, seeking shade, and wearing protective clothing when outdoors on
sunny days were the primary outcomes, assessed by frequency scales. Overall, the hypothesized
mediators had the strongest explanatory potential for sunscreen use and the weakest - for seeking
shade. Education level emerged as a potential mediator for sunscreen use and wearing protective
clothing (both p<0.05). Perceived physical health status and involvement with social networks
showed mediating effects only for sunscreen use (p<0.05). Our findings have descriptive
importance and could inform intervention development as well as sun safety studies on causal
mechanisms. The results should be replicated with large Latino samples and longitudinal
datasets, encompassing a variety of psychosocial measures.

60
Introduction
Sun safety constitutes avoiding ultraviolet radiation (UVR) overexposure and is a key
behavioral component for skin cancer prevention regardless of skin type (U.S. Department of
Health and Human Services, 2000; American Cancer Society, 2008). Suggested sun-safe
practices include applying sunscreen, wearing protective clothing, and avoiding midday sun
exposure (U.S. Department of Health and Human Services, 2000; American Academy of
Dermatology, 2008). No racial/ethnic group, however, appears to meet current sun safety
recommendations (National Cancer Institute, 2005; Coups et al., 2008).

Moreover, recent
findings indicate that Latinos were at greater risk than non-Latino whites of experiencing
sunburn, while both groups were almost equally likely to report ≥4 sunburn episodes during the
preceding year (Saraiya et al., 2007).
Latinos generally experience more advanced melanoma tumors with a poorer prognosis
than non-Latino whites (Cress et al., 1997; Cormier et al., 2006; Cockburn et al., 2006; Hu et al.,
2006), possibly due to lower awareness about risk factors or symptoms (Olsen et al., 1993;
Pipitone et al., 2002; Byrd-Miles et al., 2007), lack of health insurance (Roetzheim et al., 1999;
De Navas-Walt et al., 2007), insufficient efforts for melanoma screening (Cockburn et al., 2006),
and delay in seeking treatment (Olsen et al., 1993; Friedman et al., 1994; Cormier et al., 2006).
Findings from the 1988-2001 California Cancer Registry highlighted steadily increasing invasive
melanoma incidence among Latinos (Cockburn et al., 2006). Sun safety initiatives, however,
often produce only modest behavioral effects among non-Latino whites (Mahler et al., 1997;
Baum et al., 1998; Weinstock et al., 2002) and Latinos might be even harder to reach.

Acculturation and health
Latinos’ acculturation has been established as an independent risk factor for a number of

61
health practices, including increased substance use and cancer incidence, and worsening of
dietary practices (Kaplan et al., 1990; Vega et al., 1994; Lara et al., 2005; Abraido-Lanza et al.,
2006; Barcenas et al., 2007). Regarding leisure-time UVR exposure, research suggests that beach
use might be higher among US-born and/or English-speaking Latinos than among their Mexico-
born, Spanish-speaking counterparts, after controlling for socioeconomic status (SES) (Wolch et
al., 2004). Further, recent findings documented that Latinos’ acculturation has a negative
association with seeking shade or wearing protective clothing when outdoors on sunny days
(Andreeva et al., In review). In order to further knowledge about the relationship between
acculturation and sun-safe behaviors, we explored the potential mediating effects of several
sociodemographic constructs. Generally, mediating variables improve the understanding of
causality and are best validated with experimental studies (MacKinnon et al., 2007).
Nonetheless, researchers advocate the utility of non-experimental, correlational analyses for
advancing the theoretical and empirical rationale that could guide interventions (Shrout et al.,
2002).

Potential mediators between acculturation and sun-safe behaviors
While the causal mechanisms of the effects of acculturation on health are not fully
understood (Hunt et al., 2004; Abraido-Lanza et al., 2006), acculturation has been positively
associated with healthcare access (Wells et al., 1989; Abraido-Lanza et al., 2004) which, in turn,
has been associated with sunscreen use, skin cancer examinations (Santmyire et al., 2001) and

sunburn (Hall et al., 2003). The potential mediating effect of healthcare access on health
practices might be due to having a regular healthcare provider (Moreland et al., 2006; Cairns et
al., 2006) thus a regular source of health information (Gombeski Jr et al., 1982; Wathen, 2006), a
favorable attitude towards provider communication behavior (Rutten et al., 2006), and a source of

62
recommendation for screening (Cairns et al., 2006). Further, research suggests that self-
perceptions of health status improve with acculturation (Markides et al., 1991; Shetterly et al.,
1996; Franzini et al., 2004; Lara et al., 2005; Kandula et al., 2007). Perceived health status, in
turn, might influence the amount of time spent outdoors. For example, sunburn frequency - an
estimate of UVR overexposure, dose-adjusted for skin type (Cockburn et al., 2006) - was higher
among individuals reporting excellent/good health than among others (Hall et al., 2003).
Studies also note that as Latinos acculturate, they become increasingly involved with a
wider social network (Griffith et al., 1985; Rueschenberg et al., 1989; Franzini et al., 2004).
Moreover, the composition of one’s social network has been suggested as a mediator between
acculturation and certain risk behaviors (Unger et al., 2000). Features of the social network (e.g.,
networks formed on the basis of occupation/recreational pursuits, belonging to a health club,
having friends who sunbathe) have also been correlated with individual sun-safe behaviors
(Keesling et al., 1987; Robinson et al., 2004). Having a trusted source of health information
(Gombeski Jr et al., 1982) and the opportunity to encounter health/risk behavior influences from
one’s peers and other contacts (Unger et al., 2000) have been suggested as reasons for the
potential mediating effects of the social networks construct.
Finally, acculturation has been positively associated with cultural capital (e.g., acquisition
of new information, second language fluency, formal education) (Rueda et al., 2003; Franzini et
al., 2004; Berrigan et al., 2006; Mainous et al., 2006; Pichon et al., 2007). In turn, research has
revealed a positive association between education level and skin cancer prevention/early
detection behaviors (e.g., sunscreen use, seeking shade, skin cancer examination), possibly due to
increased awareness about risk factors and/or symptoms (Michielutte et al., 1996; Hoegh et al.,
1999; Santmyire et al., 2001; Azzarello et al., 2006). Further, studies have suggested that adults
with a university degree might be more likely to vacation at seaside resorts (thus have increased

63
UVR exposure) than those who had completed only elementary school (Bränström et al., 2004).
Overall, we hypothesized that acculturation would have a positive association with
perceived health status, education level, involvement with social networks regarding health
matters, and healthcare access, respectively. These constructs were, in turn, expected to have
behavior-specific effects on sun safety. For example, we hypothesized that better health status
would be associated with more sunscreen use and less use of shade or protective clothing. We
also hypothesized that education, healthcare access, and the social networks construct would be
associated with more sun safety across all behaviors. The selected variables were grounded in
acculturation and health behavior theories, such as Bandura’s Social Cognitive Theory (SCT)
(1986) and the acculturation model summarized by Myers and Rodriguez (2003). SCT
accommodates the impact of the sociocultural environment on health behaviors and has been used
in prior sun safety research (Stone et al., 1999; Glanz et al., 1999; Robinson et al., 2004). In
addition, Myers and Rodriguez’s model highlights the indirect effects of acculturation on health,
however, many of the mediated paths have not been empirically examined.

Methods
Data source
We utilized cross-sectional 2005 HINTS data, gathered via geographic stratification and
list-assisted random-digit dialing, using a probability-based sample of the US civilian, non-
institutionalized, adult population. One adult (aged 18+) was selected at random within each
household and was given the option of responding in English or Spanish. 5,586 individuals
completed the full interview, of whom 496 answered affirmatively to the question, “Are you
Hispanic or Latino?” and only their data were used in this study. Details on the survey’s

64
conception, testing, and implementation are published elsewhere (Nelson et al., 2004; Davis et
al., 2005). The questionnaire is available at: http://hints.cancer.gov/instrument.jsp.

Measures
Primary outcome: Sun-safe behaviors constituted the primary outcome in this study and
were assessed by five-point frequency scales (1=always; 5=never). Respondents were told that
the questions pertained to sun protection practices and were asked how often they used sunscreen,
wore long-sleeve shirts, and long pants, and stayed in the shade when outside for ≥1 hour on a
warm, sunny day. An item about wearing wide-brim hats was also available, however, due to
lack of statistically significant main effects (Andreeva et al., In review) it was omitted from the
present analyses. The HINTS items were comparable to the standardized set of survey measures
of sun exposure and sun protection habits, and are considered applicable across different
populations (Glanz et al., 2008). All selected sun safety items were reverse-coded so that higher
scores reflected higher endorsement of the behaviors. Exploratory factor analyses (EFA) were
performed in order to assess clustering within latent factors.
Primary predictor: We assessed acculturation with an index which encompassed four
items: a dichotomous item for language of interview (English or Spanish), a six-level item for
perceived comfort with the English language (1=completely comfortable; 6=does not speak
English), a continuous measure for age at US arrival (assessed among the foreign-born and
calculated by subtracting the year of birth from the year of arrival), and a continuous measure for
the duration of US residence (also assessed among the foreign-born and calculated by subtracting
the year of arrival from the year of data collection, 2005). That index had a mean value of 0.15
(SD=0.83) and a Cronbach’s alpha of 0.75. Detailed information about the acculturation measure
is available elsewhere (Andreeva et al., In review).

65
Mediators: Six variables were available for the assessment of social networks. One was a
frequency item about the number of community organizations where the participant has
membership; another item assessed whether any of these organizations provided health
information (yes/no). We combined these two items into a single measure with the following
response categories: 0=no membership, 1=membership in a non-health organization, and
2=membership in a health-related organization. The next two items assessed whether the
participant had any family and/or friends with whom he/she talked about health (yes/no) and the
frequency of talking with them (1=not very frequently to 3=very frequently). We also combined
these two items into a single measure with the following response categories: 0=no
family/friends, 1=have family/friends but do not talk to them frequently, 2=have family/friends
and talk to them somewhat frequently, 3=have family/friends and talk to them very frequently.
The next social network item was a continuous measure for the number of neighbors on whom
the participant could rely for transportation for healthcare. The last item measured the frequency
(0=never, 4=every week) with which the participant attended religious services excluding
weddings and funerals. All social network variables were evaluated with EFA in order to explore
their factor clustering and to select the most reliable measures.
Healthcare access was assessed with two items: availability of health insurance (yes/no)
and one frequency item measuring use of services by a provider during the past 12 months
(excluding emergency room visits). These two variables were also evaluated with EFA for
potential modeling as a single access factor. Perceived physical health status was assessed with a
single item whose response options ranged from 1=excellent to 5=poor. This item was reverse-
coded so that higher scores corresponded to better health. Education level was assessed with a
single item with responses ranging from 1=never attended school to 11=professional/graduate
degree. Finally, all “refused” and “don’t know/other” responses were coded as missing.

66
Statistical analysis
Univariate and bivariate linear regression analyses, and EFA were performed using SAS
Version 9.1 (SAS Institute Inc., 2004). Principal component EFA with oblique and orthogonal
rotations was used for examining the factor loadings of the sun safety, social networks and
healthcare access items, respectively, and for selecting the most reliable measures. Then we
created a measurement model specifying the relationships between the measured and latent
variables. Next, we created a structural model specifying the relationships among the latent
factors (MacKinnon et al., 2003). Since the HINTS data are cross-sectional, we made the explicit
assumption that the primary predictor (acculturation) preceded each mediator, which preceded the
outcome (i.e., each sun-safe behavior). The complete model included unidirectional paths from
acculturation to each hypothesized mediator, as well as unidirectional paths from each
hypothesized mediator to each sun-safe behavior. It also included correlations between each two
mediating constructs.
Structural equation modeling (SEM) with standardized covariance matrices as input, the
maximum likelihood function with robust estimation, and the Lagrange multiplier tests were
employed in evaluating overall model fit and parameter estimates. Model fit was evaluated with
the chi-square goodness-of-fit statistic (which could be inflated by large sample sizes), the
comparative fit index (CFI) and the root mean-square error of approximation (RMSEA), both of
which are robust to sample size biases (Hu et al., 1995; Fan et al., 1997). Conventionally, CFI
values ≥0.95 and/or RMSEA values ≤0.06 signify appropriate fit (Hu et al., 1999; Buhi et al.,
2007). It has been shown that mild to moderate data non-normality has negligible effect on SEM
fit indices and parameter estimates (Fan et al., 1997). Also, the sample size (N=496) was
unlikely to cause improper SEM solutions (Fan et al., 1997).
SEM has been employed in prior sun safety research with Caucasians for the assessment

67
of theory-based mediation models (Turrisi et al., 1999; Jackson et al., 2000). One of the
advantages of SEM over conventional regression analysis pertains to the opportunity to
simultaneously assess direct and indirect effects (Byrne, 1994; Cole et al., 2003). Advantages of
SEM over Barron and Kenny’s (1986) causal steps mediation model include the provision of
direct hypothesis tests for mediation, adaptability to multiple mediators, and greater statistical
power (MacKinnon et al., 2002; Dearing et al., 2006). All SEM analyses were performed with
EQS version 6.1 for Windows (Bentler, 2004), the significance level was set at alpha=0.05, and
the provided p-values are two-sided.

Results
Sample characteristics
The mean age of the participants in this study was 41.3 (SD=15.5) and 61% were women.
About a third (35.7%) of the participants were foreign-born, slightly over half (54.6%) were
interviewed in Spanish, and about 42% reported having less than high school education. The
prevalence of the hypothesized mediating variables is summarized in Table 5. Overall, about
60% of the participants had health insurance and about half reported good or excellent physical
health status. Regarding organization membership (defined as membership in any group,
including a church, sports league, self-help group, service club, or professional organization),
close to half of the participants (43.6%) reported membership; about a fifth of the entire sample
indicted membership in a health-related organization. More information about the
sociodemographic characteristics of the participants are available elsewhere (Andreeva et al., In
review).

68
Table 5. Prevalence of the hypothesized mediating variables among
Latino respondents in HINTS 2005 (N=496)
n %
Healthcare access
Health insurance
Yes 299 60.3
No 196 39.5
Other/missing 1 0.2
Healthcare utilization/past 12 months/non-ER
None 132 26.6
1 time 106 21.4
2-4 times 144 29.0
5-9 times 58 11.7
10+ times 54 10.9
Missing 2 0.4
Social networks
Organization membership
None 277 55.8
Non-health organization membership 119 24.0
Health organization membership 97 19.6
Missing 3 0.6
Talk to family/friends about health
None 112 22.6
Have family/friends - don’t talk frequently 141 28.4
Have family/friends - talk somewhat frequently 133 26.8
Have family/friends - talk frequently 109 22.0
Missing 1 0.2
Education level
Less than high school diploma 208 41.9
High school graduate/GED 111 22.4
Vocational/some college/Associate’s degree 100 20.2
College graduate/Bachelor’s degree 48 9.7
Master’s/Doctorate/Professional degree 27 5.4
Other/missing 2 0.4

69
Table 5. Continued
n %
Perceived physical health status
Excellent 38 7.7
Very good 82 16.5
Good 162 32.7
Fair 179 36.0
Poor 33 6.7
Other/missing 2 0.4

Preliminary analyses
The EFA of the sun safety items revealed that wearing long-sleeve shirts and long pants
clustered into a single factor (e.g., protective clothing) with an eigenvalue of 1.56. Both items
had factor loadings >0.78. As expected, seeking shade and using sunscreen did not show a clear
factor pattern. Hence, these items were assessed as individual outcome variables in separate
models. Further, two indicators were selected for modeling the social networks construct: the
item assessing membership in health/non-health organizations and the item assessing
availability/frequency of talking to family/friends about health. These indicators were chosen
because of their high factor loadings, good theoretical justification, and good prevalence
distribution in the sample. Each of these two items had a factor loading of 0.75 and the factor had
an eigenvalue of 1.12. Finally, the two healthcare access items clustered well into a single factor
with an eigenvalue of 1.47. Each of these indicators had a factor loading of 0.86. Overall, two of
the four hypothesized mediators were treated as measured variables (health status, education) and
the other two (social networks, healthcare access) - as latent factors.
Prior to exploring mediated associations, we estimated bivariate linear regression models.
In these analyses, acculturation was the independent variable and use of sunscreen, shade, and

70
protective clothing were the respective dependent variables. Each of these models reached
statistical significance, indicating that acculturation was correlated with the sun-safe behaviors of
the participants.

Mediated associations
The results of the SEM analyses for use of sunscreen, shade and protective clothing are
summarized in Figures 2, 3, and 4, respectively. In all three models, the direct path between
acculturation and each of the mediating variables attained statistical significance and was in the
positive direction. As shown in Figure 2, three of the four hypothesized mediators attained
statistical significance in their direct effects on sunscreen use. Specifically, increased levels of
education, better self-perceptions of health, and more contact with friends/family and social
organizations regarding health matters were associated with increased sunscreen use when
outdoors for ≥1 hour on sunny days. The direct effect between acculturation and sunscreen use
was attenuated and became statistically non-significant, while the indirect effects of acculturation,
education level, health status, and social networks on sunscreen use remained statistically
significant. These results served as evidence for mediation. The model had very good statistical
fit (χ
2
=12.93, p>0.16; CFI=0.993, RMSEA=0.030). In addition, all of the hypothesized
correlations between each pair of potential mediators displayed statistical significance.

71
Figure 2. Mediated effects of acculturation on sunscreen use: structural equation analysis
Sunscreen use
Healthcare access
Social networks
Health status
Education
Acculturation
.501
.216
.495
.490
.182
.093
.538
χ
2
=12.93, df=9, p>0.16
CFI=0.993, RMSEA=0.030
.227
.167
.230
-.123
.156
.210
Note: Lines and curves indicate statistically significant (p<0.05) standardized paths and correlations; relationships that
did not reach statistical significance at p<0.05 are not shown.

The SEM results for seeking shade (Figure 3) were markedly different from those for
sunscreen use. Specifically, the direct path between acculturation and seeking shade remained
statistically significant, while neither the hypothesized direct effects of the potential mediators
nor the indirect effect of acculturation on seeking shade reached statistical significance. All of
the pairwise correlations among the potential mediators were again statistically significant. The
model also had good statistical fit as evidenced by the CFI=0.998 and RMSEA=0.017.

