Journal of Adolescent Health 48 (2011) 410 – 414

www.jahonline.org

Original article

Religiosity As a Protective Factor Against HIV Risk Among Young Transgender

Women
Nadia Dowshen, M.D.a,b,*, Christine M. Forke, M.S.N., R.N., C.R.N.P.a,b,c, Amy K. Johnson, M.S.W.d,e,
Lisa M. Kuhns, Ph.D., M.P.H.e, David Rubin, M.D., M.S.C.E.b,c, and Robert Garofalo, M.D., M.P.H.d,e,f
a
Craig-Dalsimer Division of Adolescent Medicine, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania
b
Department of Pediatrics, School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania
c
Division of General Pediatrics, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania
d
Division of General Pediatrics, Children’s Memorial Hospital, Chicago, Illinois
e
Department of Research, Howard Brown Health Center, Chicago, Illinois
f
Department of Pediatrics, Feinberg School of Medicine, Northwestern University, Chicago, Illinois

Article history: Received February 2, 2010; Accepted July 27, 2010

Keywords: Adolescent; Transgender; Female; Sexual risk; HIV; Alcohol; Religion

A B S T R A C T:

Purpose: Young transgender women (YTW) face many challenges to their well-being, including homeless-
ness, joblessness, victimization, and alarming rates of HIV infection. Little has been written about factors that
might help in preventing HIV in this population. Our objective was to examine the role of religion in the lives
of YTW and its relationship to HIV risk.
Methods: This study is derived from baseline data collected for an HIV prevention intervention. A convenience sample
of YTW aged 16–25 years from Chicago were recruited consecutively and completed an audio computer-assisted
self-interview. Logistic regression models were used to evaluate the relationship between sexual risk taking (sex work,
multiple anal sex partners, unprotected receptive anal sex), alcohol use, formal religious practices (service attendance,
reading/studying scripture), and God consciousness (prayer, thoughts about God).
Results: A total of 92 YTW participated in the study, their mean age being 20.4 years; 58% were African
American, 21% white, and 22% other. On multivariate logistic regression, alcohol use was significantly
associated with sexual risk in both models, with adjusted odds ratio (OR) of 5.28 (95% confidence intervals
[CI]: 1.96 –14.26) in the Formal Practices model and 3.70 (95% CI: 1.53– 8.95) in the God Consciousness model.
Controlling for alcohol use, it was found that Formal Practices was significantly associated with sexual risk
(OR ⫽ .29, 95% CI: .11–.77), but God Consciousness was not (OR ⫽ .60, 95% CI: .25–1.47).
Conclusion: Among YTW, formal religious practices may attenuate sexual risk-taking behaviors and therefore HIV
risk. Further research is needed to explore the role of the religion in the lives of YTW as a protective asset.
䉷 2011 Society for Adolescent Health and Medicine. All rights reserved.

Male-to-female transgender youth or young transgender alence of HIV infection in this population, reported to be as
women (YTW) face many challenges to their physical and emo- high as 30% in one study [1], has been attributed to several risk
tional well-being, including homelessness, joblessness, victim- factors, including sex work, risky sexual behaviors, substance
ization, and alarming rates of HIV infection. The increased prev- abuse, and lack of access to culturally appropriate health care
that addresses the specific issues of this population including
barriers to medical and surgical therapies such as limited
Dr Dowshen was at Children’s Memorial Hospital and Howard Brown Health evidence, few trained providers, and cost of services [1–9].
Center in Chicago when the study was conducted. Although a fair amount has been written about risk factors
* Address correspondence to: Nadia Dowshen, M.D., Craig-Dalsimer Division
associated with HIV transmission, there have been few studies
of Adolescent Medicine, The Children’s Hospital of Philadelphia, Main Building,
11th Floor NW, Suite 10, 34th and Civic Center Boulevard, Philadelphia, PA 19104. exploring which factors might be protective in preventing
E-mail address: dowshenn@email.chop.edu HIV, particularly among YTW.