72
Figure 3. Mediated effects of acculturation on seeking shade: structural equation analysis

Staying in shade
Healthcare access
Social networks
Health status
Education
Acculturation
.503
.221
.456
.493
χ
2
=10.17, df=9, p>0.33
CFI=0.998, RMSEA=0.017
.227
.178
.204
-.124
.166
.234
-.178
Note: Lines and curves indicate statistically significant (p<0.05) standardized paths and correlations; relationships
that did not reach statistical significance at p<0.05 are not shown.

Regarding the relationship between acculturation and wearing protective clothing (Figure
4), our results revealed a statistically significant mediated association with education level. The
direct path between acculturation and the outcome remained statistically significant, indicating
that the potential mediated effect was partial. In addition, the relationship between acculturation
and wearing protective clothing appeared to be mediated only by the level of formal education,
such that higher education was linked to less use of protective clothing when outdoors for ≥1 hour
on sunny days. The model fit was again good (χ
2
=22.63, p>0.09; CFI=0.988, RMSEA=0.033).
The only hypothesized pairwise correlation that did not reach statistical significance was between

73
healthcare access and social networks. Finally, healthcare access did not display any mediated
associations in any of the models.

Figure 4. Mediated effects of acculturation on wearing protective clothing: structural equation
analysis
Wearing protective
clothing
Healthcare access
Social networks
Health status
Education
Acculturation
.503
.220
.450
.486
χ
2
=22.63, df=15, p>0.09
CFI=0.988, RMSEA=0.033
.227
.180
-.123
.166
.226
-.295
-.182

Note: Lines and curves indicate statistically significant (p<0.05) standardized paths and correlations;
relationships that did not reach statistical significance at p<0.05 are not shown.

Discussion
Utilizing SEM and a sample of US Latino adults, we observed that sunscreen use,
seeking shade, and wearing protective clothing when outdoors on sunny days were associated
with acculturation in distinctive ways. Specifically, the relationship between Latinos’
acculturation and their sunscreen use displayed evidence for mediation by perceived health status,

74
education level and involvement with social networks regarding health matters, supporting our
hypothesis. In that model, all mediated paths were positive, revealing that higher acculturation
levels were linked to increased levels of education, better self-perceptions of health, and more
contact with friends/family and social organizations regarding health matters. In turn, all of these
mediating constructs were associated with increased sunscreen use when outdoors for ≥1 hour on
sunny days. In contrast, the relationship between acculturation and wearing protective clothing
appeared to be mediated only by education level, such that higher education was linked to less
use of protective clothing. The analyses with seeking shade as the dependent variable did not
display any mediated associations.
Further, the direct effect of acculturation on sunscreen use was attenuated and lost
statistical significance, whereas it retained its statistical significance for seeking shade and
wearing protective clothing, respectively. These findings indicate that the selected set of
sociodemographic mediators might be more applicable to sunscreen use than to the other sun-safe
behaviors. In turn, this suggests the need for examining different causal pathways. For example,
psychosocial constructs such as body image, appearance motivation, tanning preferences, and
social norms might hold better explanatory potential regarding seeking shade or wearing
protective clothing when outdoors on sunny days (Bergenmar et al., 2001; Knight et al., 2002;
Hymowitz et al., 2006; Cafri et al., 2006). Some of these constructs (e.g., social norms, body
image) have indeed been shown to undergo changes as a result of acculturation in Latino samples
(Kranau et al., 1982; Cuellar, 1995 #107; Pepper et al., 2007). In the model regarding wearing
protective clothing, both acculturation and education level displayed statistically significant
direct, negative associations with that sun-safe behavior. Such findings are consistent with
evidence of SES-specific patterns in leisure-time behaviors and preferences (Bränström et al.,
2004; Brown et al., 2006). For example, adults with a university degree might be more likely to

75
vacation at seaside resorts (e.g., conceivably have increased UVR exposure and decreased use of
protective clothing) than those who had completed only elementary school (Bränström et al.,
2004).
Also, researchers have noted that dress customs are generally incorporated faster with
increasing acculturation than are less tangible qualities, such as values (Mendoza et al., 1981),
which suggests that acculturation might lead to a relatively quick adoption of US clothing
practices. Further, in our models for seeking shade the only statistically significant path was the
direct, negative association between acculturation and that sun-safe behavior. This finding is
consistent with recent evidence that linked acculturation with increased leisure-time UVR
exposure. Specifically, research in California has suggested that beach use might be higher
among US-born and/or English-speaking Latinos than among their Mexico-born, Spanish-
speaking counterparts, after controlling for SES (Wolch et al., 2004).
The coherence of the selected set of potential mediators was supported by the finding that
all pairwise correlations among them were in the expected direction and were statistically
significant (except the correlation between social networks and healthcare access in the protective
clothing model). In line with the literature, our results revealed that acculturation was positively
associated with self-perceptions of health, healthcare access, education level, and social networks.
Among these constructs, healthcare access was the only factor that did not display any mediated
associations in any of the models. While there is consistent evidence that acculturation leads to
increased healthcare access (Myers et al., 2003; Mainous et al., 2006), its effects on engagement
in health promoting behaviors are somewhat divergent and outcome-specific. For example,
analyses of the 2000 National Health Interview Survey (NHIS) have shown that in adjusted
models having a physician visit within the previous year and having health insurance (e.g., the
indicators used in the present study) were associated with a greater likelihood of reporting

76
sunburn during the preceding year (e.g., lack of sun safety) (Hall et al., 2003). Nonetheless, there
is also indication that having a recent general physical examination was positively associated with
sunscreen use and with reporting a recent skin cancer examination (Santmyire et al., 2001).
Overall, the impact of healthcare access on sun-safe behaviors merits further assessment.
The analyses in this study relied on SEM, which allows flexibility in specifying theory-
driven models (Buhi et al., 2007) and represents a confirmatory approach to multivariate analysis
(Byrne, 1994). Unlike traditional regression methodology, SEM affords the opportunity to study
both observed and latent constructs and thus better interpret the relationships among them,
simultaneously assess direct and indirect effects, and give unbiased estimates by explicitly
modeling measurement error (i.e., by modeling unique and common variance separately) (Byrne,
1994; Cole et al., 2003). In addition, advantages of SEM over the popular causal steps model by
Barron and Kenny (1986) are that it provides direct hypothesis tests for mediation, is easily
adaptable to multiple mediators, and has greater statistical power (MacKinnon et al., 2002;
Dearing et al., 2006). However, a true SEM model is a mathematical abstraction; thus our goal
was to produce a model that approximated the data well (Fan et al., 1997). Further, researchers
have noted that in mediation analyses, even when the data are cross-sectional - as in this study -
the underlying theory is inherently causal because it suggests a direction of influence (Rose et al.,
2004). Thus, our analyses sought to support assumptions of causality (not to prove causality) and
to better inform future mediation analyses with longitudinal data (Cole et al., 2003).

Limitations
A detailed discussion of the potential threats to the external and internal validity of the
findings due to data limitations is available elsewhere (Andreeva et al., In review). Briefly,
despite using roughly 9% of the HINTS sample, representativeness concerns are somewhat

77
mitigated because many of the demographic characteristics of our participants parallel those in
nationally-representative health behavior research. For example, weighted prevalence estimates
from the 2000 NHIS (N=4,558 Latino adults) showed that 45% of the participants had less than
high school education (versus 42% in our study) and 38% were US-born (versus 36% in our
study) (Berrigan et al., 2006). In addition, current census information reveals that 33% of US
Latinos do not have health insurance (De Navas-Walt et al., 2007) which parallels our prevalence
estimate. Nonetheless, in order to improve representativeness relative to national estimates (U.S.
Census Bureau, 2008), HINTS 2007 (data collected in 2008) will over-sample Latinos.
Some selection bias - related to sampling Latinos by landline telephones, social
desirability effects, and non-response - might have occurred. As in other telephone surveys, the
HINTS response rate was low (Davis et al., 2005; Rutten et al., 2007). Research indicates that
individuals without landline telephones might be more likely to be uninsured, and that wireless
telephone substitution was greater among those living in poverty (Blumberg et al., 2006).
Overall, our findings might be more generalizable to relatively more acculturated Latinos who
have been in the US longer and have greater resources. Non-response as well as psychometric
analyses for HINTS have been planned but are not currently available (Rutten et al., 2007).
Another study limitation pertains to the inability to understand causality with cross-
sectional data. First, it is possible that education, which we modeled as a potential mediator,
might precede and/or cause acculturation. Specifically, education may have occurred before
coming to the US or, among the foreign-born Latinos, education could affect the degree to which
they are able to acculturate. While there is strong evidence that acculturation and education are
positively associated (Alarcon et al., 1999; Franzini et al., 2004; Berrigan et al., 2006; Mainous et
al., 2006; Pichon et al., 2007; Barcenas et al., 2007) and our data were consistent with the
hypothesized models, it is important for future research to re-assess these models with

78
longitudinal data. Second, cross-sectional datasets have been used in a number of mediation
studies in the psychosocial literature, including studies on acculturation (Unger et al., 2000;
Pichon et al., 2007) and sun safety (Turner et al., 2005; Azzarello et al., 2006; Cafri et al., 2006),
and the main criticism again pertains to causality inferences (MacKinnon et al., 2007).
Researchers have observed that when all constructs were assessed at the same time, there might
be other models that could explain the data (MacKinnon et al., 2007). Whereas experimental
studies have been suggested as the best way to validate causal mechanisms (Kraemer et al., 2002;
Shrout et al., 2002; MacKinnon et al., 2007), in the case of acculturation it would not be possible
to assign participants to acculturation levels. Also, our aim was to advance the acculturation and
sun safety literatures by suggesting potential mediated associations, and not to prove causality.
Indeed, we were able to provide evidence that some mediated paths are more plausible than
others (Shrout et al., 2002). It has been noted that correlation analyses are valuable for theoretical
and empirical rationale development and for indicating possible points of intervention (Shrout et
al., 2002).
Finally, our acculturation index included items that are considered proxies of
acculturation and are often employed in acculturation research (Crespo et al., 2001; Abraido-
Lanza et al., 2005; Berrigan et al., 2006; Moran et al., 2007). Utilizing an index instead of a
dichotomous measure is consistent with modern conceptualizations of acculturation as a process
and not a dichotomy. Nonetheless, it is important to note that a bidirectional measure - with
Latino and US orientations as separate scores - would have been the measure of choice had the
data been available.

Conclusion and future directions
Whereas much of the sun safety literature pertains to non-Latino whites who exhibit the

79
highest skin cancer incidence (American Cancer Society, 2008), the changing demographics of
the US population and particularly the rapidly growing Latino segment necessitate the
understanding of the etiology and epidemiology of dermatologic disorders in non-Caucasian skin
types (Halder et al., 2003). Over 45 million Latinos live in the US (15% of the population),
primarily in areas with high UVR, such as California and Texas (U.S. Census Bureau, 2008).
The public health impact of Latinos’ health behaviors in general and their cancer-related
outcomes in particular is likely to intensify (Vega et al., 1994; Amaro et al., 2002; Hu et al.,
2004). Researchers have already advocated the need for melanoma screening among Latinos
(Cockburn et al., 2006) as well as for sun safety education (Hu et al., 2004; Byrd-Miles et al.,
2007; Andreeva et al., In review). Specifically, studies have indicated that potentially successful
efforts need to focus on low- and high-acculturated Latinos separately and be age- and SES-
targeted (Andreeva et al., In review). Due to evidence that gender did not impact the relationship
between acculturation and Latinos’ sun-safe behaviors (Andreeva et al., In review), we did not
include gender in this study. However, an important future step for sun safety research is the
assessment of effect modifiers (e.g., moderators) of the potential mediated relationships that were
observed here. Also, future research could assess the potential mediating effects of other
psychosocial constructs such as appearance motivation, body image, tanning attitudes, and social
norms on wearing protective clothing and seeking shade.
Finally, our results could inform public health research and intervention efforts aimed at
improving the sun-safe behaviors of US Latinos. For example, interventions focused on
increasing sunscreen use might need to account for the individuals’ self-perceptions of health,
their education levels, as well as their involvement in health-related organizations. The findings
could also serve to inform future mediation analyses with longitudinal data and large samples of
Latinos and other acculturating populations.

80
Chapter 5: Study 3

Acculturation and sun-safe behaviors: preliminary evidence for moderated mediation by
sociodemographic factors

Valentina A. Andreeva,
1
Amy L. Yaroch,
2
Jennifer B. Unger,
1
Myles G. Cockburn,
1

Robert Rueda,
3
Kim D. Reynolds
1

1
Department of Preventive Medicine, Keck School of Medicine, University of Southern
California, Los Angeles, CA
2
Health Promotion Research Branch, Division of Cancer Control and Population Sciences,
National Cancer Institute, Bethesda, MD
3
Department of Psychology in Education, Rossier School of Education, University of Southern
California, Los Angeles, CA

Contact: Valentina A. Andreeva
Institute for Prevention Research
Keck School of Medicine, University of Southern California
1000 South Fremont Ave., Unit 8
Alhambra CA 91803
Tel: 626-457-6613/ Fax: 626-457-4012
Email: andreeva@usc.edu

81
Chapter 5 Abstract
Recent research has revealed a negative association between Latinos’ acculturation and
their sun-safe behaviors, possibly mediated by education level, perceived health status, and social
network characteristics such as availability and frequency of contact regarding health matters.
We sought to further elucidate this relationship by exploring the moderating effects of gender and
health insurance on each mediated path. Data from 496 adult Latino respondents to the 2005
Health Information National Trends Survey were used. Acculturation, assessed by a four-item
index, was the primary predictor; sunscreen use, wearing long-sleeve shirts and long pants while
outdoors on sunny days were the primary outcomes, assessed by frequency scales. Moderated
mediation effects were tested with an established three-step causal moderation approach. The
results suggested that the mediated association between acculturation, education level and
sunscreen use might be stronger among women than men (p<0.08). Also, there was evidence that
the association between acculturation and involvement with social networks regarding health
matters might be stronger among insured than uninsured Latinos (p<0.08). Research ought to
replicate these findings with large samples of Latinos and longitudinal data. Overall, our results
suggest ways of refining the theoretical and empirical rationale for future sun safety interventions
with Latinos and for future analyses of causal mechanisms.

82
Introduction
Acculturation is considered an independent risk factor for a number of health practices
including sun safety (Kaplan et al., 1990; Vega et al., 1994; O'Malley et al., 1999; Lara et al.,
2005; Barcenas et al., 2007; Andreeva et al., In review). Sun safety constitutes actively taking
precautions against ultraviolet radiation (UVR) overexposure (by means of applying sunscreen,
seeking shade, wearing protective clothing, and avoiding midday sun exposure) and is important
for the primary prevention of skin cancer (U.S. Department of Health and Human Services, 2000;
American Cancer Society, 2008). Recent findings have revealed a negative association between
US Latinos’ acculturation and their use of sunscreen, shade and protective clothing when
outdoors on sunny days (Andreeva et al., In review). There is also evidence that
sociodemographic constructs might mediate these relationships (Andreeva et al., Manuscript in
preparation). Specifically, sunscreen use appeared to be mediated by one’s education level,
perceived physical health status and involvement with social networks. In that model, all
mediated associations were positive, revealing that higher acculturation levels were linked to
increased levels of education, better self-perceptions of health, and more contact with
friends/family and social organizations regarding health matters. In turn, all of these mediating
constructs were associated with increased sunscreen use when outdoors for ≥1 hour on sunny
days. In contrast, the relationship between acculturation and wearing protective clothing
appeared to be mediated only by the level of formal education, such that higher education was
linked to a lower endorsement of that sun-safe behavior.
Currently, no racial/ethnic group seems to meet existing sun safety recommendations
(National Cancer Institute, 2005; Coups et al., 2008). Considering the size and rapid growth of
the US Latino population, its sun-safe behaviors will have important public health education and
policy implications and merit better understanding. Generally, Latinos experience more

83
advanced skin cancer (e.g., melanoma) tumors with a poorer prognosis than non-Latino whites
(Cress et al., 1997; Cormier et al., 2006; Cockburn et al., 2006; Hu et al., 2006), possibly due to
their lower awareness about risk factors or symptoms (Olsen et al., 1993; Pipitone et al., 2002;
Byrd-Miles et al., 2007), lack of health insurance (Roetzheim et al., 1999; De Navas-Walt et al.,
2007), insufficient efforts for melanoma screening (Cockburn et al., 2006), and delay in seeking
treatment (Olsen et al., 1993; Friedman et al., 1994; Cormier et al., 2006). Nonetheless, primary
or secondary skin cancer prevention initiatives are rarely targeted at this population (Cockburn et
al., 2006). Comprehensive sun safety research with Latinos is also scarce.
Both the acculturation and sun safety literatures contain sufficient evidence about the
importance of health insurance regarding health and risk behaviors, including UVR exposure/sun
safety (Myers et al., 2003; Hall et al., 2003; Lara et al., 2005). For example, in adjusted models,
insured individuals were 15% more likely to report ≥1 sunburn episode during the preceding year,
compared to those without health insurance (Hall et al., 2003). Also, insured individuals were
more likely to report engaging in cancer screening compared to those without health insurance,
after adjustment for ethnicity, age, education, income, and quality of healthcare (Abraido-Lanza
et al., 2004). Theoretically, healthcare access has been proposed as a causal agent (e.g., mediator)
in individuals’ health practices (Myers et al., 2003). However, recent sun safety research tested
healthcare access (measured by availability of health insurance and healthcare services utilization
during the preceding year) and found no support for access acting as a mediator between
acculturation and sun safety (Andreeva et al., Manuscript in preparation). It might be a more
plausible or proximal mediator for other types of health/risk behaviors for which provider contact
is critical (e.g., cancer screening, smoking cessation) (Abraido-Lanza et al., 2004; Ranney et al.,
2006) than for sun-safe behaviors, which are dictated by a complex set of personal, sociocultural,

84
and environmental factors (Leary et al., 1993; Clarke et al., 1997; Jackson et al., 2000; Garbe et
al., 2000; Bränström et al., 2004; Thieden et al., 2005).
Some researchers have suggested socioeconomic status (SES) variables as potential
moderators for health behaviors and outcomes (Kraemer et al., 2006). SES markers have
typically included income/poverty measures and education (Chesney et al., 1982; Khan et al.,
1997; Borrayo et al., 2003; Franzini et al., 2004), however, access to services has also been
proposed as an important SES indicator (Hunt et al., 2004). Further, researchers have advocated
the need for a better understanding of the mechanisms through which SES impacts health
behaviors (Borrayo et al., 2003). A review of the acculturation literature noted that it is important
to investigate how SES factors might be associated with cultural factors (Hunt et al., 2004).
Since no evidence for mediation of sun-safe behaviors by healthcare access has been found, it is
possible that health insurance might act as a moderator of the mediated relationship between
acculturation and sun safety. For example, insured individuals might have more opportunities to
join provider-based health promotion groups/networks, and to use preventive services (Wells et
al., 1989; Abraido-Lanza et al., 2004), thus improving self-perceptions of health, compared to
their uninsured counterparts.
Further, evidence suggests that women experience more unfavorable acculturation effects
than men (Bethel et al., 2005; Zemore, 2007; Barcenas et al., 2007), which has been attributed to
baseline differences in certain behaviors (e.g., smoking) as well as gender-based social norms in
the countries of origin as well as the US (Alaniz et al., 1999; Caetano et al., 2008). Meanwhile,
across age and racial/ethnic groups, gender has consistently emerged as one of the strongest
predictors of sun-safe behaviors (especially sunscreen and shade use), with women reporting
higher rates of these behaviors than men (Keesling et al., 1987; Hall et al., 1997; Garbe et al.,
2000; Santmyire et al., 2001; Bränström et al., 2004; Stanton et al., 2004; Pichon et al., 2005;

85
Lewis et al., 2006). Since there is no evidence that the association between Latinos’ acculturation
and their sun-safe behaviors varies by gender (Andreeva et al., In review), the latter merits
attention in moderated mediation analyses. For example, it is possible that gender might
moderate the relationship between acculturation and involvement with social networks, since
social network dynamics have been shown to differ for men and women during the mainstream
integration process (Hagan, 1998). Also, the relationship between education and sun-safe
behaviors might be moderated by gender. This conjecture is based on two premises. First,
findings among college students (i.e., the same education level across gender) have revealed that
women were more likely to use sunscreen (Jerkegren et al., 1999) and had more positive attitudes
and intentions towards sun safety than men (Hillhouse et al., 1996). Second, women in general
have been shown to be more knowledgeable about skin cancer and sun-safe behaviors (Vail-
Smith et al., 1993; Hillhouse et al., 1996; Stanton et al., 2004).
In recent years, moderated mediation analyses have received attention in the literature as
useful tools for broadening the theoretical and empirical understanding of causal mechanisms
(Muller et al., 2005; Bauer et al., 2006; Edwards et al., 2007; Preacher et al., 2007). Evidence of
moderated mediation is obtained when a mediated path is shown to depend on the value of a
moderating variable, without the latter affecting the overall main effects (Muller et al., 2005).
For example, it is possible that sunscreen use might be mediated by one’s education level
(Michielutte et al., 1996; Hoegh et al., 1999) only among women, or that sun safety might be
mediated by social network characteristics (e.g., belonging to a health club) (Keesling et al.,
1987) only among individuals with health insurance.

The present study
We aimed to further the theoretical and empirical understanding of the potential causal

86
mechanisms associated with Latinos’ sun-safe behaviors, thus better inform health behavior
interventions with this population. We assessed how potential mediated paths between
acculturation and several sun-safe behaviors varied as a function of gender and health insurance.
The selected mediating constructs - education level, perceived health status, and social networks
(assessed in terms of availability and frequency of contact with friends/family and social
organizations regarding health matters) - have been suggested as possible sun safety mediators in
a sample of Latino adults (Andreeva et al., Manuscript in preparation). Specifically, we
hypothesized that the relationship between acculturation and social networks would be stronger
among men than women, while the relationship between social networks and sun-safe behaviors
would remain constant across gender. In turn, we hypothesized that the relationship between
education and sun-safe behaviors would be stronger among women than men. Regarding health
insurance, we expected that the association between acculturation and perceived health status and
involvement with social networks, respectively, would be stronger among insured compared to
uninsured Latinos. The selected constructs were grounded mainly in Bandura’s Social Cognitive
Theory (1986), which accommodates the impact of the sociocultural environment on health
behaviors and has been used in prior sun safety research (Stone et al., 1999; Glanz et al., 1999;
Robinson et al., 2004). The full theoretical model is presented in Figure 5.

87
Figure 5. Hypothetical model of moderated mediation of sun-safe behaviors among US Latinos
Sun-safe behaviors
Social networks
Education
Health status
Acculturation
Gender
Health insurance

Methods
Data source
We utilized cross-sectional 2005 HINTS data, gathered via geographic stratification and
list-assisted random-digit dialing, using a probability-based sample of the US civilian, non-
institutionalized, adult population. One adult (aged 18+) was selected at random within each
household and was given the option of responding in English or Spanish. 5,586 individuals
completed the full interview, of whom 496 answered affirmatively to the question, “Are you
Hispanic or Latino?” and only their data were used in this study. Details on the survey’s
conception, testing, and implementation are published elsewhere (Nelson et al., 2004; Davis et
al., 2005). The questionnaire is available at: http://hints.cancer.gov/instrument.jsp.

88
Measures
Primary outcome: Sun-safe behaviors constituted the primary outcome in this study and
were assessed by five-point frequency scales (1=always; 5=never). Respondents were told that
the questions pertained to sun protection practices and were asked how often they used sunscreen,
wore long-sleeve shirts, and long pants when outside for ≥1 hour on a warm, sunny day. Items
about wearing wide-brim hats and seeking shade were also available, however, due to lack of
statistically significant main and/or mediated effects (Andreeva et al., In review; Andreeva et al.,
Manuscript in preparation), they were omitted from the present analyses. The selected sun safety
items were comparable to the standardized set of survey measures of sun exposure and sun
protection habits, and are considered applicable across different populations (Glanz et al., 2008).
These variables were reverse-coded so that higher scores reflected higher endorsement of the
behaviors.
Primary predictor: We assessed acculturation with an index which encompassed four
items: a dichotomous measure for language of interview (English or Spanish), a six-level item for
perceived comfort with the English language (1=completely comfortable; 6=does not speak
English), a continuous measure for age at US arrival (assessed among the foreign-born and
calculated by subtracting the year of birth from the year of arrival), and a continuous measure for
the duration of US residence (also assessed among the foreign-born and calculated by subtracting
the year of arrival from the year of data collection, 2005). That index had a mean value of 0.15
(SD=0.83) and a Cronbach’s alpha of 0.75. Detailed information about the acculturation measure
is available elsewhere (Andreeva et al., In review).
Mediators: Perceived health status was assessed by a single item with response options
ranging from 1=excellent to 5=poor. This item was reverse-coded so that higher scores
corresponded to better health. The level of formal education was also assessed with a single item

89
with response options ranging from 1=never attended school to 11=professional/graduate degree.
The social networks construct was assessed by the mean of two items (Andreeva et al.,
Manuscript in preparation): one assessed community organization membership (0=no
membership; 1=membership in a non-health organization; 2=membership in a health-related
organization) and the other assessed having family/friends with whom to discuss health (0=no
family/friends; 1=have family/friends but do not talk to them frequently; 2=have family/friends
and talk to them somewhat frequently; 3=have family/friends and talk to them very frequently).
Moderators and covariates: Moderating variables included gender and having health
insurance (yes/no). Covariates included age and Census region of residence and were informed
by prior research (Andreeva et al., In review). Finally, all “refused” and “don’t know/other”
responses were coded as missing.

Statistical analysis
Univariate analyses of all variables were followed by preliminary t-tests assessing the
variation in each potential mediator and each outcome variable by each hypothesized moderator.
All continuous variables were centered on their respective means prior to modeling. For the
moderated mediation analyses, we followed the procedures outlined by Muller, Judd and
colleagues (2005). Specifically, moderated mediation effects were assessed by creating
interaction terms between each hypothesized moderator and acculturation in three equations:
(4) Y= α
10
+ b
11
X + b
12
Z + b
13
X*Z + e
1

(5) M= α
20
+ b
21
X + b
22
Z + b
23
X*Z + e
2

(6) Y= α
30
+ b
31
X + b
32
Z + b
33
X*Z + b
34
M + b
35
M*Z + e
3

90
In these equations, Y represents the primary outcome (sun-safe behaviors), X represents the
primary predictor (acculturation), M represents a mediator, Z represents a moderator, α represents
an intercept, and e represents an error term. The first equation assessed the overall effect of
acculturation on sun-safe behaviors, as well as whether this effect was moderated by Z. The
second equation assessed whether the effect of acculturation on the mediator (M) differed across
levels of the moderator (Z). The third equation assessed the moderated effect of the mediator on
sun-safe behaviors, and the residual moderated effect of acculturation on sun-safe behaviors, after
controlling for the mediator. According to Muller and colleagues, in order to establish that
moderated mediation exists, the following three criteria must be met: 1) in equation (1),
acculturation should have a significant effect on the sun-safe behavior, but this overall effect
should not be moderated (e.g., b
11
≠0; b
13
=0); 2) in equations (2) and (3) the effect of acculturation
on the mediator must depend on the moderator, and/or the partial effect of the mediator on the
sun-safe behavior must depend on the moderator (e.g., both b
23
and b
34
≠0 or both b
21
and b
35
≠0);
and 3) the residual direct effect of acculturation on the sun-safe behavior after controlling for the
mediator must be moderated (e.g., b
33
), however, statistical significance is not necessary.
Moderated mediation analyses entail the assumptions that the predictor and moderator are
independent, and the residuals are independent and normally distributed (Muller et al., 2005).
We conducted analyses estimating equations (1), (2) and (3) separately for each
combination of the posited mediators and moderators. To avoid redundancy, only equation (3)
was adjusted for age and region of residence. Led by prior findings (Andreeva et al., In review;
Andreeva et al., Manuscript in preparation), the model with sunscreen use as the dependent
variable was hypothesized to include all three mediators (education level, perceived health status,
and social networks) and the two moderators (gender, health insurance). The model with
protective clothing as the dependent variable was hypothesized to include only one mediator

91
(education level) and the two moderators. The level of significance for all tests was set to
alpha=0.05 and all statistical procedures were performed using SAS Version 9.1 (SAS Institute
Inc., 2004).

Results
Univariate and bivariate analyses of the mediating and outcome variables, stratified by
the levels of each moderator, are summarized in Table 6. Independent-sample t-tests revealed
statistically significant differences in use of sunscreen, long pants (both p<0.0001) and long-
sleeve shirts (p<0.02) by gender. Specifically, women were more likely to report sunscreen use,
while men were more likely to report wearing protective shirts or pants. No gender-related
differences were observed in any of the three hypothesized mediators (all p>0.34).

Table 6. Univariate and bivariate analysis of the hypothesized mediating and outcome
variables (N=496)

Gender Health insurance

Male
(N=191)
Female (N=305) Yes (N=299) No (N=196)
Mean SD Mean SD Mean SD Mean SD
Sunscreen 1.13
*
1.32 1.69
*
1.54 1.61
*
1.45 1.26
*
1.51
Long-sleeve shirt 1.71
*
1.46 1.39
*
1.36 1.41 1.32 1.68 1.52
Long pants 3.11
*
1.25 2.55
*
1.38 2.54
*
1.38 3.11
*
1.26
Education 5.20 2.39 5.33 2.42 5.88
*
2.47 4.38
*
2.00
Health Status 1.87 0.99 1.79 1.06 1.91
*
1.04 1.69
*
1.02
Social networks 1.02 0.74 1.09 0.69 1.16
*
0.68 0.92
*
0.73
*
p<0.05 from independent-sample t-tests; each sun-safe behavior coded on a 5-point scale: 0=never use to
4=always use; formal education assessed on an 11-point scale: 1=never attended school to
11=professional/doctorate degree; perceived health status assessed on a 5-point scale: 0=poor to 4=excellent;
social networks assessed by the mean of two items: organization membership (0=no membership;
1=membership in a non-health organization; 2=membership in a health-related organization) and having
family/friends with whom to discuss health (0=no family/friends; 1=have family/friends but do not talk to
them frequently; 2=have family/friends and talk to them somewhat frequently; 3=have family/friends and talk
to them very frequently).

92
In contrast, participants without health insurance reported lower health status levels
(p<0.02), lower education levels (p<0.0001), and less availability/contact with social networks
regarding health matters (p<0.001) compared to their insured counterparts. Health insurance
status also produced statistically significant differences in the sun-safe behaviors. Unlike the
uninsured, insured participants were more likely to report using sunscreen (p<0.02) but less likely
to report wearing long pants (p<0.0001) when outdoors in the sun. Consistent with prior research
(Andreeva et al., In review) and in order to minimize experimentwise Type I error and to
investigate only highly targeted variables and hypotheses (Cohen, 1990), the mean of the long-
sleeve shirt and long pants items was computed and used as a protective clothing outcome
variable. Sunscreen use was assessed as a separate dependent variable.
Results of the three-step approach for assessing moderated mediation are summarized in
Table 7 for sunscreen use and Table 8 for wearing protective clothing. In equation (1), we
observed statistically non-significant interaction terms between acculturation and gender and
health insurance, respectively (all p>0.16). Thus, there was no evidence of mediated moderation
associated with gender and health insurance in our sample and the first criterion for moderated
mediation was met (Muller et al., 2005). The results from equations (2) and (3) for sunscreen use
revealed marginally significant support for the second criterion for moderated mediation.
Specifically, the partial effect of education level on sunscreen use appeared to be moderated by
gender (e.g., b
35
=-1.12, p<0.08), with an overall statistically significant effect of acculturation on
education level (b
21
=1.38
;
p<0.0001). In this model, the association between education level and
sunscreen use was stronger among women than men. Thus, we found support for one of our
hypotheses.

93
Table 7. Analyses of moderated mediation between acculturation and sunscreen use among US Latinos

Equation 1 Equation 2 Equation 3

b SE t b SE t b SE t
X: Acculturation 0.30 0.10 2.93
^
1.38 0.14 9.71
*
0.09 0.11 0.80
Z: GE -0.56 0.14 -4.11
*
-0.11 0.19 -0.59 -0.54 0.13 -4.03
*

X*Z: Acculturation* GE -0.16 0.17 -0.97 0.32 0.23 1.38 -0.02 0.19 -1.10
M: Education 0.16 0.04 4.20
*

M*Z: Education* GE -1.12 0.07 -1.76
&

X: Acculturation 0.30 0.10 2.93
^
0.29 0.07 4.18
*
0.27 0.10 2.59
#

Z: GE -0.56 0.14 -4.11
*
0.09 0.09 0.93 -0.56 0.13 -4.22
*

X*Z: Acculturation* GE -0.16 0.17 -0.97 -0.07 0.11 -0.60 -0.15 0.17 -0.90
M: Health status 0.19 0.08 2.30
#

M*Z: Health status* GE -0.03 0.13 -0.20

X: Acculturation 0.30 0.10 2.93
^
0.18 0.05 3.75
^
0.21 0.10 2.09
#

Z: GE -0.56 0.14 -4.11
*
-0.06 0.06 -0.91 -0.54 0.13 -4.13
*

X*Z: Acculturation* GE -0.16 0.17 -0.97 -0.05 0.08 -0.61 -0.09 0.16 -0.54
M: Social networks 0.58 0.12 4.70
*

M*Z: Social networks*
GE
-0.24 0.19 -1.30

X: Acculturation 0.33 0.15 2.15
#
1.10 0.21 5.26
*
0.23 0.16 1.45
Z: HI 0.18 0.16 1.11 0.50 0.22 2.31
#
0.16 0.16 0.98
X*Z: Acculturation* HI -0.23 0.19 -1.21 0.47 0.26 1.79
&
-0.36 0.21 -1.74
M: Education 0.08 0.06 1.48
M*Z: Education* HI 0.06 0.07 0.89

X: Acculturation 0.33 0.15 2.15
#
0.16 0.10 1.54 0.28 0.15 1.88
&

Z: HI 0.18 0.16 1.11 0.06 0.11 0.53 0.23 0.16 1.45
X*Z: Acculturation* HI -0.23 0.19 -1.21 0.16 0.13 1.23 -0.22 0.19 -1.18
M: Health status 0.18 0.11 1.73
M*Z: Health status* HI -0.03 0.14 -0.21

X: Acculturation 0.33 0.15 2.15
#
0.02 0.07 0.35 0.28 0.15 1.97
#

Z: HI 0.18 0.16 1.11 0.18 0.07 2.39
#
0.16 0.16 1.03
X*Z: Acculturation* HI -0.23 0.19 -1.21 0.16 0.09 1.77
&
-0.26 0.18 -1.40
M: Social networks 0.54 0.15 3.70
^

M*Z: Social networks* HI -0.09 0.19 -0.47
*
p<0.0001;
^
p<0.01;
#
p<0.05;
&
p<0.08; GE, gender; HI, health insurance

94
Table 8. Analyses of moderated mediation between acculturation and wearing protective clothing
among US Latinos

Equation 1 Equation 2 Equation 3

b SE t b SE t b SE t
X: Acculturation -0.30 0.08 -3.98
*
1.38 0.14 9.71
*
-0.27 0.08 -3.17
^

Z: GE 0.43 0.10 4.27
*
-0.11 0.19 -0.59 0.45 0.10 4.49
*

X*Z: Acculturation* GE -0.17 0.12 -1.41 0.32 0.23 1.38 -0.06 0.15 -0.43
M: Education -0.01 0.03 -0.48
M*Z: Education* GE -0.08 0.05 -1.70

X: Acculturation -0.38 0.11 -3.29
^
1.10 0.21 5.26
*
-0.36 0.12 -2.95
^

Z: HI -0.15 0.12 -1.29 0.50 0.22 2.31
#
-0.19 0.12 -1.51
X*Z: Acculturation* HI -0.08 0.14 0.57 0.47 0.26 1.79
&
0.19 0.16 1.21
M: Education -0.01 0.04 -0.23
M*Z: Education* HI -0.05 0.05 -1.05
*
p<0.0001;
^
p<0.01;
#
p<0.05;
&
p<0.08; GE, gender; HI, health insurance

Further, in the model for sunscreen use, we found evidence of moderated mediation by
health insurance and social networks. In particular, the effect of acculturation on involvement
with social networks depended on health insurance status (b
23
=0.16, p<0.08), with a statistically
significant, partial effect of social networks on sunscreen use (b
34
=0.54, p<0.001). Among
insured participants, the relationship between acculturation and involvement with social networks
regarding health matters (in terms of availability and frequency of contact) was stronger, than
among uninsured Latinos. This finding provided support for another one of our hypotheses. No
support was found for any of the other hypothesized paths in Figure 5. In addition, we found no
evidence of moderated mediation regarding wearing protective clothing in our sample of Latinos.