1054-139X/$ - see front matter 䉷 2011 Society for Adolescent Health and Medicine. All rights reserved.
doi:10.1016/j.jadohealth.2010.07.021
 N. Dowshen et al. / Journal of Adolescent Health 48 (2011) 410 – 414 411

Religion plays an important role in the lives of many youth Participants

and has been found to be a significant protective factor associ-
ated with decreased rates of specific risk behaviors. Some studies Participants’ eligibility criteria included the following:
found that increased religiosity, meaning religious beliefs and (1) being 16 –25 years of age; (2) self-identifying as a trans-
practices, was associated with decreased rates of sexual activity woman or feminine-identified/male-born person; and (3) com-
and alcohol and tobacco use [10 –13]. For example, in one study mitting to attend the program sessions.
of youth in foster care, each unit increase of religious service Potential participants were recruited through both active
attendance was associated with a 17% and 19% decrease in sexual and passive methods by transgender-identified staff mem-
activity and cigarette use, respectively [14]. In another study of bers. Active recruitment included face-to-face outreach to
college students, increased strength in religious beliefs was associ- venues and public areas frequently visited by young trans-
ated with decreased rates of alcohol use and decreased participa- women, such as night clubs, beauty pageant events, and pop-
tion in risky sexual behavior for women, but not for men [15]. ular outdoor locations (e.g., parks). Passive recruitment in-
In conceptualizing religion, researchers have used multiple cluded posting flyers and other study materials at local LGBT
community organizations and on Chicago-based web pages
frameworks that distinguish between distal or public measures
specifically directed toward transgender youth on social net-
(i.e., religious practices, worship services, and connection to a
working sites, such as Facebook, LiveJournal, and MySpace.
religious community) versus proximal or private measures (i.e.,
Participants received $20 to compensate them for the time
personal spirituality and beliefs) of religion [16]. Several studies
required to complete the baseline assessment.
have shown associations between both proximal and distal reli-
giosity and a reduction in a range of adolescent risk behaviors
Measures
[17,18]. However, these protective associations between religi-
osity and health behaviors may not necessarily generalize to
The assessment was administered using an audio computer-
lesbian, gay, bisexual, and transgender (LGBT) youth, and little assisted self-interview. This method of conducting interviews
has been written about the role of religion in this population [19]. using a personal computer allows for privacy and standardiza-
Many LGBT youth may feel isolated from religious institutions tion of the interview, and studies have shown it to be an accept-
whose beliefs are in conflict with their gender or sexual identity, able, accurate method for measuring sexual risk behaviors
whereas others may find religious communities or beliefs to be [25,26]. The survey included measurement of general demo-
an even greater source of support than that for their peers. graphics, HIV risk behaviors, as well as psychosocial and health-
Little has been written about the specific role of religious related characteristics.
beliefs and institutions in youth HIV prevention. Generally, HIV The primary outcome was sexual risk behavior, measured
prevention programs have been classified as either individual or using items adapted from the AIDS-Risk Behavior Assessment