Discussion
We aimed to extend the theoretical and empirical understanding regarding the potential
causal mechanisms of Latinos’ sun-safe behaviors. To our knowledge, only one other sun safety
study has relied on moderated mediation analyses (Reynolds et al., In review). In that study, the

95
researchers explored moderated mediation regarding the effects of a skin cancer prevention
program for middle school youth and found that the mediating effect of knowledge was
moderated by sun sensitivity, while the mediating effect of self-efficacy was moderated by
gender. No moderated mediation studies in the acculturation literature have been identified.
Our findings suggested the presence of moderated mediation regarding the relationship
between acculturation and sunscreen use. In particular, the impact of acculturation on sunscreen
use was mediated by education level but this mediating pathway varied by gender. As
hypothesized, the link between education level and sunscreen use was stronger in women than
men. This finding is in line with prior research with non-Latino whites. Specifically, there is
evidence that among individuals with the same level of formal education, compared to men,
women were more likely to use sunscreen (Jerkegren et al., 1999), to have more positive attitudes
and intentions towards sun safety (Hillhouse et al., 1996), but also more likely to use tanning
salons (during the preceding year) (Hillhouse et al., 2000; Danoff-Burg et al., 2006). Also,
women in general have been shown to be more knowledgeable about skin cancer and sun-safe
behaviors (Vail-Smith et al., 1993; Hillhouse et al., 1996; Stanton et al., 2004). It is possible that
women are more likely than men to attend to sun safety/skin cancer information in order to
determine how to tan “safely.” Therefore, sunscreen interventions with low- or high-acculturated
Latinos might have better potential for success if they are gender- and education level-targeted.
Further, the results revealed that the impact of acculturation on sunscreen use was
mediated by involvement with social networks and this mediating pathway varied by health
insurance status. Also as hypothesized, the link between acculturation and social networks was
stronger for insured than uninsured Latinos. Insured individuals might have more opportunities
to join provider-based health promotion groups/networks, compared to their uninsured
counterparts. Thus, our findings could inform future sunscreen use efforts with Latinos by

96
suggesting the importance of targeting individuals at the same health insurance status and the
importance of capitalizing on social network characteristics, such as availability and frequency of
contact regarding health matters.
In terms of wearing protective clothing when outdoors in the sun, we did not find any
statistically significant or marginally significant evidence of moderated mediation. The model,
however, included only one mediator (education level). As suggested by prior research, for this
sun-safe behavior it might be necessary to explore the potential mediating effects of other
psychosocial constructs such as appearance motivation, body image, tanning attitudes, and social
norms (Bergenmar et al., 2001; Knight et al., 2002; Cafri et al., 2006; Hymowitz et al., 2006;
Andreeva et al., In review). Some of these constructs (e.g., social norms, body image) have
indeed been shown to undergo changes as a result of acculturation in Latino samples (Kranau et
al., 1982; Cuellar, 1995 #107; Pepper et al., 2007). In addition, country of origin/ethnicity merits
attention as a potential moderator (Ramirez et al., 2004; Perez-Escamilla et al., 2007) of the
association between acculturation and wearing protective clothing. It has been noted that specific
features of different Latino cultures (e.g., values, norms) might have independent effects on
health behaviors and outcomes (Khan et al., 1997). These variables, however, were not available
in the HINTS 2005 dataset.
Finally, in addition to the three-step linear regression approach used in this study, other
methods for assessing moderated mediation also exist, such as the piecemeal approach, the
subgroup approach, and moderated path analysis (Edwards et al., 2007). While each method has
certain deficiencies (Edwards et al., 2007), Muller and colleagues’ method appears to be fairly
flexible and very well defined, and has few known important limitations. In addition, it was
shown to be the most popular moderated mediation approach in the psychology literature
according to a recent review (Edwards et al., 2007).

97
Limitations and future directions
A detailed discussion of the potential threats to the external and internal validity of the
findings due to data limitations is available elsewhere (Andreeva et al., In review). Briefly,
despite using roughly 9% of the HINTS sample, representativeness concerns are somewhat
mitigated because many of the demographic characteristics of our participants paralleled those in
nationally-representative health behavior research. For example, weighted prevalence estimates
from the 2000 National Health Interview Survey (N=4,558 Latino adults) showed that 45% of the
participants had less than high school education (versus 42% in our study) and 38% were US-
born (versus 36% in our study) (Berrigan et al., 2006). In addition, current census information
reveals that 33% of US Latinos do not have health insurance (De Navas-Walt et al., 2007), which
parallels our prevalence estimate. Nonetheless, in order to improve representativeness relative to
national estimates (U.S. Census Bureau, 2008), HINTS 2007 (data collected in 2008) will over-
sample Latinos.
Further, psychometric and non-response analyses for HINTS 2005 have been planned but
are not currently available (Rutten et al., 2007). It is possible that some of the measures used in
this study might have poor reliability coefficients, which could weaken associations and account
for some of the null effects (e.g., in the model for wearing protective clothing). However, this
problem could not be assessed at present due to the lack of reliability values.
Regarding the three-step causal moderation approach used in this study, some limitations
pertain to the possibility that the effect of the moderator might be in the opposite direction than
the direct and mediated effects, or that the moderator might have non-linear effects (Edwards et
al., 2007). It is possible that the lack of strong statistical significance in our models might be due
to the presence of non-linear effects or to the relatively small sample size for this type of
analyses. Future research ought to replicate these findings with large samples of Latinos and with

98
longitudinal data. Also, in our models there might be a certain degree of violation of one of the
assumptions for moderated mediation analyses. Specifically, the moderator and predictor are
assumed to be independent (Muller et al., 2005), however, in our sample acculturation and health
insurance had a Pearson correlation coefficient=0.45. In order to access the potential collinearity
between these two constructs, we computed their tolerance values. Generally, tolerance values
range between 0 and 1, with lower values indicating increasing collinearity. In our models,
however, the tolerance values were relatively high (0.79). The conceptual distinction between
acculturation and health insurance, as well as the fact that their correlation was not very high,
suggested that the lack of complete empirical independence might not pose a serious threat to our
findings. Nonetheless, the findings might be affected to a certain degree by decreasing the
straightforward interpretation of the regression slopes for the two variables, as well as
compromising the potential for establishing statistical significance.
Another limitation related to the use of cross-sectional data pertains to the inability to
disentangle causation. First, it is possible that education, which we modeled as a potential
mediator, might precede and/or cause acculturation. Specifically, education may have occurred
before coming to the US or, among the foreign-born Latinos, education could affect the degree to
which they are able to acculturate. While there is strong evidence that acculturation and
education are positively associated (Alarcon et al., 1999; Franzini et al., 2004; Berrigan et al.,
2006; Mainous et al., 2006; Pichon et al., 2007; Barcenas et al., 2007), and our data were
consistent with the hypothesized models, it is important for future research to re-assess these
models with longitudinal data. Second, cross-sectional datasets have been used in a number of
mediation studies in the psychosocial literature, including studies on acculturation (Unger et al.,
2000; Pichon et al., 2007) and sun safety (Turner et al., 2005; Azzarello et al., 2006; Cafri et al.,
2006), and the main criticism again pertains to causality inferences (MacKinnon et al., 2007).

99
Researchers have observed that when all constructs were assessed at the same time, there might
be other models that could explain the data (MacKinnon et al., 2007). While experimental studies
have been suggested as the best way to validate causal mechanisms (Kraemer et al., 2002; Shrout
et al., 2002; MacKinnon et al., 2007), in the case of acculturation it would not be possible to
assign participants to acculturation levels. Also, our aim was to advance the acculturation and
sun safety literatures by suggesting potential associations of moderated mediation, and not to
prove causality. It has been noted that correlation analyses are valuable for theoretical and
empirical rationale development and for indicating possible points of intervention (Shrout et al.,
2002).
Overall, psychosocial constructs (e.g., body image, social norms, tanning attitudes) merit
future investigation as potential mediators in studies of causality of sun-safe behaviors,
particularly in regards to wearing protective clothing. In addition, it would be useful to examine
other potential moderators of the observed mediated associations (e.g., Latino
subgroups/ethnicity). Following replication, our findings could inform sun safety interventions
with Latinos by suggesting potential mediating constructs (e.g., education level, social networks)
and the populations (e.g., women, insured) for which such mediators might have increased
impact. For example, interventions aimed at increasing sunscreen use could be tailored on the
participants’ acculturation levels, could be gender- and insurance status-specific, and could
capitalize on health-related social networks to correct perceptions about “safe” tanning.
Researchers have already noted that melanoma represents a public health concern for all
races/ethnicities (Cormier et al., 2006). Considering the fact that Latinos constitute the fastest
growing segment of the US population (U.S. Census Bureau, 2006), the public health impact of
their health behaviors in general and their cancer-related outcomes in particular is likely to
intensify. Prior research has suggested that potentially successful efforts need to focus on low-

100
and high-acculturated Latinos separately and be age- and SES-targeted (Andreeva et al., In
review). The results of the present study suggest ways to further fine-tune the theoretical and
empirical foundations of future sun safety research and education programs for Latino adults.

101
Chapter 6: Discussion

Summary of findings with respect to each hypothesis
Overall, the findings of this research are consistent with evidence about the adverse effect
of acculturation on different health practices (Vega et al., 1994; Lara et al., 2005; Abraido-Lanza
et al., 2006) and are also in line with the Hispanic paradox (e.g., that Hispanic immigrants tend to
be healthier than US-born Hispanics) (Cho et al., 2004). This research is the first to document
that acculturation influences various skin cancer risk-related behaviors among Latinos.
Regarding skin cancer prevention, it appears that less acculturated Latinos have better sun safety
profiles than their more acculturated counterparts, who are more similar to non-Latino whites.

Study 1
The following three hypotheses were tested in Study 1:
Hypothesis 1: High-acculturated Latinos would report more sunscreen use when
outdoors in the sun than their low-acculturated counterparts.
Hypothesis 2: Acculturation would be negatively associated with use of shade or
protective clothing when outdoors in the sun.
Hypothesis 3: The observed effects of acculturation on Latinos’ sun-safe behaviors
would be stronger among women than men, among younger than older Latinos, among
individuals of higher versus lower SES, and more pronounced in areas with higher versus lower
UVR levels.
In bivariate linear regression models, Hypothesis 1 received support since more
acculturated participants were shown to be more likely to report using sunscreen (p<0.01).
However, that association became statistically non-significant after adjustment for

102
sociodemographic covariates in multivariable models. It is possible that in this sample of
Latinos, demographic variables such as gender, age, and SES accounted for most of the variance
in sunscreen use. In addition, the prevalence distribution of sunscreen use in the sample was
somewhat skewed, which might have compromised the potential for detecting effects.
Specifically, only 15% of the participants reported always using sunscreen versus 39% who
reported never using sunscreen when outdoors for ≥1 hour on a sunny day.
In support of Hypothesis 2, the results of the adjusted linear regression models revealed
that acculturation was negatively associated with seeking shade or wearing protective clothing
when outdoors on sunny days (all p<0.05). In other words, high-acculturated Latinos reported
less use of shade or protective clothing compared to their low-acculturated counterparts.
Researchers have noted that dress customs are generally incorporated faster with increasing
acculturation than are less tangible qualities such as values (Mendoza et al., 1981), which
suggests that acculturation might lead to a relatively quick adoption of US clothing practices.
Also, our findings are consistent with recent research evidence that beach use (i.e., leisure-time
UVR exposure) might be higher among US-born and/or English-speaking Latinos than among
their Mexico-born, Spanish-speaking counterparts, after controlling for SES (Wolch et al., 2004).
The results of Study 1 revealed some support for Hypothesis 3 since statistically
significant effects of SES and age as moderators of the association between acculturation and
sun-safe behaviors were observed. In particular, only among individuals of increased SES, higher
acculturation was associated with less use of long pants when outdoors in the sun. In that study,
SES was accessed by the level of formal education. It is possible that acculturation has a negative
impact on wearing long pants among more versus less educated Latinos because of decreased
endorsement of Latino clothing norms, increased endorsement of US norms regarding tanned
skin, or due to social status-related changes in lifestyle that permit less body coverage when

103
outdoors in the sun. Such effects are consistent with existing knowledge since education level
has been shown to influence sun exposure and sun-safe behaviors of non-Latino whites, possibly
due to specifics of leisure-time behavior patterns (Bränström et al., 2004; Brown et al., 2006). In
turn, only among younger participants, higher acculturation was associated with more sunscreen
use and less use of long-sleeve shirts. Research with diverse populations has in fact revealed that
sun-safe behaviors are positively associated with age (Stanton et al., 2004), conceivably due to
decreased importance of sensual and sociocultural factors (Leary et al., 1993; Baum et al., 1998).
In the present sample, younger Latinos might be more likely to use sunscreen than their older
counterparts because of increased leisure-time sun exposure (a construct that could not be
assessed with HINTS 2005 data).

Study 2
The following hypotheses regarding mediated associations were tested in Study 2:
Hypothesis 4: Perceived health status, level of formal education, healthcare access, and
involvement with social networks regarding health matters would display statistically significant
mediation effects between acculturation and sun-safe behaviors.
Hypothesis 5: Each sun-safe behavior would be mediated by a unique set of potential
mediating constructs.
The results of Study 2 revealed that the association between acculturation and each of the
four potential mediating constructs was statistically significant and positive, indicating that higher
acculturation was linked to better self-perceptions of health, more healthcare access, increased
levels of education, and more contact with friends/family and social organizations regarding
health matters. Also, statistically significant, positive associations were found between education
level, perceptions of health, and involvement with social networks, respectively, and sunscreen

104
use. In turn, a statistically significant, negative association was observed between education level
and wearing protective clothing. The full tests for mediation of sunscreen use supported
Hypothesis 4. Specifically, the direct effect between acculturation and sunscreen use was
attenuated and lost statistical significance, while the indirect effects of acculturation, education
level, health status, and social networks on sunscreen use remained statistically significant. In
addition, the direct path between acculturation and wearing protective clothing as well as the
indirect effects of acculturation and education level were statistically significant, indicating that
the potential mediated effect on that sun-safe behavior was partial.
Among the hypothesized mediating constructs in Study 2, healthcare access was the only
factor that did not display any mediated associations in any of the models. While there is
consistent evidence that acculturation leads to increased healthcare access (Myers et al., 2003;
Mainous et al., 2006), its effects on engagement in health promoting behaviors are somewhat
divergent and outcome-specific. For example, analyses of the 2000 NHIS have shown that in
adjusted models having a physician visit within the previous year and having health insurance
(e.g., the indicators used in Study 2) were associated with a greater likelihood of reporting
sunburn (e.g., lack of sun safety) (Hall et al., 2003). Nonetheless, there is also indication that
having a recent general physical examination was positively associated with sunscreen use and
with reporting a recent skin cancer examination (Santmyire et al., 2001). Healthcare access in
terms of contact with a provider might be a more plausible or proximal mediator for other types
of health/risk behaviors (e.g., cancer screening, smoking cessation) (Abraido-Lanza et al., 2004;
Ranney et al., 2006) than for sun-safe behaviors, which are dictated by a complex set of personal,
sociocultural, and environmental factors (Leary et al., 1993; Clarke et al., 1997; Jackson et al.,
2000; Garbe et al., 2000; Bränström et al., 2004; Thieden et al., 2005). Regarding sun-safe
behaviors, it is possible that healthcare assess might exhibit effects as a moderator of the observed

105
mediated associations. Some researchers have indeed suggested that SES variables (including
access to services) could be potential moderators of health behaviors and outcomes (Kraemer et
al., 2006).
The mediation models in Study 2 revealed some support for Hypothesis 5. Utilizing
SEM techniques, it was observed that the relationship between Latinos’ acculturation and their
sunscreen use displayed evidence for mediation by perceived health status, education level and
involvement with social networks. In that model, the direct effect of acculturation on the
outcome was attenuated and became statistically non-significant, suggesting that the selected set
of mediators had good explanatory potential for sunscreen use. In turn, the model with protective
clothing as the outcome variable displayed a statistically significant mediated association with
education level but not with the other hypothesized mediating constructs. This findings is in
unison with prior research suggesting that adults with a university degree might be more likely to
vacation at seaside resorts (e.g., conceivably have decreased use of protective clothing, decreased
use of shade, and increased leisure-time UVR exposure) than those who had completed only
elementary school (Bränström et al., 2004). In the model with wearing protective clothing,
acculturation retained a statistically significant, direct, negative effect on the outcome. This
finding indicates that there might be other potential mediators omitted from the model. For
example, it is possible that psychosocial constructs such as body image, tanning
attitudes/preferences, and social norms (Bergenmar et al., 2001; Knight et al., 2002; Cafri et al.,
2006; Hymowitz et al., 2006) might hold better explanatory potential regarding use of protective
clothing that the selected set of mediators examined in Study 2. Some of these constructs (e.g.,
social norms, body image) have indeed been shown to undergo changes as a result of
acculturation in Latino samples (Kranau et al., 1982; Cuellar, 1995 #107; Pepper et al., 2007).

106
Finally, no mediated associations were observed in the model with seeking shade as the
outcome variable. In that model, the main effect of acculturation remained statistically
significant and displayed a negative association. Specifically, higher acculturation levels were
related to less use of shade when outdoors for ≥1 hour on a sunny day, after controlling for the
presence of the mediators. As noted earlier, this finding could be attributed to increasing
adoption of US sun exposure practices with increasing acculturation (Wolch et al., 2004). Also,
the acculturation literature has revealed consistent evidence that physical activity levels increase
with acculturation (Perez-Stable et al., 1994; Crespo et al., 2001; Evenson et al., 2004; Slattery et
al., 2006; Pichon et al., 2007). Increased physical activity, however, might lead to decreased
engagement in sun-safe behaviors (Jennings et al., 2008). As with wearing protective clothing,
the lack of mediated associations for seeking shade indicates the possible omission of important
mediating constructs, such as tanning preferences and attitudes.
In developed countries and particularly among non-Latino whites, having a suntan is
generally perceived as attractive and appearance enhancing (Leary et al., 1993; Beasley et al.,
1997; Robinson et al., 1997; Knight et al., 2002; Hymowitz et al., 2006). Across time, the most
common motives for tanning have been those related to self-presentation, such as perceived
improvement in physical attractiveness and appearance, including the appearance of health
(Keesling et al., 1987; Vail-Smith et al., 1993; Leary et al., 1993; Beasley et al., 1997; Hillhouse
et al., 2000; Bergenmar et al., 2001; Knight et al., 2002). Further, findings reveal that white
Latino students were more likely to tan deeply during the preceding year than were white non-
Latino students (Ma et al., 2007). Therefore, research on the psychosocial and sociocultural
mechanisms of sun-safe behaviors among Latinos merits increased attention.

107
Study 3
The following moderated mediation hypotheses were tested in Study 3:
Hypothesis 6: The association between acculturation and involvement with social
networks would be stronger among men than women, while the association between social
networks and sun-safe behaviors would be constant across gender.
Hypothesis 7: The association between acculturation and education level would be
constant across gender, while the association between education level and sun-safe behaviors
would be stronger among women than men.
Hypothesis 8: The association between acculturation and perceived health status and
social networks, respectively, would be stronger among insured compared to uninsured Latinos,
while the association between each of these mediators and sun-safe behaviors would be constant
across health insurance status.
Overall, the findings from Study 3 suggested the presence of moderated mediation of the
relationship between acculturation and sun-safe behaviors. However, no support for Hypothesis 6
was obtained. Specifically, there was no evidence that the mediated effect for social networks
was moderated by gender for any of the outcomes. It is possible that in this sample of Latinos,
acculturation does not have a gender-specific impact on social networks. Some researchers have
suggested that social network dynamics might differ for men and women during the mainstream
integration process (Hagan, 1998). However, in that research, the conceptualization of social
networks did not include health-related features. In contrast, in the present research the social
networks construct was assessed in terms of availability and frequency of contact with
friends/family and social organizations regarding health matters. Therefore, the lack of gender-
specific effects of acculturation on involvement with social networks might be partially attributed
to the conceptualization (e.g., health-related versus occupation-related) of the latter construct.

108
In support of Hypothesis 7, the results of Study 3 revealed that the education level-
mediated association between acculturation and sunscreen use was different for men and women.
As hypothesized, the link between acculturation and education level was constant across gender,
while the link between education level and sunscreen use was stronger among women than men.
This finding is in line with prior research with non-Latino whites. Specifically, there is evidence
that among individuals with the same level of formal education, women were more likely to use
sunscreen (Jerkegren et al., 1999), to have more positive attitudes and intentions towards sun
safety (Hillhouse et al., 1996), but also more likely to use tanning salons (during the preceding
year) (Hillhouse et al., 2000; Danoff-Burg et al., 2006). Also, women in general have been
shown to be more knowledgeable about skin cancer and sun-safe behaviors (Vail-Smith et al.,
1993; Hillhouse et al., 1996; Stanton et al., 2004). Hence, it is possible that women are more
likely than men to attend to sun safety/skin cancer information in order to determine how to tan
“safely.”
Finally, Hypothesis 8 received partial support. In particular, there was evidence that the
effect of social networks as a potential mediator between acculturation and sunscreen use was
moderated by health insurance status. Among insured participants the relationship between
acculturation and involvement with social networks regarding health matters (in terms of
availability and frequency of contact) was stronger compared to their uninsured counterparts. It
is possible that having health insurance gives individuals the opportunity to join provider-based
health promotion groups/networks, thus increasing the salience of the social networks construct.
As hypothesized, the association between social networks and sunscreen use remained constant
across health insurance status.
Regarding wearing protective clothing when outdoors in the sun, the results of Study 3
did not reveal any statistically significant or marginally significant evidence of moderated

109
mediation. The model, however, included only one mediator (education level). As suggested by
prior research, for this sun-safe behavior it might be useful to explore the potential mediating
effects of other psychosocial constructs such as appearance motivation, body image, tanning
attitudes, and social norms (Bergenmar et al., 2001; Knight et al., 2002; Hymowitz et al., 2006;
Cafri et al., 2006). In addition, country of origin/ethnicity merits attention as a potential
moderator (Ramirez et al., 2004; Perez-Escamilla et al., 2007) of the association between
acculturation and wearing protective clothing. It has been noted that specific features of different
Latino cultures (e.g., values, norms) might have independent effects on health behaviors and
outcomes (Khan et al., 1997). These variables, however, were not available in the HINTS 2005
dataset.

Limitations and strengths of the present research
Limitations
This research utilized roughly 9% of the HINTS 2005 sample. The full survey sample
was predominantly (>73%) non-Latino whites. In order to improve representativeness relative to
national estimates (U.S. Census Bureau, 2008), HINTS 2007 (data collected in 2008) will over-
sample Latinos. Hence, future research could compare the updated estimates with those obtained
with HINTS 2005 in order to assess their external generalizability. Representativeness concerns,
however, are somewhat mitigated because many of the demographic characteristics of this sample
paralleled those in nationally-representative health behavior research. For example, weighted
prevalence estimates from the 2000 NHIS (N=4,558 Latino adults) showed that 45% of the
participants had less than high school education (versus 42% reported here) and 38% were US-
born (versus 36% reported here) (Berrigan et al., 2006). In addition, current census information

110
reveals that 33% of US Latinos do not have health insurance (De Navas-Walt et al., 2007) which
parallels the prevalence estimate in this research.
Nonetheless, some selection bias related to sampling Latinos by landline telephones,
social desirability effects, and non-response might have occurred. As in other telephone surveys,
the HINTS response rate was low (Davis et al., 2005; Rutten et al., 2007). Regarding non-
coverage, estimates suggest that during the period January 2004 - June 2005 (i.e., roughly the
time period when the HINTS 2005 data were collected), 7.2% of the Latino households had only
wireless telephones and 3.2% had no telephone service (Blumberg et al., 2006). Research has
shown that individuals without landline telephones might be more likely to be uninsured, and that
wireless telephone substitution was greater among those living in poverty (Blumberg et al., 2006).
Thus, the findings of the present research might be more generalizable to relatively more
acculturated Latinos who have been in the US longer and have greater resources. A selection bias
in the direction of non-coverage of the least acculturated Latinos might have occurred and the
results might underestimate true effects of acculturation on sun-safe behaviors.
Further, non-response as well as psychometric analyses for HINTS 2005 have been
planned but are not currently available (Rutten et al., 2007). It is possible that some of the
measures used in this study might have poor reliability coefficients, which could weaken
associations and account for some of the null effects (e.g., in the moderated mediation model for
wearing protective clothing). However, this problem could not be assessed at present due to the
lack of reliability values.
Several limitations pertain specifically to the acculturation measure used in this research.
First, acculturation implies a linear time sequence despite the use of cross-sectional data. The
process of acculturation reflects changes that occur as a function of time, among other factors
(e.g., degree of exposure, context) (Berry, 1980; Mendoza, 1984; LaFromboise et al., 1993).

111
Since the present research was concerned with the impact of acculturation on health behaviors in
the sun, rather than with the actual process of acculturation, the construct was treated as a fixed,
antecedent variable (Flores et al., 1995). Second, the HINTS data allowed for the assessment of
acculturation on a unidirectional scale, but a bidirectional measure - with Latino and US
orientations as separate scores - would have been the measure of choice had the data been
available. Third, in computing the acculturation index the mean of all available acculturation-
related variables for all available participants was utilized. In EFA, dichotomous variables do not
perform as well as multiple-category or continuous variables, and this might explain why nativity
did not load onto the acculturation factor and was excluded from the index. For the foreign-born
participants, the mean acculturation score included interview language, comfort with English,
duration of US residence, and age at US arrival; for their US-born counterparts, the corresponding
score included language of interview and comfort with English. Language items, however,
generally explain most of the variance in acculturation scales (Marin, 1992; Epstein et al., 1996).
Another data limitation pertains to the inability to distinguish among Latino subgroups.
Yet, out of various risk factors explored among Latinos (e.g., lack of sun safety, smoking,
drinking, poor diet, insufficient exercise), the lack of sun safety has shown the least amount of
variation across countries of origin (Ramirez et al., 2004). This might be due to Latinos’
decreased awareness of sun safety compared to other health practices. Nonetheless, Latino
ethnicity is suggested as a potential moderator for future mediation studies of sun-safe behaviors.
Research has also noted a potential confound (which could not be assessed with HINTS data)
between acculturation and skin color, such that the latter might impact one’s motivation or ability
to acculturate due to perceived socioeconomic discrimination against dark-skin individuals
(Gomez, 2000; Murguia et al., 2002). However, this does not imply that acculturation causes
people to discontinue engaging in sun-safe behaviors.

112
Further, the HINTS 2005 did not include occupation data. The employment categories
available in the survey were as follows: employed for wages, self-employed, out of work for ≥1
year, out of work for <1 year, a homemaker, a student, retired, and unable to work. Occupational
sun exposure, however, could be an important factor in Latinos’ sun-safe behaviors. For
example, in a sample of Latino farmworkers in California, only 1.5% reported always using
sunscreen (versus 15% in the present research), whereas 83% reported always/often wearing
long-sleeve shirts (versus 24% in the present research) (Salas et al., 2005). Also, many outdoor
Latino workers exhibit low acculturation (Salas et al., 2005). Thus, the exclusion of occupation
data is acknowledged as a limitation of this research and its inclusion is suggested for future sun
safety studies with Latinos.
Several limitations particular to Studies 2 and 3 must also be noted. Since causality could
not be fully understood with cross-sectional data, it is possible that education, which was modeled
as a potential mediator, might precede and/or cause acculturation. Specifically, education may
have occurred before coming to the US or, among the foreign-born Latinos, education could
affect the degree to which they are able to acculturate. While there is strong evidence that
acculturation and education are positively associated (Alarcon et al., 1999; Franzini et al., 2004;
Berrigan et al., 2006; Mainous et al., 2006; Pichon et al., 2007; Barcenas et al., 2007), and the
data were consistent with the hypothesized models, it is important for future research to re-assess
these models with longitudinal data. Cross-sectional datasets have been used in a number of
mediation studies in the psychosocial literature, including studies on acculturation (Unger et al.,
2000; Pichon et al., 2007) and sun safety (Turner et al., 2005; Azzarello et al., 2006; Cafri et al.,
2006), and the main criticism again pertains to causality inferences (MacKinnon et al., 2007).
Researchers have observed that when all constructs were assessed at the same time, there might
be other models that could explain the data (MacKinnon et al., 2007). Whereas experimental

113
studies have been suggested as the best way to validate causal mechanisms (Kraemer et al., 2002;
Shrout et al., 2002; MacKinnon et al., 2007), in the case of acculturation it would not be possible
to assign participants to acculturation levels. Also, the aims of Studies 2 and 3 were to advance
the acculturation and sun safety literatures by suggesting potential mediated associations, and not
to prove causality. Indeed, the results of Study 2 revealed evidence that some mediated paths are
more plausible than others (Shrout et al., 2002). It has been noted that correlation analyses are
valuable for theoretical and empirical rationale development and for indicating possible points of
intervention (Shrout et al., 2002).
Finally, in addition to the three-step linear regression approach used in Study 3, other
methods for assessing moderated mediation also exist, such as the piecemeal approach, the
subgroup approach, and moderated path analysis (Edwards et al., 2007). While each method has
certain deficiencies (Edwards et al., 2007), Muller and colleagues’ (2005) method appeared to be
fairly flexible and very well defined, and has few known important limitations. In addition, it
appeared to be the most popular moderated mediation approach in the psychology literature
according to a recent review (Edwards et al., 2007). Nonetheless, some limitations of the chosen
method pertain to the possibility that the effect of the moderator might be in the opposite
direction than the direct and mediated effects, or that the moderator might have non-linear effects
(Edwards et al., 2007). It is possible that the lack of strong statistical significance in the models
might be due to the presence of non-linear effects or to the relatively small sample size for this
type of analyses. Future research ought to replicate these findings with large samples of Latinos
and with longitudinal data.
Also, in the models in Study 3 there might be a certain degree of violation of one of the
assumptions for moderated mediation analyses. Specifically, the moderator and predictor are
assumed to be independent (Muller et al., 2005), however, in the present sample acculturation and

114
health insurance had a Pearson correlation coefficient=0.45. In order to access the potential
collinearity between these two constructs, their tolerance values were computed. Generally,
tolerance values range between 0 and 1, with lower values indicating increasing collinearity. In
these models, however, the tolerance values were relatively high (0.79). The conceptual
distinction between acculturation and health insurance, as well as the fact that their correlation
was not very high, suggested that the lack of complete empirical independence might not pose a
serious threat to the overall results. Nonetheless, the findings might be affected to a certain
degree by decreasing the straightforward interpretation of the regression slopes for the two
variables, as well as compromising the potential for establishing statistical significance.

Strengths of the present research
This research was guided by empirical evidence from the acculturation and sun safety
literatures as well as a testable theoretical framework. It utilized data from a national survey and
represented the first effort to connect the acculturation and sun safety literatures. An advantage
of studying acculturation-related issues among Latinos is entailed in the existence of a wide range
of acculturation levels among them (Perez-Escamilla et al., 2007). By asking novel questions
pertaining to the role of acculturation in sun-safe behaviors, and by employing a comprehensive
assessment of direct, mediated, and moderated associations, the findings of this research could
better inform public health research and education efforts targeted at Latinos. This research
addressed health practices of the largest and fastest growing minority population in the US,
whose health behaviors and outcomes will continue to have an increasing socioeconomic impact
on the US public health system.
A specific strength of Study 2 is the use of SEM, which allows flexibility in specifying
theory-driven models (Buhi et al., 2007) and represents a confirmatory approach to multivariate

115
analysis (Byrne, 1994). Unlike traditional regression analysis, SEM affords the opportunity to
study both observed and latent constructs and thus better interpret the relationships among them,
simultaneously assess direct and indirect effects, and give unbiased estimates by explicitly
modeling measurement error (i.e., by modeling unique and common variance separately) (Byrne,
1994; Cole et al., 2003). In addition, advantages of SEM over the popular causal steps model by
Barron and Kenny (1986) are that it provides direct hypothesis tests for mediation, is easily
adaptable to multiple mediators, and provides greater statistical power (MacKinnon et al., 2002;
Dearing et al., 2006).
Finally, the acculturation index used in this research included items that are considered
proxies of acculturation and are often employed in acculturation studies (Crespo et al., 2001;
Abraido-Lanza et al., 2005; Berrigan et al., 2006; Moran et al., 2007). Utilizing an index instead
of a dichotomous measure is consistent with modern conceptualizations of acculturation as a
process and not a dichotomy.

Overall conclusion and implications for public health research and education
Sun safety education initiatives often produce only modest behavioral effects among non-
Latino whites (Mahler et al., 1997; Baum et al., 1998; Weinstock et al., 2002) and it might be
even harder to reach Latinos due to their lower awareness about risk factors or symptoms (Olsen
et al., 1993; Pipitone et al., 2002; Byrd-Miles et al., 2007). The findings of this research suggest
that potentially successful efforts need to focus on low- and high-acculturated Latinos separately
and be age- and SES-targeted. Considering evidence from the literature as well as factors
associated with low acculturation (e.g., being monolingual, lack of healthcare coverage, limited
exposure to media-based health information, familism, manual occupations) (Ruiz et al., 1992;
Cuellar et al., 1995; Marquez et al., 2006; Mainous et al., 2006), sun safety messages for low-

116
acculturated Latinos could utilize informal, Spanish-language strategies, such as posting or
distributing leaflets advocating use of shading structures, sunscreen, and protective clothing in
bus stations, churches and community centers. Such messages could also emphasize the
importance of having a healthy family and could be incorporated in interactive activities at Latino
cultural events. The findings also echo recommendations for outdoor worker protection (e.g.,
discussions about UVR, eye-catching sun safety posters, supervisor training) (International
Commission on Non-Ionizing Radiation Protection, 2007) since many outdoor Latino workers
exhibit low acculturation (Salas et al., 2005).
Overall, education efforts with low-acculturated Latinos could be resource-efficient since
the aim would be to reinforce existing sun-safe behaviors. In contrast, the findings of this
research indicate that education efforts with high-acculturated Latinos might require more
resources due to the need for behavior modification. Such initiatives could use the English
language, emphasize modifying individual (versus family) sun-safe behaviors, and could be
delivered in health clinics (due to increased healthcare access) and through the media, similar to
initiatives targeted at the whole US population. Indeed, research with Latinos has documented
that language acculturation could predict media exposure after controlling for demographic
variables; media exposure, in turn could predict cancer screening and symptom knowledge after
controlling for language acculturation and demographic factors (Ruiz et al., 1992).
Regarding specific sun-safe behaviors, the results of the present research suggest that
interventions focused on increasing sunscreen use, for example, might need to account for the
individuals’ self-perceptions of health, their education levels, and their involvement in health-
related organizations. In addition, such interventions might have better potential for success of
they are gender- and education-level targeted, since women might be more likely than men to use
sun safety/skin cancer information in order to determine how to tan “safely.”

117
The mediation and moderated mediation analyses in the present research have descriptive
importance and could further inform future sun safety intervention development as well as
longitudinal studies on mediation. Specifically, the findings indicated that some mediated paths
(e.g., between acculturation, education level, and wearing protective clothing) are more plausible
than others (e.g., between acculturation, perceived health status and seeking shade). It has been
noted that correlation analyses are valuable for theoretical and empirical rationale development
and for indicating possible points of intervention (Shrout et al., 2002). Hence, the results of the
present mediation analyses could serve to inform future studies on causal mechanisms, employing
longitudinal data and large samples of Latinos and other acculturating populations. Further, the
findings suggest potential mediating constructs (e.g., education level, social networks) and the
populations (e.g., women, insured) for which such mediators might have increased impact. Also,
future research could assess the potential mediating effects of other psychosocial constructs,
which were not available in HINTS, such as appearance motivation, body image, tanning
attitudes, and social norms on wearing protective clothing and seeking shade.
Whereas much of the sun safety literature pertains to non-Latino whites who exhibit the
highest skin cancer incidence (American Cancer Society, 2008), the changing demographics of
the US population and particularly the rapidly growing Latino segment necessitate the
understanding of the etiology and epidemiology of dermatologic disorders in non-Caucasian skin
types (Halder et al., 2003). Over 45 million Latinos live in the US (15% of the population),
primarily in areas with high UVR, such as California and Texas (U.S. Census Bureau, 2008).
Hence, the public health impact of Latinos’ health behaviors in general and their cancer-related
outcomes in particular is likely to intensify (Vega et al., 1994; Amaro et al., 2002; Hu et al.,
2004). Despite data limitations, the present research demonstrated that acculturation has an
adverse effect on Latinos’ skin cancer risk-related behaviors. Appropriate interventions with this

118
population could improve engagement in sun-safe behaviors as well as knowledge about skin
cancer signs and symptoms, and could result in decreased rates of advanced melanoma.