small-group based (i.e., peer-based education in a small group of and the CDC’s core sexual risk behavior measure. Items from
culturally similar youth) which account for the majority of effec- these measures were adapted to assess HIV-related sexual risk
tive evidence-base interventions versus community- or policy- for YTW. All questions regarding sexual behavior were worded
oriented programs (i.e., prevention messages delivered locally or to specify voluntary sexual activity. A composite sexual risk
nationally through social media) which are less common [20 – index was constructed based on responses to the three follow-
22]. The Center for Disease Control (CDC) has prioritized partner- ing items: (1) history of sex work (paying or trading male sex
ing with communities of faith in addressing the HIV epidemic, partners) in the past 3 months; (2) history of multiple male anal sex
recognizing that religious institutions may be in a unique posi- partners (two or more) in the past 3 months; (3) history of unpro-
tion to intervene with youth on both the individual and commu- tected receptive anal sex in the past 3 months. Items with an affir-
nity levels [23,24]. For YTW, in whom HIV risk is high, it is mative response scored one point, and negative responses scored
particularly important to understand how and whether religious zero. Individual item scores were summed to compute the compos-
involvement may affect sexual risk behavior. ite risk score (range: 0 –3). For the purpose of analyses, participants
The purpose of this study was to examine the role of religios- who scored zero were classified as having low sexual risk and those
ity in the lives of young transgender women and explore its scoring one or greater were classified as having high sexual risk.
relationship to sexual risk behaviors that can result in HIV trans- The principal exposure of interest was religiosity, assessed using
mission. Our main hypothesis was that increased religious prac- the Religious Background and Behaviors (RBB) questionnaire [27].
tices and beliefs would be associated with a decrease in sexual The RBB is comprised of two validated subscales, Formal Practices
and God Consciousness, which include all items and can be
risk behaviors that may lead to HIV infection.
summed to obtain the total RBB score. The Formal Practices sub-
scale includes eight items that assessed the following four con-
Methods cepts: meditation, worship service attendance, reading scriptures,
and direct experiences with God. The God Consciousness subscale
Study Design included five items reflecting three concepts: religious self-
description, prayer, and thoughts about God. Each item on the
Data were collected at an LGBT-focused health center as subscale was rated from 0 to 3, and items on the appropriate sub-
part of a larger HIV prevention intervention study for young scale were summed to obtain respective subscale scores. Items
transgender-identified women in Chicago, IL, between Janu- from each subscale are included in Table 1. Higher subscale scores
ary 2007 and November 2008. Institutional Review Board ap- reflected greater participation in religious practices (Formal Prac-
proval was obtained before study implementation, and a tice subscale) or stronger religious beliefs (God Consciousness sub-
waiver of parental consent was granted for participants aged scale). Cutpoints for each subscale were established based on me-
16 –17 years. dian values. The cutpoint for the Formal Practices subscale was
412 N. Dowshen et al. / Journal of Adolescent Health 48 (2011) 410 – 414