119
References
Aase A, Bentham G. (1994). The geography of malignant melanoma in the Nordic countries: the
implications of stratospheric ozone depletion. Geografiska Annaler: Series B 76(2):129-
139.

Abraido-Lanza AF, Chao MT, Gammon MD. (2004). Breast and cervical cancer screening among
Latinas and non-Latina whites. American Journal of Public Health 94(8):1393-1398.

Abraido-Lanza AF, Chao MT, Florez KR. (2005). Do healthy behaviors decline with greater
acculturation? Implications for the Latino mortality paradox. Social Science & Medicine
61(6):1243-1255.

Abraido-Lanza AF, Armbrister AN, Florez KR, et al. (2006). Toward a theory-driven model of
acculturation in public health research. American Journal of Public Health 96(8):1342-
1346.

Ajzen I, Fishbein M. (1980). Understanding attitudes and predicting behavior. Englewood Cliffs,
NJ, Prentice-Hall.

Alaniz ML, Treno AJ, Saltz RF. (1999). Gender, acculturation, and alcohol consumption among
Mexican Americans. Substance Use & Misuse 34(10):1407-1426.

Alarcon GS, Rodriguez JL, Benavides Jr. G, et al. (1999). Systemic lupus erythematosus in three
ethnic groups. V. Acculturation, health-related attitudes and behaviors, and disease
activity in Hispanic patients from the LUMINA cohort. LUMINA study group. Lupus in
minority populations, nature versus nurture. Arthritis Care & Research 12(4):267-276.

Amaro H, de la Torre A. (2002). Public health needs and scientific opportunities in research on
Latinas. American Journal of Public Health 92(4):525-529.

American Academy of Dermatology. (2008). Be sun smart. Accessed on May 28, 2008 from:
http://www.aad.org/public/sun/smart.html.

American Cancer Society. (2008). Cancer Facts & Figures, 2008. Atlanta, GA, American Cancer
Society.

Andreeva VA, Cockburn MG, Yaroch AL, et al. (Manuscript in preparation). Acculturation and
sun-safe behaviors among Latinos: preliminary evidence for mediation by health status,
education level, and involvement with social networks.

Andreeva VA, Unger JB, Yaroch AL, et al. (In review). Acculturation and sun-safe behaviors
among US Latinos: findings from the 2005 Health Information National Trends Survey.
American Journal of Public Health.

120
Anez LM, Paris Jr. M, Bedregal LE, et al. (2005). Application of cultural constructs in the care of
first generation Latino clients in a community mental health setting. Journal of
Psychiatric Practice 11(4):221-230.

Arcia E, Skinner M, Bailey D, et al. (2001). Models of acculturation and health behaviors among
Latino immigrants to the US. Social Science & Medicine 53(1):41-53.

Atillasoy ES. (2003). Chapter 216. Skin cancers. The Merck manual of medical information.
Beers MH, Fletcher AJ, Jones TV, Porter R, Berkwits M, Kaplan JL, eds. New York,
Pocket Books.

Autier P, Boniol M, Dore J-F. (2007). Sunscreen use and increased duration of intentional sun
exposure: still a burning issue. International Journal of Cancer 121(1):105.

Azzarello LM, Dessureault S, Jacobsen PB. (2006). Sun-protective behavior among individuals
with a family history of melanoma. Cancer Epidemiology, Biomarkers and Prevention
15(1):142-145.

Bandura A. (1986). Social foundations of thought and action: A Social Cognitive Theory.
Englewood Cliffs, NJ, Prentice Hall.

Barcenas CH, Wilkinson AV, Strom SS, et al. (2007). Birthplace, years of residence in the United
States, and obesity among Mexican-American adults. Obesity 15(4):1043-1052.

Baron RM, Kenny DA. (1986). The moderator-mediator variable distinction in social
psychological research: conceptual, strategic, and statistical considerations. Journal of
Personality and Social Psychology 51(6):1173-1182.

Bauer DJ, Preacher KJ, Gil KM. (2006). Conceptualizing and testing random indirect effects and
moderated mediation in multilevel models: new procedures and recommendations.
Psychological Methods 11(2):142-163.

Baum A, Cohen L. (1998). Successful behavioral interventions to prevent cancer: the example of
skin cancer. Annual Review of Public Health 19:319-33.

Beasley TM, Kittel BS. (1997). Factors that influence health risk behaviors among tanning salon
patrons. Evaluation & The Health Professions 20(4):371-388.

Becker MH. (1974). The Health Belief Model and personal health behavior. Health Education
Monographs 2(4):234-273.

Bentler PM. (2004). EQS 6.1 structural equations program manual. Encino, CA, Multivariate
Software, Inc.

Bergenmar M, Brandberg Y. (2001). Sunbathing and sun-protection behaviors and attitudes of
young Swedish adults with hereditary risk for malignant melanoma. Cancer Nursing
24(5):341-350.

121
Berrigan D, Dodd K, Troiano RP, et al. (2006). Physical activity and acculturation among adult
Hispanics in the United States. Research Quarterly for Exercise and Sports 77(2):147-
157.

Berry JW. (1980). Acculturation as varieties of adaptation. Acculturation: theory, models and
some new findings (pp. 9-25). Padilla AM, ed. Boulder, CO, Westview Press.

Berwick M, Wiggins C. (2006). The current epidemiology of cutaneous malignant melanoma.
Frontiers in Bioscience 11:1244-1254.

Bethel JW, Schenker MB. (2005). Acculturation and smoking patterns among Hispanics: a
review. American Journal of Preventive Medicine 29(2):143-148.

Bigler C, Feldman J, Hall E, et al. (1996). Pigmented basal cell carcinoma in Hispanics. Journal
of the American Academy of Dermatology 34(5 Pt 1):751-752.

Black WC, Goldhahn Jr. RT, Wiggins C. (1987). Melanoma within a southwestern Hispanic
population. Archives of Dermatology 123(10):1331-1334.

Blumberg SJ, Luke JV, Cynamon ML. (2006). Telephone coverage and health survey estimates:
evaluating the need for concern about wireless substitution. American Journal of Public
Health 96(5):926-931.

Boggild AK, From L. (2003). Barriers to sun safety in a Canadian outpatient population. Journal
of Cutaneous Medicine and Surgery 7(4):292-299.

Borrayo EA, Jenkins SR. (2003). Feeling frugal: socioeconomic status, acculturation, and cultural
health beliefs among women of Mexican descent. Cultural Diversity and Ethnic Minority
Psychology 9(2):197-206.

Borrell LN. (2005). Racial identity among Hispanics: implications for health and well-being.
American Journal of Public Health 95(3):379-381.

Bränström R, Ullén H, Brandberg Y. (2004). Attitudes, subjective norms and perception of
behavioural control as predictors of sun-related behaviour of Swedish adults. Preventive
Medicine 39(5):992-999.

Bränström R, Kristjansson S, Ullén H. (2006). Risk perception, optimistic bias, and readiness to
change sun related behaviour. European Journal of Public Health 16(5):492-497.

Brown TT, Quain RD, Troxel AB, et al. (2006). The epidemiology of sunburn in the US
population in 2003. Journal of the American Academy of Dermatology 55(4):577-583.

Buhi ER, Goodson P, Neilands TB. (2007). Structural equation modeling: a primer for health
behavior researchers. American Journal of Health Behavior 31(1):74-85.

122
Buriel R. (1984). Chapter 7. Integration with traditional Mexican-American culture and
sociocultural adjustment. Chicano psychology, 2nd ed,( pp. 95-130). Martinez JL,
Mendoza RH, eds. Orlando, FL, Academic Press, Inc.

Burnam MA, Hough RL, Karno M, et al. (1987). Acculturation and lifetime prevalence of
psychiatric disorders among Mexican Americans in Los Angeles. Journal of Health and
Social Behavior 28(1):89-102.

Byrd-Miles K, Toombs EL, Peck GL. (2007). Skin cancer in individuals of African, Asian, Latin
American, and American Indian descent: differences in incidence, clinical presentation,
and survival compared to Caucasians. Journal of Drugs in Dermatology 6(1):10-16.

Byrne BM. (1994). Structural equation modeling with EQS and EQS/Windows. Basic concepts,
applications, and programming. Thousand Oaks, CA, Sage Publications, Inc.

Caetano R, Ramisetty-Mikler S, Wallisch LS, et al. (2008). Acculturation, drinking, and alcohol
abuse and dependence among Hispanics in the Texas-Mexico border. Alcoholism:
Clinical & Experimental Research 32(2):314-321.

Cafri G, Thompson JK, Jacobsen PB. (2006). Appearance reasons for tanning mediate the
relationship between media influence and UV exposure and sun protection. Archives of
Dermatology 142(8):1067-1069.

Cairns CP, Viswanath K. (2006). Communication and colorectal cancer screening among the
uninsured: data from the Health Information National Trends Survey (United States).
Cancer Causes & Control 17(9):1115-1125.

Cantero PJ, Richardson JL, Baezconde-Garbanati L, et al. (1999). The association between
acculturation and health practices among middle-aged and elderly Latinas. Ethnicity &
Disease 9(2):166-180.

Castanedo-Cazares JP, Torres-Alvarez B, Medellín-Pérez ME, et al. (2006). Knowledge and
attitudes of Mexican population concerning solar radiation [Article in Spanish]. Gaceta
Médica de México 142(6):451-455.

Chesney AP, Chavira JA, Hall RP, et al. (1982). Barriers to medical care of Mexican-Americans:
the role of social class, acculturation, and social isolation. Medical Care 20(9):883-891.

Cho Y, Frisbie WP, Rogers RG. (2004). Nativity, duration of residence, and the health of
Hispanic adults in the United States. International Migration Review 38(1):184-211.

Christos PJ, Oliveria SA, Masse LC, et al. (2004). Skin cancer prevention and detection by
nurses: attitudes, perceptions, and barriers. Journal of Cancer Education 19(1):50-57.

Clarke VA, Williams T, Arthey S. (1997). Skin type and optimistic bias in relation to the sun
protection and suntanning behaviors of young adults. Journal of Behavioral Medicine
20(2):207-222.

123
Clement R. (1986). Second language proficiency and acculturation: an investigation of the effects
of language status and individual characteristics. Journal of Language and Social
Psychology 5(4):271-290.

Clement R, Baker SC, MacIntyre PD. (2003). Willingness to communicate in a second language.
The effects of context, norms, and vitality. Journal of Language and Social Psychology
22(2):190-209.

Cockburn MG, Zadnick J, Deapen D. (2006). Developing epidemic of melanoma in the Hispanic
population of California. Cancer 106(5):1162-1168.

Cohen J. (1990). Things I have learned (so far). American Psychologist 45(12):1304-1312.

Cole DA, Maxwell SE. (2003). Testing mediational models with longitudinal data: questions and
tips in the use of structural equation modeling. Journal of Abnormal Psychology
112(4):558-577.

Comas-Diaz L. (2001). Hispanics, Latinos, or Americanos: the evolution of identity. Cultural
Diversity and Ethnic Minority Psychology 7(2):115-120.

Cormier JN, Xing Y, Ding M, et al. (2006). Ethnic differences among patients with cutaneous
melanoma. Archives of Dermatology 166(17):1907-1914.

Coups EJ, Ostroff JS. (2005). A population-based estimate of the prevalence of behavioral risk
factors among adult cancer survivors and noncancer controls. Preventive Medicine
40(6):702-711.

Coups EJ, Manne SL, Heckman CJ. (2008). Multiple skin cancer risk behaviors in the U.S.
population. American Journal of Preventive Medicine 34(2):87-93.

Crespo CJ, Smit E, Carter-Pokras O, et al. (2001). Acculturation and leisure-time physical
inactivity in Mexican American adults: results from NHANES III, 1988-1994. American
Journal of Public Health 91(8):1254-1257.

Cress RD, Holly EA. (1997). Incidence of cutaneous melanoma among non-Hispanic whites,
Hispanics, Asians, and blacks: an analysis of California Cancer Registry data, 1988-1993.
Cancer Causes & Control 8(2):246-252.

Crimmins EM, Soldo BJ, Kim JK, et al. (2005). Using anthropometric indicators for Mexicans in
the United States and Mexico to understand the selection of migrants and the "Hispanic
paradox." Social Biology 52(3/4):164-177.

Crimmins EM, Kim JK, Alley DE, et al. (2007). Hispanic paradox in biologic risk profiles.
American Journal of Public Health 97(7):1305-1310.

124
Cuellar I, Arnold B, Gonzalez G. (1995). Cognitive referents of acculturation: assessment of
cultural constructs in Mexican Americans. Journal of Community Psychology 23(4):339-
356.

Danoff-Burg S, Mosher CE. (2006). Predictors of tanning salon use. Behavioral alternatives for
enhancing appearance, relaxing and socializing. Journal of Health Psychology 11(3):511-
518.

Davis T, Park I, Covell J, et al. (2005). Health Information National Trends Survey (HINTS)
2005: Final report. Rockville, MD, National Cancer Institute.

De Navas-Walt C, Proctor BD, Lee CH. (2006). Income, poverty, and health insurance coverage
in the United States: 2005. U.S. Census Bureau, Current Population Reports, P60-231.
Washington, DC, U.S. Government Printing Office.

De Navas-Walt C, Proctor BD, Smith J. (2007). Income, poverty, and health insurance coverage
in the United States: 2006. U.S. Census Bureau, Current Population Reports.
Washington, DC, U.S. Government Printing Office.

Dearing E, Hamilton LC. (2006). V. Contemporary advances and classic advice for analyzing
mediating and moderating variables. Monographs of the Society for Research in Child
Development 71(3):88-103.

Derose KP, Baker DW. (2000). Limited English proficiency and Latinos' use of physician
services. Medical Care Research & Review 57(1):76-91.

Diffey BL. (1998). Ultraviolet radiation and human health. Clinics in Dermatology 16(1):83-89.

Dornelles S, Goldim J, Cestari T. (2004). Determination of the minimal erythema dose and
colorimetric measurements as indicators of skin sensitivity to UV-B radiation.
Photochemistry & Photobiology 79(6):540-544.

Douglass HM, McGee R, Williams S. (1997). Sun behaviour and perceptions of risk for
melanoma among 21-year-old New Zealanders. Australian and New Zealand Journal of
Public Health 21(3):329-334.

Edwards JR, Lambert LS. (2007). Methods for integrating moderation and mediation: a general
analytical framework using moderated path analysis. Psychological Methods 12(1):1-22.

Eide MJ, Weinstock MA. (2005). Association of UV index, latitude and melanoma incidence in
nonwhite populations - US Surveillance, Epidemiology, and End Results (SEER)
program, 1992-2001. Archives of Dermatology 141(4):477-481.

Elo IT, Turra CM, Kestenbaum B, et al. (2004). Mortality among elderly Hispanics in the United
States: past evidence and new results. Demography 41(1):109-128.

125
Epstein JA, Botvin GJ, Dusenbury L, et al. (1996). Validation of an acculturation measure for
Hispanic adolescents. Psychological Reports 79(3 Pt.1):1075-1079.

Espenshade TJ, Fu H. (1997). An analysis of English-language proficiency among U.S.
immigrants. American Sociological Review 62(April):288-305.

Evenson KR, Sarmiento OL, Ayala GX. (2004). Acculturation and physical activity among North
Carolina Latina immigrants. Social Science & Medicine 59(12):2509-2522.

Fan X, Wang L, Thompson B. (1997). Effects of data nonnormality and other factors on fit
indices and parameter estimates for true and misspecified SEM models. Paper presented
at the American Educational Research Association Annual Meeting, Chicago, IL, March
25.

Farabee D, Wallisch L, Maxwell JC. (1995). Substance use among Texas Hispanics and non-
Hispanics: who's using, who's not, and why. Hispanic Journal of Behavioral Sciences
17(4):523-536.

Fears TR, Guerry IV D, Pfeiffer RM, et al. (2006). Identifying individuals at high risk of
melanoma: a practical predictor of absolute risk. Journal of Clinical Oncology
24(22):3590-3595.

Federal Register. (1997). Revisions to the standards for the classification of federal data on race
and ethnicity. Washington, DC, Office of Management and Budget.

Fitzpatrick TB. (1988). The validity and practicality of sun-reactive skin types I through VI.
Archives of Dermatology 124(6):869-871.

Flannery WP, Reise SP, Yu J. (2001). An empirical comparison of acculturation models.
Personality and Social Psychology Bulletin 27(8):1035-1045.

Flores ET, Castro FG, Fernandez-Esquer ME. (1995). Social theory, social action, and
intervention research: implications for cancer prevention among Latinos. Journal of the
National Cancer Institute Monographs 18:101-108.

Franzini L, Ribble JC, Keddie AM. (2001). Understanding the Hispanic paradox. Ethnicity &
Disease 11(3):496-518.

Franzini L, Fernandez-Esquer ME. (2004). Socioeconomic, cultural, and personal influences on
health outcomes in low income Mexican-origin individuals in Texas. Social Science &
Medicine 59(8):1629-1646.

Friedman LC, Bruce S, Weinberg AD, et al. (1994). Early detection of skin cancer: racial/ethnic
differences in behaviors and attitudes. Journal of Cancer Education 9(2):105-110.

Gandini S, Sera F, Cattaruzza MS, et al. (2005). Meta-analysis of risk factors for cutaneous
melanoma: II. Sun exposure. European Journal of Cancer 41(1):45-60.

126
Garbe C, Buettner PG. (2000). Predictors of the use of sunscreen in dermatological patients in
Central Europe. Preventive Medicine 31(2 Pt 1):134-139.

Glanz K, Lew RA, Song V, et al. (1999). Factors associated with skin cancer prevention practices
in a multiethnic population. Health Education & Behavior 26(3):344-359.

Glanz K, Yaroch AL, Dancel M, et al. (2008). Measures of sun exposure and sun protection
practices for behavioral and epidemiological research. Archives of Dermatology
144(2):217-222.

Gloster Jr. HM, Neal K. (2006). Skin cancer in skin of color. Journal of the American Academy of
Dermatology 55(5):741-760.

Godlee F. (1991). Dangers of ozone depletion. British Medical Journal 303(6813):1326-1328.

Goel MS, McCarthy EP, Phillips RS, et al. (2004). Obesity among US immigrant subgroups by
duration of residence. Journal of the American Medical Association 292(23):2860-2867.

Gombeski Jr WR, Ramirez AG, Kautz JA, et al. (1982). Communicating health information to
urban Mexican Americans: sources of health information. Health Education Quarterly
9(4):293-309.

Gomez C. (2000). The continual significance of skin color: an exploratory study of Latinos in the
Northeast. Hispanic Journal of Behavioral Sciences 22(1):94-103.

Gonzalez Cervantes JG, Mora Tiscareno A, Beltran Ortega A. (1990). 261 cases of cutaneous
malignant melanoma, general characteristics and prognostic factor values [Article in
Spanish]. Revista del Instituto Nacional de Cancerologia (Mexico) 36(3):1103-1112.

Gordon MM. (1964). Assimilation in American life. The rise of race, religion, and national
origins. New York, Oxford University Press.

Greene K, Brinn LS. (2003). Messages influencing college women's tanning bed use: statistical
versus narrative evidence format and a self-assessment to increase perceived
susceptibility. Journal of Health Communication 8(5):443-461.

Gregory-Mercado KY, Staten LK, Ranger-Moore J, et al. (2006). Fruit and vegetable
consumption of older Mexican-American women is associated with their acculturation
level. Ethnicity & Disease 16(1):89-95.

Griffith J, Villavicencio S. (1985). Relationships among acculturation, sociodemographic
characteristics and social supports in Mexican American adults. Hispanic Journal of
Behavioral Sciences 7(1):75-92.

Guendelman S, Abrams B. (1995). Dietary intake among Mexican-American women:
generational differences and a comparison with white non-Hispanic women. American
Journal of Public Health 85(1):20-25.

127
Hagan JM. (1998). Social networks, gender, and immigrant incorporation: resources and
constraints. American Sociological Review 63(1):55-67.

Halder RM, Nootheti PK. (2003). Ethnic skin disorders overview. Journal of the American
Academy of Dermatology 48(6 Suppl):S143-S148.

Hall HI, May DS, Lew RA, et al. (1997). Sun protection behaviors of the U.S. white population.
Preventive Medicine 26(4):401-407.

Hall HI, Saraiya M, Thompson T, et al. (2003). Correlates of sunburn experiences among US
adults: results of the 2000 National Health Interview Survey. Public Health Reports
118(6):540-549.

Halpern AC, Altman JF. (1999). Genetic predisposition to skin cancer. Current Opinions in
Oncology 11(2):132-138.

Hayes Bautista DE, Chapa J. (1987). Latino terminology: conceptual bases for standardized
terminology. American Journal of Public Health 77(1):61-68.

Hillhouse JJ, Stair III AW, Adler CM. (1996). Predictors of sunbathing and sunscreen use in
college undergraduates. Journal of Behavioral Medicine 19(6):543-561.

Hillhouse JJ, Turrisi R, Kastner M. (2000). Modeling tanning salon behavioral tendencies using
appearance motivation, self-monitoring, and the Theory of Planned Behavior. Health
Education Research 15(4):405-414.

Himmelgreen DA, Perez-Escamilla R, Martinez D, et al. (2004). The longer you stay, the bigger
you get: length of time and language use in the U.S. are associated with obesity in Puerto
Rican women. American Journal of Physical Anthropology 125(1):90-96.

Hoegh HJ, Davis BD, Manthe AF. (1999). Sun avoidance practices among non-Hispanic white
Californians. Health Education & Behavior 26(3):360-368.

Housman TS, Feldman SR, Williford PM, et al. (2003). Skin cancer is among the most costly of
all cancers to treat for the Medicare population. Journal of the American Academy of
Dermatology 48(3):425-429.

Hu L, Bentler PM. (1995). Evaluating model fit. Structural equation modeling: concepts, issues,
and applications (pp.76-99). Hoyle RH, ed. Thousand Oaks, CA, Sage Publications, Inc.

Hu L, Bentler PM. (1999). Cut-off criteria for fit-indexes in covariance structure analysis:
conventional criteria versus new alternatives. Structural Equation Modeling 6:1-55.

Hu S, Ma F, Collado-Mesa F, et al. (2004). UV radiation, latitude, and melanoma in US
Hispanics and blacks. Archives of Dermatology 140(7):819-824.

128
Hu S, Soza-Vento RM, Parker DF, et al. (2006). Comparison of stage at diagnosis of melanoma
among Hispanic, black, and white patients in Miami-Dade County, Florida. Archives of
Dermatology 142(6):704-708.

Hunt KJ, Resendez RG, Williams K, et al. (2003). All-cause and cardiovascular mortality among
Mexican-American and non-Hispanic white older participants in the San Antonio Heart
Study - evidence against the "Hispanic paradox". American Journal of Epidemiology
158(11):1048-1057.

Hunt LM, Schneider S, Comer B. (2004). Should "acculturation" be a variable in health research?
A critical review of research on US Hispanics. Social Science & Medicine 59(5):973-986.

Hutchison R. (1987). Ethnicity and urban recreation: whites, blacks, and Hispanics in Chicago's
public parks. Journal of Leisure Research 19(3):205-222.

Hymowitz MB, Hayes BB, Maury JJ, et al. (2006). Evaluation of medical students' knowledge,
attitudes, and personal practices of sun protection and skin self-examination. Archives of
Dermatology 142(4):523-524.

International Agency for Research on Cancer. (1992). Volume 55. Solar and ultraviolet radiation.
IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Lyon, France,
IARC.

International Commission on Non-Ionizing Radiation Protection. (2007). Protecting workers
from ultraviolet radiation. Vecchia P, Heitanen M, Stuck BE, van Deventer E, Niu S,
eds. Oberschleissheim, Germany, International Commission on Non-Ionizing Radiation
Protection.

Jackson KM, Aiken LS. (2000). A psychosocial model of sun protection and sunbathing in young
women: the impact of health beliefs, attitudes, norms, and self-efficacy for sun
protection. Health Psychology 19(5):469-478.

Jacobs EA, Karavolos K, Rathouz PJ, et al. (2005). Limited English proficiency and breast and
cervical cancer screening in a multiethnic population. American Journal of Public Health
95(8):1410-1416.

Jennings E, Whiteley J, Marcus-Blank B, et al. (2008). Physical activity and sun protection
behaviors in a randomized controlled physical activity trial [Abstract]. Annals of
Behavioral Medicine 35(Suppl):S15.

Jerant AF, Johnson JT, Sheridan CD, et al. (2000). Early detection and treatment of skin cancer.
American Family Physician 62(2):357-368.

Jerkegren E, Sandrieser L, Brandberg Y, et al. (1999). Sun-related behaviour and melanoma
awareness among Swedish university students. European Journal of Cancer Prevention
8(1):27-34.

129
Jurkowski JM, Johnson TP. (2005). Acculturation and cardiovascular disease screening practices
among Mexican Americans living in Chicago. Ethnicity & Disease 15(3):411-417.

Kaidbey KH, Agin PP, Sayre RM, et al. (1979). Photoprotection by melanin - a comparison of
black and Caucasian skin. Journal of the American Academy of Dermatology 1(3):249-
260.

Kandula NR, Lauderdale DS, Baker DW. (2007). Differences in self-reported health among
Asians, Latinos, and non-Hispanic whites: the role of language and nativity. Annals of
Epidemiology 17(3):191-198.

Kaplan MS, Marks G. (1990). Adverse effects of acculturation: psychological distress among
Mexican American young adults. Social Science & Medicine 31(12):1313-1319.

Keesling B, Friedman HS. (1987). Psychosocial factors in sunbathing and sunscreen use. Health
Psychology 6(5):477-493.

Kelfkens G, de Gruijl FR, van der Leun JC. (1991). Tumorogenesis by short-wave ultraviolet A:
papillomas versus squamous cell carcinomas. Carcinogenesis 12(8):1377-1382.

Khan LK, Sobal J, Martorell R. (1997). Acculturation, socioeconomic status, and obesity in
Mexican Americans, Cuban Americans, and Puerto Ricans. International Journal of
Obesity 21(2):91-96.

Knight JM, Kirincich AN, Farmer ER, et al. (2002). Awareness of the risks of tanning lamps does
not influence behavior among college students. Archives of Dermatology 138(10):1311-
1315.

Kraemer HC, Wilson T, Fairburn CG, et al. (2002). Mediators and moderators of treatment
effects in randomized clinical trials. Archives of General Psychiatry 59(10):877-883.

Kraemer HC, Frank E, Kupfer DJ. (2006). Moderators of treatment outcomes: clinical, research,
and policy importance. Journal of the American Medical Association 296(10):1286-1289.

Kranau EJ, Green V, Valencia-Weber G. (1982). Acculturation and the Hispanic woman:
attitudes toward women, sex-role attribution, sex-role behavior, and demographics.
Hispanic Journal of Behavioral Sciences 4(1):21-40.

LaFromboise T, Coleman HLK, Gerton J. (1993). Psychological impact of biculturalism:
evidence and theory. Psychological Bulletin 114(3):395-412.

Landrine H, Klonoff EA. (2004). Culture change and ethnic-minority health behavior: an operant
theory of acculturation. Journal of Behavioral Medicine 27(6):527-555.

Lara M, Gamboa C, Kahramanian MI, et al. (2005). Acculturation and Latino health in the United
States: a review of the literature and its sociopolitical context. Annual Review of Public
Health 26:367-397.

130
Leary MR, Jones JL. (1993). The social psychology of tanning and sunscreen use: self-
presentational motives as a predictor of health risk. Journal of Applied Social Psychology
23(17):1390-1406.

Lewis EC, Mayer JA, Slymen DJ. (2006). Postal workers' occupational and leisure-time sun
safety behaviors (United States). Cancer Causes & Control 17(2):181-186.

Lopez VA, Castro FG. (2006). Participation and program outcomes in a church-based cancer
prevention program for Hispanic women. Journal of Community Health 31(4):343-362.

Ma F, Collado-Mesa F, Hu S, et al. (2007). Skin cancer awareness and sun protection behaviors
in white Hispanic and white non-Hispanic high school students in Miami, Florida.
Archives of Dermatology 143(8):983-988.

Mack TM, Floderus B. (1991). Malignant melanoma risk by nativity, place of residence at
diagnosis, and age at migration. Cancer Causes & Control 2(6):401-411.

MacKie RM. (2006). Long-term health risk to the skin of ultraviolet radiation. Progress in
Biophysics and Molecular Biology 92(1):92-96.

MacKinnon DP, Lockwood CM, Hoffman JM, et al. (2002). A comparison of methods to test
mediation and other intervening variable effects. Psychological Methods 7(1):83-104.

MacKinnon DP, Lockwood CM. (2003). Advances in statistical methods for substance abuse
prevention research. Prevention Science 4(3):155-171.

MacKinnon DP, Fairchild AJ, Fritz MS. (2007). Mediation analysis. Annual Review of
Psychology 58:593-614.

Madronich S, de Gruijl FR. (1994). Stratospheric ozone depletion between 1979 and 1992:
implications for biologically active ultraviolet-B radiation and non-melanoma skin cancer
incidence. Photochemistry & Photobiology 59(5):541-546.

Maher JE, Boysun MJ, Rohde K, et al. (2005). Are Latinos really less likely to be smokers?
Lessons from Oregon. Nicotine & Tobacco Research 7(2):283-287.

Mahler HIM, Fitzpatrick B, Parker P, et al. (1997). The relative effects of a health-based versus
an appearance-based intervention designed to increase sunscreen use. American Journal
of Health Promotion 11(6):426-429.

Mainous AG, Majeed A, Koopman RJ, et al. (2006). Acculturation and diabetes among
Hispanics: evidence from the 1999-2002 National Health and Nutrition Examination
Survey. Public Health Reports 121:60-66.

Manne S, Fasanella N, Connors J, et al. (2004). Sun protection and skin surveillance practices
among relatives of patients with malignant melanoma: prevalence and predictors.
Preventive Medicine 39(1):36-47.

131
Manne S, Lessin S. (2006). Prevalence and correlates of sun protection and skin self-examination
practices among cutaneous malignant melanoma survivors. Journal of Behavioral
Medicine 29(5):419-434.

Manrai LA, Manrai AK. (1995). Effects of cultural-context, gender, and acculturation on
perceptions of work versus social/leisure time usage. Journal of Business Research
32(2):115-128.

Marin G, Marin BV. (1991). Research with Hispanic populations. Newbury Park, CA, Sage
Publications, Inc.

Marin G. (1992). Issues in the measurement of acculturation among Hispanics. Psychological
testing of Hispanics (pp.235-251). Geisinger KF, ed. Washington, DC, American
Psychological Association.

Markides KS, Coreil J. (1986). The health of Hispanics in the Southwestern United States: an
epidemiologic paradox. Public Health Reports 101(3):253-265.

Markides KS, Ray LA, Stroup-Benham CA, et al. (1990). Acculturation and alcohol consumption
in the Mexican American population of the Southwestern United States: findings from
NHANES 1982-84. American Journal of Public Health 80(Suppl.):42-46.

Markides KS, Lee DJ. (1991). Predictors of health status in middle-aged and older Mexican
Americans. Journal of Gerontology: Social Sciences 46(5):S243-S249.

Markides KS, Eschbach K. (2005). Aging, migration, and mortality: current status of research on
the Hispanic paradox. Journals of Gerontology: Series B 60B(Special issue II):68-75.

Marks R, Dorevitch AP, Mason G. (1990). Do all melanomas come from "moles"? A study of the
histological association between melanocytic naevi and melanoma. Australasian Journal
of Dermatology 31(2):77-80.

Marks R. (1996). Squamous cell carcinoma. The Lancet 347(9003):735-738.

Marquez DX, McAuley E. (2006). Gender and acculturation influences on physical activity in
Latino adults. Annals of Behavioral Medicine 31(2):138-144.

Masel MC, Rudkin LL, Peek MK. (2006). Examining the role of acculturation in health behaviors
of older Mexican Americans. American Journal of Health Behavior 30(6):684-699.

McManus W, Gould W, Welch F. (1983). Earnings of Hispanic men: the role of English language
proficiency. Journal of Labor Economics 1(2):101-130.

Mendoza RH, Martinez JL. (1981). The measurement of acculturation. Explorations in Chicano
psychology (pp.71-82). Baron Jr. A, ed. New York, Praeger Publishers.

132
Mendoza RH. (1984). Chapter 5. Acculturation and sociocultural variability. Chicano
Psychology, 2nd ed (pp.61-75). Martinez JL, Mendoza RH, eds. Orlando, FL, Academic
Press, Inc.

Michielutte R, Dignan MB, Sharp PC, et al. (1996). Skin cancer prevention and early detection
practices in a sample of rural women. Preventive Medicine 25(6):673-683.

Molgo M, Castillo C, Valdes R, et al. (2005). Sun exposure behaviors and knowledge among
Chileans [Article in Spanish]. Revista Medica de Chile 133(6):662-666.

Monzo LD, Rueda R. (2006). A sociocultural perspective on acculturation. Latino immigrant
families negotiating diverse discipline practices. Education and Urban Society 38(2):188-
203.

Moran A, Diez Roux AV, Jackson SA, et al. (2007). Acculturation is associated with
hypertension in a multiethnic sample. American Journal of Hypertension 20(4):354-363.

Moreland S, Engelman K, Greiner KA, et al. (2006). Papanicolaou testing among Native
American and Hispanic populations. Ethnicity & Disease 16(1):223-227.

Muller D, Judd CM, Yzerbyt VY. (2005). When moderation is mediated and mediation is
moderated. Journal of Personality and Social Psychology 89(6):852-863.

Murguia E, Saenz R. (2002). An analysis of the Latin Americanization of race in the United
States: a reconnaissance of color stratification among Mexicans. Race & Society 5:85-
101.

Myers HF, Rodriguez N. (2003). Acculturation and physical health in racial and ethnic
minorities. Acculturation: advances in theory, measurement, and applied research
(pp.163-185). Chun KM, Organista PB, Marin G, eds. Washington, DC, American
Psychological Association.

Myers LB, Horswill MS. (2006). Social cognitive predictors of sun protection intention and
behavior. Behavioral Medicine 32(2):57-63.

National Cancer Institute. (2005). Cancer Trends Progress Report - 2005 Update. Bethesda, MD,
National Cancer Institute.

National Cancer Institute. (2007). HINTS briefs, number 6. Bethesda, MD, National Cancer
Institute.

Naylor MF, Rigel DS. (2005). Chapter 6. Current concepts in sunscreens and usage. Cancer of
the skin (pp.71-84). Rigel DS, Friedman RJ, Dzubow LM, Reintgen DS, Bystryn, JC,
Marks R, eds. Philadelphia, Elsevier, Inc.

Neff JA, Hoppe SK. (1993). Race/ethnicity, acculturation, and psychological distress: fatalism
and religiosity as cultural resources. Journal of Community Psychology 21(1):3-20.

133

Negy C, Woods DJ. (1992a). A note on the relationship between acculturation and socioeconomic
status. Hispanic Journal of Behavioral Sciences 14(2):248-251.

Negy C, Woods DJ. (1992b). The importance of acculturation in understanding research with
Hispanic-Americans. Hispanic Journal of Behavioral Sciences 14(2):224-247.

Nelson DE, Kreps GL, Hesse BW, et al. (2004). The Health Information National Trends Survey
(HINTS): development, design, and dissemination. Journal of Health Communication
9(5):443-460.

Neuhouser ML, Thompson B, Coronado GD, et al. (2004). Higher fat intake and lower fruit and
vegetables intakes are associated with greater acculturation among Mexicans living in
Washington State. Journal of the American Dietetic Association 104(1):51-57.