Table 1 Fisher’s exact tests for categorical variables and Student’s t-tests
Religiosity measure: RBBQ subscale items and Chronbach ␣ values or ANOVA for continuous variables. Logistic regression analysis
Subscale name and respective items Cronbach ␣ or was done to predict high sexual risk as a function of formal
correlation coefficient religious practices and God consciousness, while controlling for
age, race, educational level, employment status, and alcohol use.
God Consciousness subscale (proximal religiosity) ␣ ⫽ .69
Which of the following best describes you at Because of the high correlation between the Formal Practices and
the present time? God Consciousness subscales (r ⫽ .6), they were tested in sepa-
I do not believe in God. rate models. The first model included Formal Practices and not
I believe we can’t really know about God. God Consciousness, whereas the second model included God
I don’t know what to believe about God.
I believe in God, but I’m not religious.
Consciousness but not Formal Practices.
I believe in God and practice religion. Output of all models included adjusted odds ratios and 95%
For the past year, how often have you done confidence intervals (CI). The overall model fit was assessed
the following? using log likelihood estimates, and models were compared using
Thought about God
the Akaike information criterion (AIC) which is a goodness of fit
Prayed
Have you ever in your life done the following? estimate for model selection. A lower AIC indicates a better
Thought about God fitting model [29]. Analyses were carried out using STATA ver-
Prayed sion 11 [30].
Formal Practices subscale (distal religiosity) ␣ ⫽ .67
For the past year, how often have you done the
Results
following?
Meditated
Attended worship service A total of 92 young transgender women participated in the
Read-studied scriptures, holy writings study. The mean age of participants was 20.4 ⫾ 2.19 (range:
Had direct experiences of God
16 –25) years. Characteristics of women in the sample are pre-
Have you ever in your life done the following?
Meditated
sented in Table 2.
Attended worship service Scores on the sexual risk index ranged from 0 to 3: 39 (42%)
Read or studied scriptures, holy writings participants reported 0 risk, 22 (24%) reported 1 risk, 21 (23%)
Had direct experiences of God reported 2 risks, and 10 (11%) reported 3 risks. When sexual risk
Total Scale ⫽ God Conscious subscale ⫹ Formal r ⫽ .60a
was categorized into a dichotomous variable comparing no risk
Practices subscale
to any risk, 39 (42%) participants were classified as low sexual
a
Correlation between the God Conscious subscale and Formal Practices
risk and 53 (58%) as high sexual risk.
subscale.
When examining religiosity, subscales within the Formal
Practices subscale and the God Consciousness subscale showed
established as ⱕ12 (low) and ⬎12 (high). The cutpoint for the God reasonable internal consistency in this sample, with Cronbach’s
Consciousness subscale was established as ⱕ18 (low) and ⬎18 ␣ values of .67 and .69, respectively. The mean score on the
(high). Formal Practices subscale was 13.0 ⫾ 9.9 (range: 0 –36), and on
Alcohol has been identified as an independent risk factor for
unprotected sex [28] that may lead to HIV infection. In addition,
Table 2
alcohol use and religious involvement are inversely correlated Sample characteristics (N ⫽ 92)
[12,14]. The confounding effects of alcohol were assessed by the
Characteristics n %
following question from the CDC substance use measure: “How
often have you gotten drunk or buzzed from alcohol in the past Age
three months?” Participant responses were then dichotomized 16–20 44 48
into those who have been drunk or buzzed versus those who 21–25 48 52
Race/ethnicity
have not been so in the past 3 months. African American 53 58
Other independent variables included patient demographics. White 19 21
Participants reported age, race/ethnicity, highest level of educa- Other 20 22
tion obtained, and current employment status. Age was consid- Highest level of education
11th grade or less 31 34
ered as a continuous variable. Because of small cell sizes, race/
12th grade, GED or higher 61 66
ethnicity responses were divided into the following three Current employment
categories: African American, other minority race/ethnicity Not employed 70 76
(multiracial, Asian, Latina), and Caucasian. Educational level was Employed 22 24
originally collected as a six-category variable ranging from grade Drunk or buzzed in past 3 months
No 38 41
8 through college. For the purpose of analysis and to assist with
Yes 54 59
interpretation, participants were classified as having completed Sexual risk index
high school or not completed high school. Low risk 39 42
High risk 53 58
Religion
Analysis
God consciousness
Low (ⱕ18) 51 56
Data were described using frequencies for categorical vari- High (⬎18) 40 44
ables, and means (with standard deviations) for continuous vari- Formal practices
ables. Univariate analyses examining the association between Low (ⱕ12) 46 51
High (⬎12) 44 49
independent variables and risky sexual behavior included ␹2 and
 N. Dowshen et al. / Journal of Adolescent Health 48 (2011) 410 – 414 413

Table 3 [17,31,32]. In our study, it is unclear why God Consciousness or

Multivariate models exploring associations between formal religious practice proximal religiosity was not protective in the same way that
and God consciousness on the odds of sexual risk among young transgender
Formal Practices or distal religiosity was. Perhaps for these youth
women
who are so often stressed and isolated, the relationship between
Formal Practices God Consciousness
formal practices and deceased sexual risk behavior may be me-
model model
diated by structural and personal factors like social support,
OR CI OR CI
community connectedness, and stress relief through group wor-
Religiosity ship. By contrast, proximal religiosity— one’s particular beliefs or
Formal practices .28* .10–.77* – – religious identification—might be a less important factor in de-
God consciousness – – .44 .17–1.10
termining sexual risk because particular religious doctrines or
Alcohol use 5.44* 1.97–15.0* 3.92* 1.57–9.78*
beliefs may be viewed as personally irrelevant or judgmental,
* Values are statistically significant (p ⬍ .05). given the challenges that YTW often face.
In a recent review of evidence-based interventions for HIV
prevention among at-risk adolescents, the five programs chosen
the God Consciousness it was 15.0 ⫾ 5.9 (range: 0 –22). Using the as the most effective were all small group-based or individual
established cutpoints, 44 (49%) participants were classified as interventions that focused on personal level principles, including
having high Formal Practices, and 40 (44%) as having high God believing in yourself, committing to change, and being prepared
Consciousness. [21]. However, none of these interventions addressed structural
In response to the alcohol use item, 54 (59%) participants factors, such as how membership and regular contact with a
reported alcohol consumption in the past 3 months, whereas 38 religious community may support one’s ability to maintain or
(41%) reported no alcohol consumption in the past 3 months. avoid certain behaviors. As such, our finding that participation in
None of the demographic variables or total RBB score was
a religious community is associated with decreased sexual risk
significantly associated with sexual risk on univariate compari-
behavior may point toward innovative HIV prevention strategies
sons, although their relationship was in the expected direction.
that have not been previously prioritized.
Those with higher sexual risk were more likely to use alcohol
There are several limitations to this study. First, small sample
(70% vs. 30%, p ⫽ .003).
size may have resulted in insufficient power to detect an associ-
Regression results are presented in Table 3. Alcohol use was
ation between God consciousness and sexual risk. Thus, our
significantly associated with high sexual risk in both models.
results should be confirmed in a larger cohort. Second, the prob-
Controlling for alcohol use, Formal Practices was significantly
able oversampling of the highest risk individuals (since recruit-
associated with sexual risk, but God Consciousness was not.
ment was centered around high-risk venues or places where
When comparing models, the Formal Practices (AIC ⫽ 122)
at-risk youth seek services), and that the majority of the partici-
model had a better overall fit than the God Consciousness (AIC ⫽
pants were urban, ethnic youth being a minority, may limit its
129) model, as indicated by the lower AIC value.
generalizability to all YTW. Third, dichotomization of continuous
The final model included only alcohol use and Formal Prac-
tice/God Consciousness, given the lack of statistical association variables may lead to loss of information. Fourth, the cross-
with demographic characteristics. For the condensed Formal sectional design demonstrates association but not necessarily
Practices model, alcohol use (OR ⫽ 5.28, 95% CI: 1.96 –14.26) and causality. Finally, this study does not include any qualitative data
formal practices (OR ⫽ .29, 95% CI: .11–.77) remained significant, about the experience of religious practices and beliefs among
with point estimates and confidence intervals similar to the YTW which may be very different from other LGBT youth or
overall model. Likewise, for the condensed God Consciousness youth in general.
model, results for alcohol use (OR ⫽ 3.70, 95% CI: 1.53– 8.95) and Despite these limitations, this is one of the first studies to our
God consciousness (OR ⫽ .60, 95% CI: .25–1.47) were similar to knowledge that demonstrates a protective factor associated with
the full model. When comparing the condensed models to the decreased HIV-related sexual risk behavior for YTW. Although
full models including demographics, the condensed formal prac- some religious doctrines or communities may not be supportive
tices model (AIC ⫽ 112) and condensed God consciousness of transgender youth, our data affirm recent efforts by the CDC to
model (AIC ⫽ 120) demonstrated a better fit. prioritize the involvement of religious institutions in HIV pre-
vention efforts, particularly in the African American community
Discussion [32]. The finding that distal religiosity, which operates on both an
individual and community level, decreased sexual risk behavior
Although YTW generally represent a population isolated from suggests that working with communities of faith to develop HIV
typical societal and cultural institutions, such as the education prevention interventions targeting both individual behavior and
system and formal employment, this study is one of the first to community support may be particularly effective. Future re-
have reported high levels of religious involvement among these search, both qualitative and quantitative, would require to ex-
youth, when compared with previously reported data of the RBB amine in greater depth how religious communities and practices
for other populations [27]. may be a source of support and lead to improved health out-
Our study suggests that formal religious practices (religious comes for this vulnerable population.
service attendance and reading/studying scripture) may attenu-
ate sexual risk taking behavior (sex work, multiple partners,
Acknowledgments
unprotected anal receptive sex) and therefore HIV risk, among
YTW. Several authors have identified potential mediators, such This project was done with funding provided by Center for Disease
as social support or forgiveness, that may explain the connection Control and Prevention, 5UR6PS000396-02; U22PS000486-02. The
between religion and positive health or behavioral outcomes authors would like to acknowledge Howard Brown Health Center and
414 N. Dowshen et al. / Journal of Adolescent Health 48 (2011) 410 – 414

Broadway Youth Center staff for their dedication to this research and [15] Poulson RL, Eppler MA, Satterwhite TN, et al. Alcohol consumption, strength
of religious beliefs, and risky sexual behavior in college students. J Am Coll
most importantly the youth who participated in the study. Health 1998;46:227–32.
[16] Rew L, Wong YJ. A systematic review of associations among religiosity/
spirituality and adolescent health attitudes and behaviors. J Adolesc Health
References 2006;38:433– 42.
[17] Cotton S, Zebracki K, Rosenthal SL, et al. Religion/spirituality and adolescent
[1] Garofalo R, Deleon J, Osmer E, et al. Overlooked, misunderstood and at-risk: health outcomes: A review. J Adolesc Health 2006;38:472– 80.
Exploring the lives and HIV risk of ethnic minority male-to-female trans- [18] Cotton S, Berry D. Religiosity, spirituality, and adolescent sexuality. Adolesc
gender youth. J Adolesc Health 2006;38:230 – 6. Med State Art Rev 2007;18:471– 83, vi.
[2] Sanchez NF, Sanchez JP, Danoff A. Health care utilization, barriers to care, [19] Rostosky SS, Danner F, Riggle ED. Is religiosity a protective factor against
and hormone usage among male-to-female transgender persons in New substance use in young adulthood? Only if you’re straight! J Adolesc Health
York City. Am J Public Health 2009;99:713–9. 2007;40:440 –7.
[3] Melendez RM, Exner TA, Ehrhardt AA, et al. Health and health care among [20] DiClemente RJ, Wingood GM. Expanding the scope of HIV prevention for
male-to-female transgender persons who are HIV positive. Am J Public adolescents: Beyond individual-level interventions. J Adolesc Health 2000;
Health 2006;96:1034 –7. 26:377– 8.
[4] Nemoto T, Operario D, Keatley J, et al. HIV risk behaviors among male-to- [21] Rotheram-Borus MJ, Swendeman D, Flannery D, et al. Common principles
female transgender persons of color in San Francisco. Am J Public Health embedded in effective adolescent HIV prevention programs. AIDS Behav
2004;94:1193–9. 2009;13:387–98.
[5] Sevelius JM, Reznick OG, Hart SL, Schwarcz S. Informing interventions: The [22] Rotheram-Borus MJ, Swendeman D, Flannery D, et al. Common factors in
importance of contextual factors in the prediction of sexual risk behaviors effective HIV prevention programs. AIDS Behav 2009;13:399 – 408.
among transgender women. AIDS Educ Prev 2009;21:113–27. [23] CDC. National Prevention Information Network: Faith partnership for
[6] Wilson EC, Garofalo R, Harris RD, et al. Transgender female youth and sex HIV Prevention. Available at: http://www.cdcnpin.org/scripts/
work: HIV risk and a comparison of life factors related to engagement in sex campaign/cdc.asp#six.
work. AIDS Behav 2009;13:902–13. [24] CDC. Consultation on faith and HIV prevention. Atlanta, GA: CDC, 2006.
[7] Operario D, Soma T, Underhill K. Sex work and HIV status among transgen- [25] Metzger DS, Koblin B, Turner C, et al. Randomized controlled trial of audio
computer-assisted self-interviewing: Utility and acceptability in longitudi-
der women: Systematic review and meta-analysis. J Acquir Immune Defic
nal studies. HIVNET Vaccine Preparedness Study Protocol Team. Am J Epi-
Syndr 2008;48:97–103.
demiol 2000;152:99 –106.
[8] Grossman AH, D’Augelli AR. Transgender youth: Invisible and vulnerable. J
[26] Morrison-Beedy D, Carey MP, Tu X. Accuracy of audio computer-assisted
Homosex 2006;51:111–28.
self-interviewing (ACASI) and self-administered questionnaires for the as-
[9] Mayer KH, Bradford JB, Makadon HJ, et al. Sexual and gender minority
sessment of sexual behavior. AIDS Behav 2006;10:541–52.
health: What we know and what needs to be done. Am J Public Health
[27] Connors GJT, Scott J, Miller WR. Measure of religious background and
2008;98:989 –95. behavior for use in behavior change research. Psych Add Behav 1996;10:
[10] Turner-Musa J, Lipscomb L. Spirituality and social support on health behav- 90 – 6.
iors of African American undergraduates. Am J Health Behav 2007;31:495– [28] Leigh BC, Stall R. Substance use and risky sexual behavior for exposure to
501. HIV: Issues in methodology, interpretation, and prevention. Am Psychol
[11] Brown TL, Parks GS, Zimmerman RS, Phillips CM. The role of religion in 1993;48:1035– 45.
predicting adolescent alcohol use and problem drinking. J Stud Alcohol [29] Akaike H. Data analysis by statistical models. No To Hattatsu 1992;24:
2001;62:696 –705. 127–33.
[12] Sinha JW, Cnaan RA, Gelles RJ. Adolescent risk behaviors and religion: [30] STATA Data Analysis and Statistical Software. College Station, TX: Stata Corp
Findings from a national study. J Adolesc 2007;30:231– 49. LP, 2009.
[13] Rostosky SS, Regnerus MD, Wright ML. Coital debut: The role of religiosity [31] Hill PC, Pargament KI. Advances in the conceptualization and measurement
and sex attitudes in the Add Health Survey. J Sex Res 2003;40:358 – 67. of religion and spirituality. Implications for physical and mental health
[14] Scott LD Jr, Munson MR, McMillen JC, et al. Religious involvement and its research. Am Psychol 2003;58:64 –74.
association to risk behaviors among older youth in foster care. Am J Com- [32] Ellison CG, Levin JS. The religion-health connection: Evidence, theory, and
munity Psychol 2006;38:223–36. future directions. Health Educ Behav 1998;25:700 –20.