Nicol I, Gaudy C, Gouvernet J, et al. (2007). Skin protection by sunscreen is improved by explicit
labeling and providing free sunscreen. Journal of Investigational Dermatology 127(1):41-
48.

Nole G, Johnson AW. (2004). An analysis of cumulative lifetime solar ultraviolet radiation
exposure and the benefits of daily sun protection. Dermatologic Therapy 17(Suppl 1):57-
62.

O'Malley AS, Kerner J, Johnson AE, et al. (1999). Acculturation and breast cancer screening
among Hispanic women in New York City. American Journal of Public Health
89(2):219-227.

Olsen SJ, Frank-Stromborg M. (1993). Cancer prevention and early detection in ethnically
diverse populations. Seminars in Oncology Nursing 9(3):198-209.

Palloni A, Arias E. (2004). Paradox lost: explaining the Hispanic adult mortality advantage.
Demography 41(3):385-415.

Palmer RC, Fernandez ME, Tortolero-Luna G, et al. (2005). Acculturation and mammography
screening among Hispanic women living in farmworker communities. Cancer Control
12(Suppl. 2):21-27.

Patel KV, Eschbach K, Ray LA, et al. (2004). Evaluation of mortality data for older Mexican
Americans: implications for the Hispanic paradox. American Journal of Epidemiology
159(7):707-715.

Pepper AC, Ruiz SY. (2007). Acculturation's influence on antifat attitudes, body image and
eating behaviors. Eating Disorders 15(5):427-447.

Perez-Escamilla R, Putnik P. (2007). The role of acculturation in nutrition, lifestyle, and
incidence of Type 2 diabetes among Latinos. The Journal of Nutrition 137(4):860-870.

134
Perez-Stable EJ, Sabogal F, Otero-Sabogal R, et al. (1992). Misconceptions about cancer among
Latinos and Anglos. Journal of the American Medical Association 268(22):3219-3223.

Perez-Stable EJ, Marin G, Marin BV. (1994). Behavioral risk factors: a comparison of Latinos
and non-Latino whites in San Francisco. American Journal of Public Health 84(6):971-
976.

Pichon LC, Mayer JA, Slymen DJ, et al. (2005). Ethnoracial differences among outdoor workers
in key sun-safety behaviors. American Journal of Preventive Medicine 28(4):374-378.

Pichon LC, Arredondo EM, Roesch S, et al. (2007). The relation of acculturation to Latinas'
perceived neighborhood safety and physical activity: a structural equation analysis.
Annals of Behavioral Medicine 34(3):295-303.

Pipitone M, Robinson JK, Camara C, et al. (2002). Skin cancer awareness in suburban
employees: a Hispanic perspective. Journal of the American Academy of Dermatology
47(1):118-123.

Polednak AP. (1997). Gender and acculturation in relation to alcohol use among Hispanic
(Latino) adults in two areas of the northeastern United States. Substance Use & Misuse
32(11):1513-1524.

Preacher KJ, Rucker DD, Hayes AF. (2007). Addressing moderated mediation hypotheses:
theory, methods, and prescriptions. Multivariate Behavioral Research 42(1):185-227.

Ramirez AG, Suarez L, Chalela P, et al. (2004). Cancer risk factors among men of diverse
Hispanic or Latino origins. Preventive Medicine 39(2):263-269.

Ranney L, Melvin C, Lux L, et al. (2006). Systematic review: smoking cessation intervention
strategies for adults and adults in special populations. Annals of Internal Medicine
145(11):845-856.

Reynolds KD, Chou C-P, Duan L, et al. (In review). Moderated mediation in skin cancer
prevention for middle school children. Prevention Science.

Robinson JK, Rigel DS, Amonette RA. (1997). Trends in sun exposure knowledge, attitudes, and
behaviors: 1986 to 1996. Journal of the American Academy of Dermatology 37(2 Pt.
1):179-186.

Robinson JD, Silk KJ, Parrott RL, et al. (2004). Healthcare providers' sun protection promotion
and at-risk clients' skin cancer prevention outcomes. Preventive Medicine 38(3):251-257.

Roetzheim RG, Pal N, Tennant C, et al. (1999). Effects of health insurance and race on early
detection of cancer. Journal of the National Cancer Institute 91(16):1409-1415.

135
Rose BM, Holmbeck GN, Coakley RM, et al. (2004). Mediator and moderator effects in
developmental and behavioral pediatric research. Developmental and Behavioral
Pediatrics 25(1):58-67.

Rosenstock I. (1974). Historical origins of the Health Belief Model. Health Education
Monographs 2(4):328-335.

Ross SA, Sánchez JL. (1990). Recreational sun exposure in Puerto Rico: trends and cancer risk
awareness. Journal of the American Academy of Dermatology 23(6 Pt 1):1090-1092.

Rubin AI, Chen EH, Ratner D. (2005). Basal-cell carcinoma. The New England Journal of
Medicine 353(21):2262-2269.

Rueda R, Monzo LD, Arzubiaga A. (2003). Academic instrumental knowledge: deconstructing
cultural capital theory for strategic intervention approaches. Current Issues in Education
[On-line] 6(14).

Rueschenberg E, Buriel R. (1989). Mexican American family functioning and acculturation: a
family systems perspective. Hispanic Journal of Behavioral Sciences 11(3):232-244.

Ruiz MS, Marks G, Richardson JL. (1992). Language acculturation and screening practices of
elderly Hispanic women. The role of exposure to health-related information from the
media. Journal of Aging & Health 4(2):268-281.

Rutten LJF, Augustson E, Wanke K. (2006). Factors associated with patients' perceptions of
health care providers' communication behavior. Journal of Health Communication
11(Suppl 1):135-146.

Rutten LJF, Moser RP, Beckjord EB, et al. (2007). Cancer communication: Health Information
National Trends Survey. Washington, DC, National Cancer Institute.

Sabogal F, Marin G, Otero-Sabogal R, et al. (1987). Hispanic familism and acculturation: what
changes and what doesn't. Hispanic Journal of Behavioral Sciences 9(4):397-412.

Saladi RN, Persaud AN. (2005). The causes of skin cancer: a comprehensive review. Drugs of
Today 41(1):37-53.

Salas R, Mayer JA, Hoerster KD. (2005). Sun-protective behaviors of California farmworkers.
Journal of Occupational and Environmental Medicine 47(12):1244-1249.

Santmyire BR, Feldman SR, Fleischer AB. (2001). Lifestyle high-risk behaviors and
demographics may predict the level of participation in sun protection behaviors and skin
cancer primary prevention in the United States. Results of the 1998 National Health
Interview Survey. Cancer 92(5):1315-1324.

Saraiya M, Hall IH, Uhler RJ. (2002). Sunburn prevalence among adults in the United States,
1999. American Journal of Preventive Medicine 23(2):91-97.

136

Saraiya M, Balluz L, Wen XJ, et al. (2007). Sunburn prevalence among adults - United States,
1999, 2003, and 2004. Morbidity and Mortality Weekly Report 56(21):524-528.

SAS Institute Inc. (2004). SAS, version 9.1. Cary, NC, SAS Institute, Inc.

Scerri L, Aquilina S, Amato Gauci A, et al. (2002). Sun awareness and sun protection practices in
Malta. Journal of the European Academy of Dermatology and Venereology 16(1):47-52.

Sciamanna CN, Zampi A, Weinstock MA. (2002). Comparison of sun protection behaviors
among physicians and patients. American Journal of Health Promotion 16(6):319-322.

Shah M, Zhu K, Potter J. (2006). Hispanic acculturation and utilization of colorectal cancer
screening in the United States. Cancer Detection and Prevention 30(3):306-312.

Shetterly SM, Baxter J, Mason LD, et al. (1996). Self-rated health among Hispanic vs. non-
Hispanic white adults: the San Luis Valley health and aging study. American Journal of
Public Health 86(12):1798-1801.

Shoveller JA, Lovato CY, Peters L, et al. (2000). Canadian national survey on sun exposure &
protective behaviours: outdoor workers. Canadian Journal of Public Health 91(1):34-35.

Shrout PE, Bolger N. (2002). Mediation in experimental and nonexperimental studies: new
procedures and recommendations. Psychological Methods 7(4):422-445.

Sinclair C. (2007). Vitamin D - an emerging issue in skin cancer control. Implications for public
health practice based on the Australian experience. Recent Results in Cancer Research
174:197-204.

Slattery ML, Sweeney C, Edwards S, et al. (2006). Physical activity patterns and obesity in
Hispanic and non-Hispanic white women. Medicine & Science in Sports & Exercise
38(1):33-41.

Smith DP, Bradshaw BS. (2006). Rethinking the Hispanic paradox: death rates and life
expectancy for US non-Hispanic white and Hispanic populations. American Journal of
Public Health 96(9):1686-1692.

Sorlie PD, Backlund E, Johnson NJ, et al. (1993). Mortality by Hispanic status in the United
States. Journal of the American Medical Association 270(20):2464-2468.

Stanton WR, Janda M, Baade PD, et al. (2004). Primary prevention of skin cancer: a review of
sun protection in Australia and internationally. Health Promotion International
19(3):369-378.

Stiller MJ. (2003). Chapter 214. Sunlight and skin damage. The Merck manual of medical
information (pp.1230-1233). Beers MH, Fletcher AJ, Jones TV, Porter R, Berkwits M,
Kaplan JL, eds. New York, Pocket Books.

137

Stone VB, Parker V, Quarterman M, et al. (1999). The relationship between skin cancer
knowledge and preventive behaviors used by parents. Dermatology Nursing 11(6):411-
426.

Strickland PT, Vitasa BC, West SK, et al. (1989). Quantitative carcinogenesis in man: solar
ultraviolet B dose dependence of skin cancer in Maryland watermen. The Journal of the
National Cancer Institute 81(24):1910-1913.

Sue S. (2003). Foreword. Acculturation: advances in theory, measurement, and applied research.
Chun KM, Organista PB, Marin G, eds. Washington, DC, American Psychological
Association.

Surveillance Epidemiology and End Results Program. (2008). Melanoma of the skin. Cancer Stat
Fact Sheets. Bethesda, MD, National Cancer Institute.

Tadokoro T, Kobayashi N, Zmudzka BZ, et al. (2003). UV-induced DNA damage and melanin
content in human skin differing in racial/ethnic origin. The FASEB Journal 17(9):1177-
1179.

Thieden E, Philipsen PA, Sandby-Moller J, et al. (2005). Sunburn related to UV radiation
exposure, age, sex, occupation, and sun bed use based on time-stamped personal
dosimetry and sun behavior diaries. Archives of Dermatology 141(4):482-488.

Trinidad DR, Gilpin EA, Messer K, et al. (2006). Trends in smoking among Hispanic women in
California: relationship to English language use. American Journal of Preventive
Medicine 31(3):257-260.

Turner LR, Mermelstein RJ. (2005). Psychosocial characteristics associated with sun protection
practices among parents of young children. Journal of Behavioral Medicine 28(1):77-90.

Turra CM, Goldman N. (2007). Socioeconomic differences in mortality among U.S. adults:
insights into the Hispanic paradox. Journal of Gerontology: Social Sciences
62B(3):S184-S192.

Turrisi R, Hillhouse J, Gebert C, et al. (1999). Examination of cognitive variables relevant to
sunscreen use. Journal of Behavioral Medicine 22(5):493-509.

U.S. Census Bureau. (2006). Nation's population one-third minority. Washington, DC, U.S.
Census Bureau.

U.S. Census Bureau. (2008). Table 4: Estimates of the population by race and Hispanic origin for
the United States and States: July 1, 2007 (SC-EST2007-04). Washington, DC, U.S.
Census Bureau, Population Division.

138
U.S. Department of Health and Human Services. (2000). Healthy People 2010. 2nd Ed. With
understanding and improving health and objectives for improving health. Washington,
DC, U.S. Government Printing Office.

Unger JB, Cruz TB, Rohrbach LA, et al. (2000). English language use as a risk factor for
smoking initiation among Hispanic and Asian American adolescents: evidence for
mediation by tobacco-related beliefs and social norms. Health Psychology 19(5):403-410.

Unger JB, Gallaher P, Shakib S, et al. (2002). The AHIMSA acculturation scale: a new measure
of acculturation for adolescents in a multicultural society The Journal of Early
Adolescence 22(3):225-251.

Urbach F. (1997). Ultraviolet radiation and skin cancer of humans. Journal of Photochemistry
and Photobiology B: Biology 40(1):3-7.

Vail-Smith K, Felts WM. (1993). Sunbathing: college students' knowledge, attitudes, and
perceptions of risk. Journal of American College Health 42(1):21-26.

Vázquez-Botet M, Latoni D, Sánchez JL. (1990 ). Malignant melanoma in Puerto Rico [Article in
Spanish]. Boletín de la Asociación Médica de Puerto Rico 82(10):454-457.

Vega WA, Amaro H. (1994). Latino outlook: good health, uncertain prognosis. Annual Review of
Public Health 15:39-67.

Wathen CN. (2006). Health information seeking in context: how women make decisions
regarding hormone replacement therapy. Journal of Health Communication 11(5):477-
493.

Weinstock MA, Rossi JS, Redding CA, et al. (2002). Randomized controlled community trial of
the efficacy of a multicomponent stage-matched intervention to increase sun protection
among beachgoers. Preventive Medicine 35(6):584-592.

Wells KB, Golding JM, Hough RL, et al. (1989). Acculturation and the probability of use of
health services by Mexican Americans. Health Services Research 24(2):237-257.

Wolch J, Zhang J. (2004). Beach recreation, cultural diversity and attitudes toward nature.
Journal of Leisure Research 36(3):414-443.

World Health Organization. (2006). Global disease burden from solar ultraviolet radiation.
Geneva, World Health Organization.

Wright CY, Reeder AI. (2005). Youth solar ultraviolet radiation exposure, concurrent activities
and sun-protective practices: a review. Photochemistry & Photobiology 81(6):1331-1342.

Wu ZH, Black SA, Markides KS. (2001). Prevalence and associated factors of cancer screening:
why are so many older Mexican American women never screened? Preventive Medicine
33(4):268-273.

139

Zane N, Mak W. (2003). Major approaches to the measurement of acculturation among ethnic
minority populations: a content analysis and an alternative empirical strategy.
Acculturation: advances in theory, measurement, and applied research (pp.39-60). Chun
KM, Organista PB, Marin G, eds. Washington, DC, American Psychological Association.

Zanetti R, Gafa L, Franceschi S, et al. (1999). Estimate of the proportion of skin tumors
attributable to sun exposure in 3 Italian populations [Article in Italian]. Epidemiologia e
Prevenzione 23(4):416-422.

Zea MC, Asner-Self KK, Birman D, et al. (2003). The abbreviated multidimensional
acculturation scale: empirical validation with two Latino/Latina samples. Cultural
Diversity and Ethnic Minority Psychology 9(2):107-126.

Zemore SE. (2007). Acculturation and alcohol among Latino adults in the United States: a
comprehensive review. Alcoholism: Clinical & Experimental Research 31(12):1968-
1990.

Abstract (if available)

Abstract Acculturation has emerged as an independent risk factor for different health practices across immigrant groups. Prompted by evidence of increasing melanoma incidence among Latinos, the present research investigated the impact of acculturation on Latinos’ sun safety. Data from 496 Latino respondents to the 2005 Health Information National Trends Survey were analyzed with multiple linear regression and structural equation models. Sunscreen use, seeking shade and wearing protective clothing were the primary outcomes, assessed by frequency scales. Acculturation was the primary predictor, assessed by a composite index. The direct, mediated, and moderated effects of acculturation on sun-safe behaviors were investigated in three studies. Hypothesized mediators included health status, education, and social networks

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Psychosocial and cultural factors in the primary prevention of melanoma targeted to multiethnic children

PDF

Sociocultural stress, coping and substance use among Hispanic/Latino adolescents

PDF

Exploring the role of peer influence, linguistic acculturation, and social networks in substance use

PDF

Multilevel sociodemographic correlates of the health and healthcare utilization of childhood cancer survivors

PDF

Multilevel influences of care engagement and long-term survival among childhood, adolescent, and young adult cancer survivors

PDF

Examination of sun awareness behavior, knowledge, and attitudes before and after a sun awareness program in 4th to 8th grade students

PDF

Motivation and the meanings of health behavior as factors associated with eating behavior in Latino youth

PDF

Cultural risk and protective factors for tobacco use behaviors and depressive symptoms among American Indian adolescents in California

PDF

Psychosocial and behavioral ractors associated with emotional eating in adolescents

PDF

Genetic variants and smoking progression in Chinese adolescents

PDF

Childhood cancer survivorship: parental factors associated with survivor's follow-up care behavior and mental health

PDF

Competing risks: the role of the perceived consequences of refusing to share injection equipment among injection drug users

PDF

Exploring geography and multi-level factors associated with disparities in HPV vaccination among Latino adolescents in Los Angeles County

PDF

Social self-control and adolescent substance use

PDF

Factors and correlates of sexual behaviors among young adults from continuation high schools

PDF

Problematic alcohol use in Hispanic emerging adults: the role of perceived discrimination, cultural identity, and salient cultural values

PDF

Subclinical atherosclerosis in overweight Latino youth: influence of cardiometabolic risk factors

PDF

The role of depression symptoms on social information processing and tobacco use among adolescents

PDF

A sociocultural and developmental approach to intimate partner violence among a sample of Hispanic emerging adults

PDF

Acculturation team-based clinical program: pilot program to address acculturative stress and mental health in the Latino community

Asset Metadata

Creator Andreeva, Valentina A. (author)

Core Title The role of acculturation in the sun-safe behaviors of US Latino adults in the United States

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Preventive Medicine (Health Behavior)

Publication Date 09/19/2008

Defense Date 07/09/2008

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag acculturation,Latinos,OAI-PMH Harvest,skin cancer prevention,sun safety

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Reynolds, Kim D. (committee chair), Cockburn, Myles G. (committee member), Rueda, Robert S. (committee member), Unger, Jennifer B. (committee member), Yaroch, Amy L. (committee member)

Creator Email andreeva@usc.edu,valentina.a.andreeva@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m1610

Unique identifier UC1117042

Identifier etd-Andreeva-2268 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-95590 (legacy record id),usctheses-m1610 (legacy record id)

Legacy Identifier etd-Andreeva-2268.pdf

Dmrecord 95590

Document Type Dissertation

Rights Andreeva, Valentina A.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu