University of Southern California Dissertations and Theses

Not just talk: observed communication in adolescent friendship and its implications for health risk behavior

(USC Thesis Other)

Not just talk: observed communication in adolescent friendship and its implications for health risk behavior

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content i

Not Just Talk: Observed Communication in Adolescent Friendship
and Its Implications for Health Risk Behavior

Estibaliz Iturralde, M.A.
Doctoral Dissertation
May 2015

Dissertation Committee
Gayla Margolin, Ph.D., Chair
Julie A. Cederbaum, Ph.D.
Steven R. Lopez, Ph.D.
Franklin R. Manis, Ph.D.
Shrikanth Narayanan, Ph.D.
!
! ii!
Table of Contents

Acknowledgments 1

General Introduction 2
References 5

Research Papers
Close Adolescent Friends’ Health-Related Talk: A Dyadic Analysis 9
Abstract 10
Background 11
Close Friends’ Communication and Health Behavior 12
The Present Study 15
Method 17
Participants 17
Procedures 18
Measures 20
Analytic Approach 22
Results 24
Validity of the Discussion Task 24
Dyadic Confirmatory Factor Analysis of Talk Scores 24
Talk and Health Descriptives 25
Associations Between Talk Variables and Health Indices 28
Discussion 30
Limitations and Strengths 34
Clinical Implications and Summary 35
References 37
Table 1 46
Table 2 47
Table 3 48
Figure 1 49
Figure 2 50

Emotional Disclosure and Support Between Close Adolescent Friends: 51
Does Intimacy Promoting Talk Relate to Sexual Risk Behavior?
Abstract 52
Background 53
Current Study 55
Method 56
Participants 56
Procedures 57
Measures 58
Analytic Approach 60
Results 61
Descriptive Statistics and Bivariate Associations 61
Intimacy Promoting Talk and Pregnancy 63
Intimacy Promoting Talk and STI Risk 63
Intimacy Promoting Talk Versus Other Peer Context Variables 64
! iii!
Table of Contents (continued)

Discussion 64
References 68
Table 1 76
Table 2 77
Table 3 78
Table 4 79

General Discussion and Conclusions 80
Health-Related Talk 81
Intimacy Promoting Talk 82
Clinical Implications 84
Limitations and Future Directions 85
Conclusions 86
References 88

Appendix A. Discussion Topics 93
Appendix B. Peer Discussion Task Script 94
Appendix C. Post-Discussion Questionnaire 97
Appendix D. Coding Manual Excerpts 98
Appendix E. Coder Sheet 109
Appendix F. Substance Use Items 110
Appendix G. Peer Norm Items 112
Appendix H. Grades Item 114
Appendix I. Sexual Risk Behavior Items 115
Appendix J. Quantitative Approach to Dyadic Data 117
Appendix K. CFA Mplus Code 122
1
Acknowledgments
Data collection and preparation of these two manuscripts were supported by grant
number F31HD069147 from the Eunice Kennedy Shriver National Institute Of Child Health &
Human Development awarded to Estibaliz Iturralde and grant number NIH-NICHD R01
HD046807 awarded to Gayla Margolin. Current and past members of The USC Family Studies
Project team played an integral role in collecting and organizing the data used in these studies,
including (in alphabetical order): Brian Baucom, Diana Bennett, Larissa Borofsky, Claire
Burgess, Sarah Duman, Elyse Guran, Kristene Hossepian, Ilana Kellerman, Kelly Miller,
Michelle Ramos, Aubrey Rodriguez, Darby Saxbe, Lauren Spies Shapiro, Adela Timmons, and
Katrina Vickerman. Our diligent coders spent more than 200 hours watching adolescent friends
talk to each other; thank you Claire Burgess, Arielle Gillman, Julie Guerin, Kristene Hossepian,
Angela Izmirian, Alexandra Martinez, Tina Merlino, Isaac Rottman, and especially Sandra
Elmgren and Jennifer Wong for their key role in piloting the coding system. Gratitude also goes
to the youth participating in the USC Family Studies Project for their openness and willingness
to engage in this research. I would also like to thank Julie Cederbaum, Steve Lopez, Frank
Manis, and Shri Narayanan for their helpful comments and guidance during this project.
I am indebted to my research advisor Gayla Margolin, whose patience and generosity of
spirit have helped me time and again, even beyond her role as an incisive thinker and mentor.
My family, friends, and schoolmates on both coasts and internationally have been an invaluable
source of support. Finally, I must thank my husband William Rogan who continues to amaze me
more each day. You have more than earned your honorary spousal Ph.D. This thesis is dedicated
to our baby girl Winifred, who arrived just in time for the final phase of its production and has
made the process of studying child and adolescent development all the more dazzling.
2
General Introduction
This dissertation is premised on the idea that what youth say to their friends and how they
say it provide meaningful information about their health and risk behavior. Adolescence is a
period of increased autonomy, which presents opportunities and challenges as youth make the
transition to adulthood. Unintended pregnancy, sexually transmitted infections, and alcohol,
tobacco, and marijuana use are prevalent in adolescence, are often inter-correlated, and pose
dangers to long-term health and safety (Casares, Lahiff, Eskenazi, & Halpern-Felsher, 2010;
Finer, 2010; Forhan et al., 2009; Jessor & Jessor, 1977; Johnston, O’Malley, Miech, Bachman, &
Schulenberg, 2014; McCambridge, McAlaney, & Rowe, 2011; Meier et al., 2012). The
increasing prominence of peer relationships in adolescence also presents opportunities and
challenges. An ecological model of adolescent development posits that close relationships, such
as those with friends, shape individuals’ health-related behavior (Bronfenbrenner, 1979). Often
mentioned but poorly understood in the adolescent risk behavior literature is the exchange of
information and influence between peers on health issues. Primary prevention programs have
increasingly focused on addressing peer group-based norms and interpersonal behaviors
(Jackson, Geddes, Haw, & Frank, 2012; Wolfe, Crooks, Chiodo, Hughes, & Ellis, 2012), but
more investigation is needed to understand the peer context of health behavior.
The two studies presented here seek to examine the role of peer talk in adolescent health
and risk behavior by using direct observation of youth aged 15 to 20. A community sample of
111 ethnically diverse, male and female adolescents and their gender-matched close friends
participated in videotaped, ecologically valid discussions of topics important to their lives.
Topics included interpersonal problems, substance use beliefs, personal goals, romantic
relationships, and friendships, and yielded a remarkably candid behavioral sample of how
3
adolescents talk to close friends. A coding system was developed specifically for this work,
allowing for measurement of communication processes with significance for health behavior.
Youth also responded to questionnaires regarding their current substance use, their sexual risk
history, their school grades, and the norms of their peer group. Body mass index was assessed in-
lab as an additional indicator of health-compromising behavior.
One major aim of the current research was to investigate how youth discuss health
concepts with a friend, with attention to both risk promoting and health promoting aspects of
talk. In studies that have examined observed behavior, the adolescent health literature has
historically focused on risky forms of peer influence (Allen, Porter, & McFarland, 2006; Cohen
& Prinstein, 2006; Dishion & Tipsord, 2011). However, other research has suggested that friends
attempt to promote healthy behaviors among one another (Brady, Morrell, Song, & Halpern-
Felsher, 2013; Buckley, Chapman, Sheehan, & Reveruzzi, 2014). Evidence suggests that health
beliefs play a role in a variety of health and risk domains (Wiefferink et al., 2006) and that
speaking aloud one’s intentions appears to increase likelihood of adopting healthier practices
(Miller & Rollnick, 2013), yet there has been little investigation of how health beliefs and other
cognitions are co-constructed in the peer group. It was hypothesized that health-related talk
would fit a two-factor model of risk promotion and health promotion constructs, and that these
factors would map onto actual health and risk behavior.
A second aim of this project was to examine the role of mutual, emotionally supportive
communication between friends as it relates to sexual risk behavior. Development of intimate
relationships with peers is an important developmental task of adolescence (Collins & Sroufe,
1999). In their coded observations of youth, Gottman and Mettetal (1986) found that, compared
to younger children, adolescents engage in more self-disclosure and “mindreading” (voicing of
4
the peer’s perspective), and that these behaviors distinguish friend dyads from those who are
strangers. There is some research suggesting that intimate friendship processes have a protective
role in the domain of adolescent sexual risk behavior (Elkington, Bauermeister, & Zimmerman,
2011; Lando-King et al., in press; Miller, Notaro, & Zimmerman, 2002). However, contrasting
associations have also been found (Benda & Corwyn, 1999; Salazar et al., 2007), and this
question has not been previously studied using behavioral observation methods. The current
research, therefore, investigated youths’ intimacy promoting talk, which was defined as
reciprocal use of emotional disclosure and empathic support. Intimacy promoting talk between
friends was hypothesized to have an inverse (protective) association with sexual risk behavior.
The present research, in addition to its rare application of behavioral observation to these
questions, was also unique in its use of multiple levels of data to shed light on adolescents’ peer
relationships and health behavior. Multilevel analyses allowed for the simultaneous
consideration of individual- and dyad-level phenomena, including measures assessed directly
with both members of the participating friend pairs. In keeping with an ecological perspective,
analyses also considered the role of contextual variables including risky and prosocial peer
norms, the risk behavior of the co-participating friend, neighborhood poverty, and individual
background characteristics such as gender, age, and race/ethnicity. Taken together, these two
studies provide a novel glimpse into peer processes that may be a source of risk as well as
resilience in adolescents’ lives.

5
References for General Introduction
Allen, J. P., Porter, M. R., & McFarland, F. C. (2006). Leaders and followers in adolescent close
friendships: Susceptibility to peer influence as a predictor of risky behavior, friendship
instability, and depression. Development and Psychopathology, 18(01), 155-172.
doi:10.10170S0954579406060093
Benda, B. B., & Corwyn, R. F. (1999). Developmental differences in theories of sexual behavior
among rural adolescents residing in AFDC families. Deviant Behavior, 20(4), 359-385. doi:
10.1080/016396299266452
Brady, S. S., Morrell, H. E., Song, A. V., & Halpern-Felsher, B. L. (2013). Longitudinal study of
adolescents’ attempts to promote and deter friends’ smoking behavior. Journal of
Adolescent Health, 53(6), 772-777. doi:10.1016/j.jadohealth.2013.06.022
Bronfenbrenner, U. (1979). The ecology of human development. Cambridge, MA: Harvard
University Press.
Buckley, L., Chapman, R. L., Sheehan, M. C., & Reveruzzi, B. N. (2014). In their own words:
Adolescents’ strategies to prevent friends’ risk taking. The Journal of Early Adolescence,
34(4), 539-561. doi:10.1177/0272431613496637
Casares, W. N., Lahiff, M., Eskenazi, B., & Halpern-Felsher, B. L. (2010). Unpredicted
trajectories: The relationship between race/ethnicity, pregnancy during adolescence, and
young women’s outcomes. Journal of Adolescent Health, 47(2), 143-150.
doi:10.1016/j.jadohealth.2010.01.013
Cohen, G. L., & Prinstein, M. J. (2006). Peer contagion of aggression and health risk behavior
among adolescent males: An experimental investigation of effects on public conduct and
private attitudes. Child Development, 77(4), 967-983. doi:10.1111/j.1467-
6
8624.2006.00913.x
Collins, W. A., & Sroufe, L. A. (1999). Capacity for intimate relationships: A developmental
construction. In W. Furman, B. B. Brown, & C. Feiring (Eds.), The development of
romantic relationships in adolescence (pp. 125-147). New York, NY: Cambridge
University Press.
Dishion, T. J., & Tipsord, J. M. (2011). Peer contagion in child and adolescent social and
emotional development. Annual Review of Psychology, 62, 189-214.
doi:10.1146/annurev.psych.093008.100412
Elkington, K. S., Bauermeister, J. A., & Zimmerman, M. A. (2011). Do parents and peers
matter? A prospective socio-ecological examination of substance use and sexual risk
among African American youth. Journal of Adolescence, 34(5), 1035-1047.
doi:10.1016/j.adolescence.2010.11.004
Finer, L. B. (2010). Unintended pregnancy among U.S. adolescents: Accounting for sexual
activity. Journal of Adolescent Health, 47(3), 312-314.
doi:10.1016/j.jadohealth.2010.02.002
Forhan, S. E., Gottlieb, S. L., Sternberg, M. R., Xu, F., Datta, S. D., McQuillan, G. M., …
Markowitz, L. E. (2009). Prevalence of sexually transmitted infections among female
adolescents aged 14 to 19 in the United States. Pediatrics, 124(6), 1505-1512.
doi:10.1542/peds.2009-0674
Gottman, J. M., & Mettetal, G. (1986). Speculations about social and affective development:
Friendship and acquaintanceship through adolescence. In J. M. Gottman & J. G. Parker
(Eds.), Conversations of friends: Speculations on affective development (pp. 192-237).
New York, NY: Cambridge University Press.
7
Jackson, C., Geddes, R., Haw, S., & Frank, J. (2012). Interventions to prevent substance use and
risky sexual behaviour in young people: A systematic review. Addiction, 107(4), 733-747.
doi:10.1111/j.1360-0443.2011.03751.x
Jessor, R., & Jessor, S. L. (1977). Problem behavior and psychosocial development—A
longitudinal study of youth. New York, NY: Academic Press.
Johnston, L. D., O’Malley, P. M., Miech, R. A., Bachman, J. G., & Schulenberg, J. E. (2014).
Monitoring the Future national results on drug use: 1975-2013: Overview, key findings on
adolescent drug use. Ann Arbor, MI: Institute for Social Research, The University of
Michigan.
Lando-King, E. A., McRee, A. L., Gower, A. L., Shlafer, R. J., McMorris, B. J., Pettingell, S., &
Sieving, R. E. (in press). Relationships between social-emotional intelligence and sexual
risk behaviors in adolescent girls. Journal of Sex Research.
doi:10.1080/00224499.2014.976782
McCambridge, J., McAlaney, J., & Rowe, R. (2011). Adult consequences of late adolescent
alcohol consumption: A systematic review of cohort studies. PLoS Medicine, 8(2),
e1000413. doi:10.1371/journal.pmed.1000413
Meier, M. H., Caspi, A., Ambler, A., Harrington, H., Houts, R., Keefe, R. S., … Moffitt, T. E.
(2012). Persistent cannabis users show neuropsychological decline from childhood to
midlife. Proceedings of the National Academy of Sciences, 109(40), E2657-E2664.
doi:10.1073/pnas.1206820109
Miller, A. L., Notaro, P. C., & Zimmerman, M. A. (2002). Stability and change in internal
working models of friendship: Associations with multiple domains of urban adolescent
functioning. Journal of Social and Personal Relationships, 19(2), 233-259.
8
doi:10.1177/0265407502192004
Miller, W. R., & Rollnick, S. (2013). Motivational interviewing: Helping people change (3rd
ed.). New York, NY: Guilford Press.
Salazar, L. F., Crosby, R. A., Diclemente, R. J., Wingood, G. M., Rose, E., Sales, J. M., &
Caliendo, A. M. (2007). Personal, relational, and peer-level risk factors for laboratory
confirmed STD prevalence among low-income African American adolescent females.
Sexually Transmitted Diseases, 34(10), 761-766. doi:10.1097/01.olq.0000264496.94135.ac
Wiefferink, C. H., Peters, L., Hoekstra, F., Dam, G. T., Buijs, G. J., & Paulussen, T. G. (2006).
Clustering of health-related behaviors and their determinants: Possible consequences for
school health interventions. Prevention Science, 7(2), 127-149. doi:10.1007/s11121-005-
0021-2
Wolfe, D. A., Crooks, C. V., Chiodo, D., Hughes, R., & Ellis, W. (2012). Observations of
adolescent peer resistance skills following a classroom-based healthy relationship program:
A post-intervention comparison. Prevention Science, 13(2), 196-205. doi:10.1007/s11121-
011-0256-z

DYADIC HEALTH-RELATED TALK 9

Close Adolescent Friends’ Health-Related Talk: A Dyadic Analysis

Esti Iturralde
University of Southern California

DYADIC HEALTH-RELATED TALK 10
Abstract
The current study sought to examine how adolescents talk to a close friend about health-
related beliefs and how both risk promoting and health promoting talk relate to actual health and
functioning. A coding system was developed to assess the talk behavior observed among 111
pairs of gender-matched close friends (100 females and 122 males total, aged 15-20) during a 30-
minute dyadic, multi-topic discussion. Dyadic confirmatory factor analysis was used to assess
the presence of underlying risk promotion and health promotion factors. Health indices included
past-month substance use, past pregnancy involvement, overweight/obesity, having good grades,
and affiliation with risky or prosocial peers. Talk data fit the hypothesized two dyadic factor
model. One additional behavior, disgust talk (e.g., “cigarettes are gross”) did not load strongly on
either factor. As expected, risk promoting talk was generally related in a positive direction to
substance use and risky peers, but showed null or unexpected associations with other indices.
Consistent with hypotheses, health promoting talk was generally associated negatively with risk
indices (binge drinking, cigarettes, overweight/obesity for females) and positively with adaptive
functioning indices (good grades, prosocial peers for females). Disgust talk was negatively
related to cigarette smoking, marijuana use, and for males, risky peer affiliation, but was
unrelated to other indices. Health-related talk appears closely aligned with youths’ actual health
and functioning, but is seldom measured in studies of adolescent health or risk behavior. Further
assessment of both health promoting and risk promoting communication may illuminate the role
of peer processes in health decision-making.
Keywords: adolescent risk behavior, peer relationships, deviancy training, health
communication, behavioral observation, dyadic analysis, social learning, positive youth
development
DYADIC HEALTH-RELATED TALK 11
Close Adolescent Friends’ Health-Related Talk: A Dyadic Analysis
Adolescence is a key stage in which to shape attitudes and behavior on a wide range of
health issues. Health behavior patterns initiated in adolescence have lifelong impacts on
morbidity and mortality (Williams, Holmbeck, & Greenley, 2002). Experimentation with risk
behavior, such as drinking alcohol and use of tobacco or marijuana, becomes more frequent
during this time period and can set the stage for patterns of abuse and long-lasting threats to
health and safety (Johnston, O’Malley, Miech, Bachman, & Schulenberg, 2014; McCambridge,
McAlaney, & Rowe, 2011; Meier et al., 2012). Substance use is likely to coincide with other risk
behaviors, acquisition of poor health habits, endorsement of risk-oriented health beliefs, and
social determinants of risk such as academic problems and affiliation with risky peers (Jessor &
Jessor, 1977; León, Carmona, & García, 2010; Peters et al., 2009). Likewise, youth who have
low involvement in risk behavior tend to hold more health-oriented values and concerns, take
better care of their health, perform better in school, and affiliate with prosocial peers (Costa,
Jessor, Fortenberry, & Donovan, 1996; Jessor, Turbin, & Costa, 1998; León et al., 2010;
Rosengard et al., 2001). Although some youth fit one pattern more closely than the other, a mix
of protective and risk factors are present for many youth, suggesting the value of
multidimensional models incorporating both facets (Kia-Keating, Dowdy, Morgan, & Noam,
2011; R⊘ysamb, Rse, & Kraft, 1997).
Peers appear to play an important role in health decision-making. Social learning theory
emphasizes the particular impacts of observing others’ behavior (Bandura, 1977), and
adolescents are known to be particularly swayed by their peers’ behavior and to take more risks
in their peers’ presence (Gardner & Steinberg, 2005). Unsurprisingly, a strong predictor of
adolescents’ risk behavior are the norms, i.e., attitudes, preferences, and actual behavior, of their
DYADIC HEALTH-RELATED TALK 12
peer group (Wiefferink et al., 2006). What is less understood is how these norms are
communicated in everyday language among adolescents and their friends. Communication
among peers is itself a kind of behavior that can model and reinforce risky ideation, which can
then be predictive of risky activities (Dishion, Spracklen, Andrews, & Patterson, 1996). Many
studies addressing the role of peer norms do not measure communication processes per se,
despite theoretical and some empirical evidence of its importance as a mechanism (Lapinski &
Rimal, 2005; Real and Rimal, 2007). Closer study of peer communication in the context of
health and risk behavior may be beneficial given that many interventions with adolescents
involve a peer relational dimension, for example, skill-building for avoiding risk behavior in peer
situations or the use of peer leaders to disseminate health information (Jackson, Geddes, Haw, &
Frank, 2012; Wolfe, Crooks, Chiodo, Hughes, & Ellis, 2012). The current study sought to
examine how adolescents talk to a close friend about health-related beliefs and how both risk
promoting and health promoting talk relate to actual health and functioning.
Close Friends’ Communication and Health Behavior
Close friendship offers a unique context in which social learning processes can unfold.
Adolescents gravitate towards friends who are similar to them (Brechwald & Prinstein, 2011),
but also, the emotional rewards of friendship create more salient interactions, which may
intensify the effects of modeling. Close friendship is noteworthy for its intimate and reciprocal
nature (Newcomb & Bagwell, 1995). Friendship quality strengthens conformity with close
friends’ risk behavior (Urberg, Luo, Pilgrim, & Degirmencioglu, 2003), and, close friends’
(versus general peers’) behavior is a stronger predictor of youths’ own risk behavior (Ali &
Dwyer, 2011; Simons-Morton & Farhat, 2010; Yanovitzky, Stewart, & Lederman, 2006). From a
methodological standpoint, it is useful for researchers to observe how close friends talk to one
DYADIC HEALTH-RELATED TALK 13
another because the high degree of intimacy and similarity between them make it more likely
that what they say will be honest and will relate meaningfully to their actual behavior.
This closeness may also highlight particular kinds of talk that could be seen as health
promoting. Health promoting talk may not be received favorably within the wider peer context,
but may be more acceptable among close friends. Based on social network theory, close friends
are seen as more mutually available and interested in helping one another compared to
acquaintances (Granovetter, 1983). Adolescents themselves say that they are more likely to
intervene with peers’ risk behavior when they share a close friendship, and that a primary
approach they would take would be through talking (Buckley, Chapman, Sheehan, & Reveruzzi,
2014). Youth also say that talking to friends is a primary strategy for reducing their own risk
behavior (Metrik, Frissell, McCarthy, D’Amico, & Brown, 2003). However, little work has
examined how youth talk to each other about these issues.
Risk promoting talk. So far, research using behavioral observation of youths’ health-
related communication has largely focused on risk promoting forms of talk. One particularly
studied phenomenon among youth with antisocial behavior is deviancy training (for a review,
see Dishion & Tipsord, 2011). Such youth use more rule-breaking talk when conversing with a
close friend relative to youth without antisocial behavior. This rule-breaking talk is reciprocally
exchanged among friends, consists of deviant suggestions and attitudes that violate the law or
societal norms, and is often reinforced through positive affect cues. Deviancy training has been
observed during laboratory discussion tasks structured around topics that are both related and
ostensibly unrelated to risk behavior (e.g., problem-solving; see Dishion et al., 1996; Piehler &
Dishion, 2007). Deviancy training predicts growth of risk behavior over time, suggesting that it
is one mechanism by which problem behavior escalates among youth (Patterson, Dishion, &
DYADIC HEALTH-RELATED TALK 14
Yoerger, 2000). Much of the deviancy training work has focused on clinical, male, and
Caucasian samples; it is not clear how influential this process may be in a more typical and
diverse sample. Given the range of risk level that was expected in the current study, we focused
on risk promoting talk that was specific to substances (positive substance talk) since substance
use is common in this age group. Communication behavior that appeared to reinforce risky
beliefs (risk-positive responses) was also assessed.
Health promoting talk. There is evidence that adolescents do attempt to dissuade their
friends from engaging in health risk behavior. For example, a recent prospective study found that
adolescents were actually more likely to discourage rather than encourage smoking among their
friends; interestingly, more discouragement was voiced by youth with past smoking history
versus none (Brady, Morrell, Song, & Halpern-Felsher, 2013), a finding that concurs with
skeptical views of the assumed centrality of peer pressure (Arnett, 2007). One strategy used by
youth when discouraging friends from risk behavior is to focus on health issues. Qualitative
work describes attempts by friends to deter each other from risk behavior by discussing
perceived harms of the behavior and the importance of adhering to individual values (Buckley et
al., 2014; Nichter, Vuckovic, Quintero, & Ritenbaugh, 1997). In line with this work, the current
study assessed two types of potentially protective health-focused talk, cautious health talk, which
consists of statements emphasizing the benefits of avoiding risk behavior based on individual
values, and health promoting responses, in which youth express approval of a friends’ health-
cautious statements or disapproval of risky ones.
Disgust, i.e., the “gross” smell, taste, or other sensory experience associated with risk
behavior, is a frequently discussed reason among youth as to why a behavior should be avoided;
disgust talk has been prominent in studies of tobacco use in particular (McLeod et al., 2008;
DYADIC HEALTH-RELATED TALK 15
Nichter et al., 1997). Affect-oriented theories of cognition and behavior emphasize the role of
strong emotional reactions, including negative ones such as disgust, in guiding action, i.e., by
conditioning an aversive association with the activity and related social situations (Albert &
Steinberg, 2011). It is not known how disgust talk in a dyadic context may be related to health or
functioning. The current study, therefore, measured youths’ disgust talk related to risk behavior
to allow for examination of how this communication behavior appears in a dyadic context and
alongside other health-related talk.
The positive youth development perspective posits the importance of health promoting
behavior that capitalizes on developmentally normative aspects of adolescence (Catalano,
Hawkins, Berglund, Pollard, & Arthur, 2002). Not all protective processes are the direct inverse
of risky ones; some may satisfy alternative objectives. Bergin and colleagues (2003) found that
adolescents themselves characterized prosocial peers as not only those engaged in overtly helpful
behaviors, but also those who contributed to positive relational experiences, such as through
humor and play. Talk that is interpersonally oriented but not focused on health or risk per se may
be both socially appropriate and relevant to health. Therefore, in this study youth were assessed
in their use of interpersonal value talk, which consisted of statements of appreciation for peers in
terms of positive relational characteristics that did not involve risk issues. Of interest was how
this talk related to other, more overt kinds of health-oriented talk, as well as to youths’ actual
health behavior.
The Present Study
The present study sought to examine health-related talk among close adolescent friends
through use of behavioral observation, and to assess how this talk related to a broad selection of
health and risk indices. In keeping with past research examining multiple domains of adolescent
DYADIC HEALTH-RELATED TALK 16
health and risk behavior (León, et al., 2010; Turbin, Jessor, & Costa, 2000; R⊘ysamb, et al.,
1997), a broad array of health indices were selected, encompassing not only risk behavior such
as substance use and pregnancy, but also other health-compromising behavior (obesity). Past
work has found overlap among variables predicting both risk behavior and health-compromising
behavior among adolescents (Peters et al., 2009). Furthermore, an efficient approach to
intervention design relies on targeting processes that are common to multiple health issues, rather
than approaching each health or risk category separately (Wiefferink et al., 2006). Most indices
examined here correspond to current public health objectives for improved adolescent and young
adult health (Healthy People 2020, 2014), specifically past-month binge-drinking, cigarette
smoking, and marijuana use, pregnancy involvement, and being obese or overweight. Based on
prior research on predictors of risk behavior, we also assessed several social determinants of
health: school achievement and affiliation with risky and prosocial friends (Kia-Keating et al.,
2011).
A coding system was developed to measure distinct talk behaviors employed by friends.
Dyadic confirmatory factor analysis (Kenny, Kashy, & Cook, 2006) was used to test the
underlying latent structure of these individual-level codes with an a priori model in mind. It was
hypothesized that the data would best fit a two-factor model with inversely correlated factors of
risk promotion and health promotion. Hypothesized risk promotion indicators were talk
behaviors that favored substance use or encouraged risk-oriented talk in the friend. Health
promotion indicators, on the other hand, were hypothesized to be multi-faceted, comprised of
risk-specific talk and talk about normative interpersonal interests. An alternative single-factor
model was tested, as was pattern invariance across gender for factor loadings and correlations. It
was also hypothesized that risk promoting talk would be positively associated with substance
DYADIC HEALTH-RELATED TALK 17
use, pregnancy, overweight, and risky peer affiliation, and negatively associated with good
grades and prosocial peer affiliation, whereas health promoting talk would have associations in
the opposite directions.
Method
Participants
The current study included data from 111 pairs of close adolescent friends (mean age =
17.62; SD = 1.32; 61 male and 50 female dyads) who were participants in a larger multi-wave
longitudinal study of community families drawn from a major metropolitan area in the western
United States. Families were recruited through newspaper advertising, flyers, and word-of-mouth
to take part in a study of youth functioning. Inclusion criteria at the initial wave of data collection
with families were the following: the mother, father, and one child from each family had to be
able to participate; parents had to have 3 or more years living together with their child; and,
participants had to be able to complete study measures in English. The current data were
collected when original study youth were aged 15 to 20.
In the present study, each original study youth recruited a close, gender-matched friend of
similar age to participate. About 88% of participants rated the friend as one of my 5 closest or
closer; 80.2% had known the friend for 3 or more years; 51.4% of dyads reported being the same
age whereas another 45.0% indicated 1 to 2 years of age difference. The current sample excluded
20 original study youth who, despite participating in the study wave, were not eligible because:
15 were not able to arrange for a close friend to participate; 4 submitted questionnaire data
electronically from home but did not participate in the lab discussion task; and, 1 refused to be
recorded in the lab discussion task.
Participants were ethnically and socioeconomically diverse. Of the total 222 participating
youth, 31.5% were non-Hispanic/Latino Caucasian, 31.5% indicated primarily Hispanic/Latino
DYADIC HEALTH-RELATED TALK 18
descent, 25.2% identified with African-American or mixed African-American race/ethnicity,
5.0% identified as Asian or Pacific Islander, and 6.8% were of another race/ethnicity. Youth
varied considerably on a zip code-based measure of neighborhood poverty; they ranged from
having 0 to 46.1% of families in their home neighborhood living under the federal poverty level
(median = 8.5%).
Procedures
Participating friends visited the laboratory as a pair. Prior to the visit, for those youth
under age 18, study personnel telephoned the youth’s parent or guardian to obtain consent. A
consent procedure was also conducted with the participants at the beginning of the visit. Due to
the sensitive nature of the health data collected for the study, youth were informed that their
answers on self-report questionnaire measures would not be shared with their families or the
friend joining them on the lab visit. The friends were seated at workstations in separate rooms to
be given privacy to complete computer-administered questionnaires.
After completing questionnaires, the friends were reunited for the discussion task, a
modified version of that used by the Oregon Youth Study (Capaldi & Clark, 1998; Dishion,
Andrews, & Crosby, 1995). Youth were asked to engage in a discussion on multiple topics, each
lasting 5 minutes. Youth were reminded that information they disclosed would not be shared
with family members. The experimenter left the room after reading each of 6 topics and the
interaction was digitally recorded.
Discussion topics were devised to elicit a range of talk typical of adolescents. In Topics 1
and 2, youth took turns discussing a current problem with a person close to them. In Topic 3,
youth discussed their beliefs about the use of alcohol, tobacco, marijuana, and “other drugs.”
Topic 4 concerned youths’ major individual goals for the next year. Topic 5 was on dating,
DYADIC HEALTH-RELATED TALK 19
whereas Topic 6 concerned opinions about their peer group. Three subsequent topics were not
analyzed in the present study. All study procedures were approved by the host university’s
Institutional Review Board.
The Friend Talk coding system was developed for the larger research project to assess
talk content and behavior during the discussion task. The system drew from some of the content
definitions featured in the micro-social Topic Code system (Peterson, Piehler, & Dishion, 2006)
but was designed to assess a wide variety of talk not examined by previous studies. After a
review of the literature, an initial coding manual was developed. A macro-coding strategy was
employed. Each behavior was rated for each individual dyad member per 5-minute topic interval.
A random 5% of the discussion recordings were selected to pilot the coding system. Two
research assistants were trained in the initial system and, working separately, generated coded
and qualitative observations from the pilot recordings. Once a criterion level of reliability for
each code was reached (intra-class correlation ≥ .80), the coding manual was finalized.
A small group of research assistants were trained in the system. Each interaction was
coded by two raters working independently who were not familiar with participants’
questionnaire ratings of health behavior. Raters watched each 5-minute discussion segment
twice. During the first viewing, raters assigned scores to one youth for that segment. They then
rewound the segment, watched again, and assigned scores to the second youth. Occasionally,
more challenging segments were watched a third time to help raters make scoring decisions.
Raters switched the sequencing of youth to be scored within each 5-minute segment. The initial
youth to be scored during the first segment was counterbalanced across raters.
Except where specified, scores took into account both the frequency and intensity of the
observed behavior, were made on a 0 to 3 scale (0 = none, 1 = some, 2 = a moderate amount, 3 =
DYADIC HEALTH-RELATED TALK 20
a lot), and were averaged across topic intervals to produce individual-level scores. Eight codes
were used in the present study. Table 1 presents examples of talk behaviors and inter-rater
reliabilities (intraclass correlation using summary scores).
Measures
Discussion validity. Raters scored each youth’s level of responsiveness during the
discussion using the engagement code, based on cues such as verbal assent and facilitation, eye
contact, and body posture. A second code, faking, was used if the youth overwhelmingly used
artificial speech or behavior. Unlike the other codes, faking was scored as either present (1) or
not present (0) per 5-minute-interval; these ratings were summed across intervals to generate a
faking score. Youth themselves also rated discussion validity via two items: How honest or frank
were you? and, How similar was this discussion to other discussions you have had with your
friend? Responses were made on a 5-point scale (0 = not at all, 1 = slightly, 2 = somewhat, 3 =
moderately, 4 = very).
Risk promotion talk. Two codes were used. Positive substance talk was coded when
youth expressed positive attitudes towards the use of alcohol, tobacco, marijuana, or other
substances. Statements could offer rationalizations for why substance use was acceptable for
themselves or peers and omitted ideas about safety or self-restraint. The risk-positive response
code was used when youth reacted to friends’ positive talk about risk behavior (whether with
substances or otherwise) with favorable or encouraging statements. Risk-positive responses
could also include positive affect cues such as smiling or laughing. Scores for both codes were
log-transformed prior to data reduction to address right skew.
Health promotion talk. Four codes were used. Cautious health talk included statements
favoring the avoidance of risk behavior based on health beliefs and personal values. If youth
DYADIC HEALTH-RELATED TALK 21
disapproved of risk behavior because of negative sensory aspects, a code for disgust talk was
applied instead. The health promoting response code was the average of one score rating positive
responses to health-positive talk and another score rating negative responses to risk-positive talk,
thus capturing talk that encouraged friends’ healthier choices. Such responses could be verbal or
non-verbal. Interpersonal value talk was scored when youth spoke favorably of peer
relationships based on relationship-oriented characteristics such as shared interests, positive
interpersonal dynamics, or similar goals, and without reference to any risk behavior. All codes
were normally distributed except for disgust talk, which was log-transformed prior to data
reduction due to right skew.
Health indices. A broad array of measures were selected to assess risk behavior, health,
and related social determinants, and were dichotomized based on a criterion value. Past-month
substance use items were adapted from the CDC Youth Surveillance Questionnaire (Eaton et al.,
2008), which assessed days of binge drinking (consuming 5 or more drinks), days of smoking
cigarettes, and times using marijuana (0 = none; 1 = any). Past pregnancy involvement was
assessed via one item inquiring if the youth had ever been pregnant or caused someone to
become pregnant, as appropriate to gender (0 = no, 1 = yes). Based on height and weight
measured after the peer discussion, youth were coded as overweight or obese if their body mass
index for age (in months) was greater than or equal to 85
th
percentile based on gender-stratified
CDC growth charts (Kuczmarski et al., 2002).
Having risky friends and having prosocial friends was assessed using items adapted from
the Peer Behavior Inventory (Prinstein, Boergers, & Spirito, 2001). Items used a 0 (none) to 4
(almost all) scale. The 7 items for risky friends assessed how many of youths’ closest friends
engage in substance use, unprotected sex, and delinquency (Cronbach’s α = .87). The 12
DYADIC HEALTH-RELATED TALK 22
prosocial items tapped the helpfulness of friends, such as by assisting peers, working, or doing
chores (α = .83). The mean of each scale was taken and then dichotomized using a cut-point of 3,
signifying that more than half of friends engaged in the given behaviors on average.
Youth were asked with 1 item to characterize their most recent school grades. Having
good grades was coded as 1 (mostly Bs or better) and 0 (no better than Bs and Cs).
Demographic variables. Gender was effect coded as .5 for male and -.5 for female.
Whole number age was used as a covariate. Neighborhood poverty was the percentage of
families living below the federal poverty level in the youth’s home neighborhood, based on the
zip code-indexed dataset of the American Community Survey (U.S. Census Bureau, 2011);
percentages were log-transformed for correlation analyses to mitigate right skew. In regression
analyses, race/ethnicity was represented through weighted coding of three variables: Caucasian,
Latino, and African-American, with all other ethnicities treated as the reference group (see
Cohen, Cohen, West, & Aiken, 2003).
Analytic Approach
Means, standard deviations, and counts were computed for study variables. Intra-dyad
correlations were estimated as a measure of similarity between dyad members. As individual
participants’ data were clustered within dyads, analyses took into account this non-independence
(Kenny et al., 2006). An added consideration was that the dyads were theoretically
“indistinguishable,” i.e., dyad members could not be ordered by a structural variable such as
gender, age, or role (Griffin & Gonzalez, 1995). If unaddressed, indistinguishability can bias
estimates in correlational analysis because the ordering of dyad members is arbitrary.
Per guidelines posed by Olsen and Kenny (2006) and Peugh and colleagues (2013), a
dyadic confirmatory factor analysis (CFA) was estimated using a dyad-level data structure. Due
DYADIC HEALTH-RELATED TALK 23
to high intra-dyad correlations among the talk variables, dyad-level factors were modeled using
individual-level indicators (as in a common fate model; see Peugh et al., 2013). Non-
independence was resolved through the addition of equality constraints for paths corresponding
across dyads, whereas indistinguishability bias was corrected through an adjustment to fit indices
based on a null and saturated model. The saturated model also served as an empirical test of
indistinguishability. Additional paths were added to allow for shared intrapersonal variance (see
Peugh et al., 2013) and interpersonal covariances related to possible dependencies among codes
in the coding system (e.g., a rater might be more likely to score Youth 1 on risk-positive
response if Youth 2 used positive substance talk simply because of the nature of the coding
system). Model fit was evaluated based on commonly used criteria for adequate fit: factor
loadings ≥ .5; chi square p value ≥ .05; comparative fit index (CFI) > .95; and root mean square
error of approximation (RMSEA) ≤ .08 (Hu & Bentler, 1999; MacCallum, Browne, & Sugawara,
1996).
Mean scores based on factor indicators were calculated and used in logistic regression
analyses to test hypotheses regarding associations between talk behavior and health indices. For
these regression analyses, robust standard errors using sandwich estimation were computed to
handle non-independence (Muthén & Muthén, 2012). Regression analyses also adjusted for age,
gender, race/ethnicity, and neighborhood poverty. Intra-dyad and individual-level correlations
were computed using a pairwise, “checkerboard,” data structure (see Griffin & Gonzalez, 1995).
CFA and logistic regression analyses were performed using MPlus version 7 (Muthén & Muthén,
2012), whereas group differences and correlations were tested using SPSS version 22 (IBM
Corp., 2013).

DYADIC HEALTH-RELATED TALK 24
Results
Validity of the Discussion Task
According to youth, the discussions were remarkably candid and authentic. A total of
72.5% of youth reported being very honest during the discussion. Most youth (68.9%) rated the
discussions as moderately or very similar to other discussions they have had with their friend in
the past. Raters likewise perceived youth as being quite engaged in the discussion (mean = 2.39,
SD = .35). Of the total 666 5-minute discussion segments that were recorded, only 14 (2.1%)
involving 6 dyads were scored by at least one rater as suggestive of faking, typically due to a
sarcastic tone that suggested a lack of candidness. Raters’ assessment of faking itself appeared
valid, as this score was inversely correlated with youth-rated honesty (r = -.28, p < .001). Dyads
with faking did not substantially differ from other youth on talk behavior variables, and so data
from these dyads were maintained in analyses.
Dyadic Confirmatory Factor Analysis of Talk Scores
A saturated model was estimated to test if dyads were distinguishable based on
recruitment differences (original study youth versus invited friend). The model fit well, meaning
that youth were indistinguishable on this characteristic for the given talk variables (χ
2
(42) =
53.96, p = .10; RMSEA = .05; CFI = .98). The hypothesized model was then tested. In an initial
model (Model 1), the indicators for risk promotion were positive substance talk and risk-positive
responses, whereas indicators for health promotion were cautious health talk, health promoting
responses, interpersonal value talk, and disgust talk. Model 1 had good fit (χ
2
(20) = 27.33, p =
.13; RMSEA = .06; CFI = .99), however, disgust talk was not a strong indicator for the health
promotion factor (β = .23). Therefore, in a second model (Model 2), this variable was removed as
an indicator and allowed to correlate with the two factors and across dyad members. Model 2
DYADIC HEALTH-RELATED TALK 25
demonstrated adequate fit (χ
2
(20) = 29.45, p = .08; RMSEA = .07; CFI = .98) with strong factor
loadings (βs = .58–.85). A strong negative correlation was found between the risk promotion and
health promotion factors (r = -.50, p < .001). Disgust talk was correlated negatively with the risk
promotion factor (r = -.18, p < .05) and positively with the health promotion factor (r = .22, p <
.01). Disgust talk was also strongly intercorrelated across dyad members (r = .53, p < .001).
Two alternatives to Model 2 were also tested. First, a single-factor model with the same
indicators was estimated. This model had poor fit (χ
2
(22) = 92.36, p < .001; RMSEA = .17; CFI
= .88), suggesting that a two-factor model was superior. Additionally, a multi-group model was
estimated to test invariance across gender of the factor loadings and correlations in the two-factor
model. When the two gender groups were constrained to be equal, the model fit did not
significantly worsen from a model in which each group’s factor loadings and correlations were
free to vary (χ
2
difference
(9) = 15.27, p = .08), suggesting a similar pattern for males and females.
Thus, Model 2 was accepted as the measurement model and is shown in Figure 1.
Data reduction. Based on the confirmatory factor analysis, dyadic mean scores were
computed using the 4 scores for risk promotion (2 per dyad member) and 6 scores for health
promotion (3 per dyad member). Each resulting scale was internally consistent (risk promotion α
= .84, health promotion α =.81). In addition, a dyad-level mean score was computed for disgust
talk given its strong intra-dyad correlation. In subsequent analyses, disgust talk was also tested as
an individual-level rather than dyad-level variable; as this did not make a notable difference,
results with the dyadic score are presented.
Talk and Health Descriptives
Descriptive statistics for talk and health variables are presented in Table 2. Interpersonal
value talk and cautious health talk were used at the highest levels and by the most youth, 93.2%
DYADIC HEALTH-RELATED TALK 26
and 95.9%, respectively. Positive substance talk and health promoting responses were observed
at lower levels but were still used by 80.2% and 88.7% of youth, respectively. Less common but
still pervasive were disgust talk (73.9% of dyads) and risk-positive responses (53.6% of
individual youth).
With regards to health variables, a minority of youth endorsed a criterion level of risk
behavior. The substance and overweight/obese categories were relatively more endorsed than
pregnancy involvement or having risky friends. There was large overlap across risk categories.
Sixty percent of youth met at least one of the six risk criteria. If youth met at least one, they were
56% likely to meet another risk criterion or more. Overlap was not uniform across all risk
categories. If youth used any substance in the past month, likelihood for use of one or more other
substances was 55.3% and for overweight/obesity 32.9%, but likelihood of having risky friends
or past pregnancy involvement was each 11.8%. Roughly half of youth endorsed criterion levels
of the positive indices. Youth who met one positive criterion were 46.4% likely to meet the other
one.
As shown in Table 2, friends were highly similar to one another on talk behavior and
health indices. For example, the intra-dyad correlation of r = .70 for positive substance talk
signifies that 70% of the variation in scores on this behavior was explained by membership in a
given dyad alone (Kenny et al., 2006). In terms of risk behavior categories, most dyads consisted
of youth who matched on not meeting the criterion. Among those youth who met the criterion,
about half were in matched dyads, whereas the other approximately half were in mixed dyads.
For example, of those 55 youth with past-month binge drinking, 30 were configured in 15
matched dyads, whereas the other 25 were with a friend who did not binge drink in the past
month. Youth had a 76.1% chance of meeting one or more risk criteria if their friend also did so,
DYADIC HEALTH-RELATED TALK 27
but only a 36.4% chance if the friend did not. As for positive indices, being paired with a friend
who earned good grades or had prosocial friends made youth 80.7% likely to also meet one of
these criteria, but only 57.1% likely if the friend did not.
Gender differences. Male dyads used more risk promoting talk (t = -2.24, p < .05) and
female dyads used more disgust talk (t = 4.49, p < .001). There were no gender differences on
health promoting talk. To assess gender differences on health indices, dyad-level chi-square
statistics were computed based on how dyad members matched on the criterion (both 0, mixed,
or both 1). Females were more likely to have good grades (χ
2
(2) = 7.51, p < .05). No other
significant gender differences in talk behavior or health indices were found.
Race/ethnicity differences. Talk and health variables were compared across the major
ethnic groupings in the sample (Caucasian, Latino, African-American, and other). For
categorical variables (race/ethnicity and health indices), each dyad was coded as 0 (neither youth
endorsed), 1 (one of two endorsed), and 2 (both endorsed) to allow for the calculation of dyad-
level Pearson’s r correlations or chi-square statistics, as appropriate. Risk promoting talk was
unrelated to race/ethnicity. Health promoting talk was more common among Caucasian (r = .29,
p < .01) and other race/ethnicity dyads (r = .20, p < .05), and less common among Latino dyads
(r = -.31, p < .01). Disgust talk was positively associated with African-American dyad
membership (r = .21, p < .05). Caucasian youth were more likely to report having good grades
(χ
2
(4) = 14.66, p < .01) and prosocial friends (χ
2
(4) = 16.12, p < .01), and were less likely to be
overweight or obese (χ
2
(4) = 15.74, p < .01). Latino dyads were more likely to report marijuana
use (χ
2
(4) = 11.31, p < .05). African-American dyads were more likely to report past pregnancy
involvement (χ
2
(4) = 14.74, p < .01) and to be overweight or obese (χ
2
(4) = 11.63, p < .05),
and were less likely to report binge drinking (χ
2
(4) = 12.33, p < .05) and cigarette smoking (χ
2

DYADIC HEALTH-RELATED TALK 28
(4) = 13.71, p < .01). No other significant race/ethnicity associations with health indices were
found, including for the other race/ethnicity category. Associations were also tested separately
for males and females, but as gender differences were not found, overall associations are
presented.
Bivariate associations with age and neighborhood poverty. Older youth used more
risk promoting talk (r = .30, p < .001) and less health promoting talk (r = -.30, p < .001). Age
was also related positively to binge drinking (r = .19, p < .05), cigarette smoking (r = .17, p <
.05), and pregnancy involvement (r = .15, p < .05), and related negatively to having good grades
(r = -.17, p < .05). Neighborhood poverty was inversely associated with health promoting talk (r
= -.20, p < .05) and having good grades (r = -.15, p < .05). Correlations tested separately for
males and females did not show significant differences.
Associations Between Talk Variables and Health Indices
Logistic regression was used to examine associations between the three dyad-level talk
behavior scores and the various individual-level health indices. Talk variables were standardized
prior to analysis. As shown in Table 3, odds ratios were computed to present the added or
reduced likelihood associated with a +1SD change in each talk variable and were adjusted for
gender, age (centered), race/ethnicity, and neighborhood poverty (centered). A talk-by-gender
interaction term was also included to examine differential effects of being male versus being
female. When there was evidence of an interaction effect, simple slopes were calculated for
males and females using the centering technique for logistic regression described by Dawson
(2014). Simple slopes for significant interactions were plotted based on a representative window
of talk scores (0 to 1SD above the mean).

DYADIC HEALTH-RELATED TALK 29
Consistent with risk promotion hypotheses, higher levels of risk promoting talk were
associated with greater likelihood of individuals’ binge drinking (AOR = 3.34, 95% CI = 2.05-
5.46), cigarette smoking (AOR = 3.27, 95% CI = 2.17-4.92), marijuana use (AOR = 4.36, 95%
CI = 2.60-7.31), and having risky friends (AOR = 1.97, 95% CI = 1.05-3.70); a trend-level
positive association was also found for pregnancy (AOR = 1.66, 95% CI = .96-2.86, p = .07).
Risk promoting talk was not related to being overweight or obese, or having good grades. There
was no main effect for having prosocial friends, although a significant gender interaction and
follow-up simple slope test indicated an unexpected positive association specifically for males
(AOR = 1.34, 95% CI = 1.00-1.79; see Figure 2a). No other gender interactions for risk
promoting talk were found.
In line with health promotion hypotheses, health promoting talk was associated with
decreased likelihood of binge-drinking (AOR = .61, 95% CI = .40-.94) and cigarette smoking
(AOR = .53, 95% CI = .33-.86) but not marijuana use; trend-level negative associations with past
pregnancy (AOR = .57, 95% CI = .30-1.05, p = .07) and having risky friends were also found
(AOR = .56, 95% CI = .30-1.04, p = .07). There was no main effect for overweight/obesity,
however, a significant gender interaction term and follow-up simple slope test revealed that for
females, health promoting talk was associated with lower likelihood of being overweight or
obese (AOR = .52, 95% CI = .30-.92; see Figure 2b). Health promoting talk was also associated
with greater likelihood of having good grades (AOR = 1.53, 95% CI = 1.07-2.19) and prosocial
friends (AOR = 1.71, 95% CI = 1.27-2.31). The protective association for prosocial friends was
stronger for females with a significant simple slope (AOR = 2.44, 95% CI = 1.51-3.94; see
Figure 2c). No other gender effects for health promoting talk were found.

DYADIC HEALTH-RELATED TALK 30
Disgust talk was also associated with lower likelihood of past-month substance use
(cigarettes AOR = .42, 95% CI = .23-.78; marijuana AOR = .65, 95% CI = .43-.98; and at a p =
.08 level, binge drinking, AOR = .67, 95% CI = .43-1.05) and of having risky friends (AOR =
.46, 95% CI = .22-.98). The protective effect for risky friends was marginally stronger for males
(p = .06) with a significant simple slope for males (AOR = .21, 95% CI = .05-.91; see Figure 2d).
No significant associations were found for past pregnancy, overweight/obesity, good grades, or
having prosocial friends. No other significant gender interactions for disgust talk were found.
Discussion
The purpose of this study was to investigate adolescents’ health-related talk with a close
friend and its association with current health and functioning. To our knowledge, this is the first
study to use behavioral observation to measure both risk promoting and health promoting talk
among adolescent peers. Past work has emphasized the role of “deviancy training” (a mutually
reinforced use of “rule breaking” talk between friends) in predicting adolescent risk behavior
(Dishion & Tipsord, 2011). In line with this work, the current study found that positive talk
about substances and mutual encouragement of risky ideas was associated largely with
substance-related risk indices, including greater affiliation with peers engaged in substance use
among other problem behaviors. A marginal positive association with past pregnancy
involvement was also found. Contrary to expectation, risk promoting talk was not inversely
related to positive indices; indeed, among males there was a weak, paradoxical association with
increased likelihood of having prosocial friends. Thus school achievement and the prosocial
tendencies of a given youth’s peer group may not be well differentiated by the use of risk
promoting talk as measured in this study.

DYADIC HEALTH-RELATED TALK 31
The most common talk among participants promoted avoidance of or alternatives to risk
behavior. This talk was inversely associated with health risk domains, specifically past-month
binge-drinking and cigarette use, and (only for females) being overweight or obese, and was
positively associated with adaptive functioning, specifically getting good grades and (especially
for females) having prosocial friends. Some of this talk addressed health behavior directly, i.e.,
by expressing or reinforcing risk-averse beliefs. Yet, talk that merely expressed normative
interpersonal values, without any reference to health issues, loaded on the same factor as
explicitly health-oriented talk. Additionally, disgust talk, which did not strongly load on this
same factor, was inversely associated with past-month cigarette and marijuana use, and
(especially for males) having risky friends.
The findings of this study are important because they illustrate the diverse repertoire of
health-related talk that can be observed among youth with their friends. Although seldom
studied, communication among close friends is important because it is a potential mechanism
through which youth come to understand the conventions of their peer group with regards to
health risk. These peer norms are known to be an important predictor of risk behavior (Fishbein,
2008), but their mode of transmission in the peer group is not well understood (Bell & Cox,
2015). Communication may be a missing link. Talk is a mechanism through which friends co-
construct their understanding of peer norms and also develop a consensus about whether or not
to align their beliefs and behavior with such norms.
This dyadic process may help explain why every type of talk in this study was strongly
inter-correlated between friends, including ostensibly idiosyncratic talk such as expressions of
disgust for particular behaviors. At the outset, similarity would be expected given the tendency
of individuals to affiliate with likeminded peers (Brechwald & Prinstein, 2011), yet it is also
DYADIC HEALTH-RELATED TALK 32
likely to arise from the communication process itself. Close friends may start out with similar
views on a health topic and then over the course of communication influence each other in
refining their attitudes and behavioral tendencies. Better understanding of this mutual influence
process, such as through longitudinal data that measure changes in communication over time,
would be useful for intervention design. Just as this dyadic process may reinforce risky norms
among some youth as described in the deviancy training literature, it may also be harnessed to
reinforce more positive alternatives.
Recent approaches to forwarding adolescent health objectives have emphasized the
importance of integrating risk processes with health promoting assets and protective factors in
understanding how to foster youths’ well-being (Kia-Keating et al., 2011). Affiliation with
prosocial peers is seen to protect youth from engagement in multiple forms of risk behavior
(Elkington, Bauermeister, & Zimmerman, 2011; Prinstein et al., 2001). However, little work has
described the nature of health promoting friendships and how protective processes might play
out in ordinary interpersonal interactions. The present study offers some new insights into the
sorts of health-oriented communication employed by youth. The findings here are consistent
with past qualitative studies suggesting that adolescents attempt to intervene with close friends
when they perceive dangers to safety or health and that they do so by using spoken
communication and by invoking personal beliefs (Buckley et al., 2014; Nichter et al., 1997). One
important benefit of the current study was that it allowed for the actual observation of youth as
they engaged in these processes and it demonstrated that these behaviors were found to correlate
not only with avoidance of health risk behavior but also positive engagement with school and
with prosocial peers. It is useful to identify protective factors that are relevant to multiple health
domains, as interventions that integrate various objectives are ultimately more efficient than
DYADIC HEALTH-RELATED TALK 33
addressing each risk category separately (Peters et al., 2009).
It is noteworthy that interpersonal value talk loaded on the same factor as these more
health-oriented types of communication. This finding demonstrates that health promoting talk is
likely to comprise more than just the opposite of talk that is favorable towards risk behavior.
Connectedness with peers has been increasingly viewed as a core competency that protects youth
from engagement in problem behavior (Guerra & Bradshaw, 2008) and youth themselves view
prosocial peers as those who enhance the relational experiences of others (Bergin et al., 2003).
Some work has shown that distal factors that have no obvious association with health, including
“conventional” relationships with peers, are as important in predicting health behavior as
proximal factors such as health-related beliefs (Jessor et al., 1998). Future research should
examine the function of interpersonal value talk vis-à-vis these concepts. It may be that talking
positively about peer relationships is simply a sign of perceived connectedness and its associated
benefits; or, perhaps the talk itself brings some additional benefit by promoting normative
interpersonal values to others.
Youths’ expressions of disgust towards risk behavior were inversely associated with
some risk indices. Disgust talk may represent a personal preference that is not uniform across
risk or health domains, or it may be influenced by social expectancies. An adolescent may be
turned off by smoking but not other risk behavior, or he may outwardly conform to peer group
consensus that tobacco but not beer is disgusting. Thus, expressions of disgust may not
correspond to more intrinsic characteristics of the person, such as risk-taking proclivities or
health-oriented beliefs. Furthermore, disgust may wane as the individual, with experience, is
exposed to more substances. However, from an intervention perspective, disgust talk may serve
as a useful strategy. Youth may be able to distance themselves from risk behavior by citing
DYADIC HEALTH-RELATED TALK 34
personal, idiosyncratic preferences, which may be more socially acceptable than citing personal
health beliefs. Future studies should examine the role of disgust talk, including its stability over
time, its association with health beliefs, and its consistency across specific risk behaviors.
Limitations and Strengths
Implications of the current study are limited by some methodological features. Health
data were collected at the same time as behavioral observations, reducing the potential for causal
inferences. Longitudinal data would clarify antecedents of health-related talk and if talk behavior
predicts later attitudinal and behavioral patterns. This study focused on close dyadic friendships
but was unable to assess the contribution of the wider peer group. Follow-up work would benefit
from taking into account communication from multiple peer relationships, such as through social
network analysis. The use of a laboratory-based discussion task poses limitations to
generalizability. Future studies could expand on this work by observing peer communication in
naturalistic settings or through electronic modalities (e.g., online social networks, text
messaging), given the increasing ubiquity of this type of communication in youths’ lives. In
addition, follow-up studies could attempt to measure risk promoting talk more broadly, beyond
the positive substance talk evaluated here, as such talk might be related to a wider array of health
indices. A final limitation to be considered is that the coding procedure, by having the rater
assess both friends, may have led to biased similarity within dyad. Youth were quite similar on
self-report measures as well, and coding bias was addressed to some degree through separate
viewings per friend and counter-balancing the order of youth across raters, but future
investigators might consider coding dyad members separately.
An important strength of the current study was its use of behavioral observation to
examine multiple forms of health-related talk between close friends, a feature that distinguishes
DYADIC HEALTH-RELATED TALK 35
it from most research examining the peer context of adolescent risk behavior. Although it could
be argued that adolescents friends would not talk openly in a lab-based environment while being
video-recorded, the validity data and the large number of dyads using positive substance talk
suggest that these conversations were authentic. This study also stands out in its assessment of
communication and self-reported risk behavior for both youth and their friends. Many other
studies rely on one participant’s perceptions of the attitudes and behavior of friends, thus
introducing problematic biases (Prinstein & Wang, 2005) as well as obscuring the role of actual
communication versus youths’ perceptions of it. Finally, this study was unique in measuring
various dimensions of talk, including talk that was related positively and negatively to risk
behavior.
Clinical Implications and Summary
Findings from the current study are relevant to the design of programs intended to
improve health behavior among youth. Peer influence processes are often implicated in
adolescent interventions, such as through a curricular focus on peer norms or through the use of
peer educators (Jackson et al., 2012). For example, one empirically supported approach to
substance use prevention is the development through role-play of adolescents’ peer resistance
strategies (Hecht, Graham, & Elek, 2006). In largely school-based programs, youth are taught to
turn down offers of drugs, alcohol, or tobacco using tactics of refusal, delay, and negotiation
(Wolfe, et al., 2012). However, obstacles to using these strategies have been noted, including
youths’ reluctance to make firm refusals within the context of close relationships (Trost, Langan,
& Kellar‐Guenther, 1999). Whereas researchers tend to favor assertive communication
strategies, youth themselves tend to rate these as less effective (Nichols, Birnel, Graber, Brooks-
Gunn, & Botvin, 2010). Furthermore, the ability to generate a greater number of resistance
DYADIC HEALTH-RELATED TALK 36
strategies, whether assertive or not, has been shown to be more effective than a particular
commitment to “just say no” (Wright, Nichols, Graber, Brooks-Gunn, & Botvin, 2004).
The current study, therefore, is relevant in that it examines actual talk behavior occurring
among youth with a range of risk behavior history. The present findings suggest a number of
additional strategies youth may be using to promote healthier behavior not only for themselves
but for their close friends. These findings also suggest a role for communication outside of
situations in which risk behavior is taking place. It is within their conversations with close
friends that youth are formulating health attitudes that can then go on to influence behavior.
Finally, the current study highlights the importance of thinking relationally when considering
adolescent health behavior. Intervention approaches tend to emphasize helping the individual
look after his or her own needs, but as evidenced here, youth appear quite attuned to the health
behavior of their friends. Interventions may benefit from capitalizing on this sense of concern
and responsibility for others.

DYADIC HEALTH-RELATED TALK 37
References
Albert, D., & Steinberg, L. (2011). Peer influences on adolescent risk behavior. In M. T. Bardo,
D. H. Fishbein, & R. Milich (Eds.), Inhibitory control and drug abuse prevention (pp.
211-226). New York, NY: Springer. doi:10.1007/978-1-4419-1268-8
Ali, M. M., & Dwyer, D. S. (2011). Estimating peer effects in sexual behavior among
adolescents. Journal of Adolescence, 34(1), 183-190.
doi:10.1016/j.adolescence.2009.12.008
Arnett, J. J. (2007). The myth of peer influence in adolescent smoking initiation. Health
Education & Behavior, 34(4), 594-607. doi:10.1177/1090198105285330
Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice Hall.
Bell, D. C., & Cox, M. L. (2015). Social norms: Do we love norms too much? Journal of Family
Theory & Review, 7(1), 28-46. doi:10.1111/jftr.12059
Bergin, C., Talley, S., & Hamer, L. (2003). Prosocial behaviours of young adolescents: A focus
group study. Journal of Adolescence, 26(1), 13-32.
Brady, S. S., Morrell, H. E., Song, A. V., & Halpern-Felsher, B. L. (2013). Longitudinal study of
adolescents’ attempts to promote and deter friends’ smoking behavior. Journal of
Adolescent Health, 53(6), 772-777. doi:10.1016/j.jadohealth.2013.06.022
Brechwald, W. A., & Prinstein, M. J. (2011). Beyond homophily: A decade of advances in
understanding peer influence processes. Journal of Research on Adolescence, 21(1), 166-
179. doi:10.1111/j.1532-7795.2010.00721.x
Buckley, L., Chapman, R. L., Sheehan, M. C., & Reveruzzi, B. N. (2014). In their own words:
Adolescents’ strategies to prevent friends’ risk taking. The Journal of Early Adolescence,
34(4), 539-561. doi:10.1177/0272431613496637
DYADIC HEALTH-RELATED TALK 38
Capaldi, D. M., & Clark, S. (1998). Prospective family predictors of aggression toward female
partners for at-risk young men. Developmental Psychology, 34(6), 1175-1188.
doi:10.1037/0012-1649.34.6.1175
Catalano, R. F., Hawkins, J. D., Berglund, M. L., Pollard, J. A., & Arthur, M. W. (2002).
Prevention science and positive youth development: Competitive or cooperative
frameworks? Journal of Adolescent Health, 31(6), 230-239. doi:10.1016/S1054-
139X(02)00496-2
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation
analysis for the behavioral sciences (3rd ed.). Mahwah, NJ: Erlbaum.
Costa, F. M., Jessor, R., Fortenberry, J. D., & Donovan, J. E. (1996). Psychosocial
conventionality, health orientation, and contraceptive use in adolescence. Journal of
Adolescent Health, 18(6), 404-416. doi:10.1016/1054-139X(95)00192-U
Dawson, J. F. (2014). Moderation in management research: What, why, when, and how. Journal
of Business and Psychology, 29(1), 1-19. doi:10.1007/s10869-013-9308-7
Dishion, T. J., Spracklen, K. M., Andrews, D. W., & Patterson, G. R. (1996). Deviancy training
in male adolescent friendships. Behavior Therapy, 27(3), 373-390. doi:10.1016/S0005-
7894(96)80023-2
Dishion, T. J., Andrews, D. W., & Crosby, L. (1995). Antisocial boys and their friends in early
adolescence: Relationship characteristics, quality, and interactional process. Child
Development, 66(1), 139-151. doi:10.1111/j.1467-8624.1995.tb00861.x
Dishion, T. J., & Tipsord, J. M. (2011). Peer contagion in child and adolescent social and
emotional development. Annual Review of Psychology, 62, 189-214.
doi:10.1146/annurev.psych.093008.100412
DYADIC HEALTH-RELATED TALK 39
Eaton, D. K., Kann, L., Kinchen, S., Shanklin, S., Ross, J., Hawkins, J., … Chyen, D. (2008).
Youth risk behavior surveillance—United States, 2007. Morbidity and Mortality Weekly
Report (MMWR), 57(4), 1-131.
Elkington, K. S., Bauermeister, J. A., & Zimmerman, M. A. (2011). Do parents and peers
matter? A prospective socio-ecological examination of substance use and sexual risk
among African American youth. Journal of Adolescence, 34(5), 1035-1047.
doi:10.1016/j.adolescence.2010.11.004
Fishbein, M. (2008). A reasoned action approach to health promotion. Medical Decision Making,
28(6), 834-844. doi:10.1177/0272989X08326092
Gardner, M., & Steinberg, L. (2005). Peer influence on risk taking, risk preference, and risky
decision making in adolescence and adulthood: An experimental study. Developmental
Psychology, 41(4), 625-635. doi:10.1037/0012-1649.41.4.625
Granovetter, M. (1983). The strength of weak ties: A network theory revisited. Sociological
Theory, 1(1), 201-233.
Griffin, D., & Gonzalez, R. (1995). Correlational analysis of dyad-level data in the exchangeable
case. Psychological Bulletin, 118(3), 430-439. doi:10.1037/0033-2909.118.3.430
Guerra, N. G., & Bradshaw, C. P. (2008). Linking the prevention of problem behaviors and
positive youth development: Core competencies for positive youth development and risk
prevention. New Directions for Child and Adolescent Development, 122, 1-17.
doi:10.1002/cd.225
Hecht, M. L., Graham, J. W., & Elek, E. (2006). The drug resistance strategies intervention:
Program effects on substance use. Health Communication, 20(3), 267-276.
doi:10.1207/s15327027hc2003_6
DYADIC HEALTH-RELATED TALK 40
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling: A
Multidisciplinary Journal, 6(1), 1-55. doi:10.1080/10705519909540118
IBM Corp. (2013). IBM SPSS Statistics for Windows, Version 22. Armonk, NY: IBM Corp.
Jackson, C., Geddes, R., Haw, S., & Frank, J. (2012). Interventions to prevent substance use and
risky sexual behaviour in young people: A systematic review. Addiction, 107(4), 733-
747. doi:10.1111/j.1360-0443.2011.03751.x
Jessor, R., & Jessor, S. L. (1977). Problem behavior and psychosocial development: A
longitudinal study of youth. New York, NY: Academic Press.
Jessor, R., Turbin, M. S., & Costa, F. M. (1998). Protective factors in adolescent health behavior.
Journal of Personality and Social Psychology, 75(3), 788-800. doi:10.1037/0022-
3514.75.3.788
Johnston, L. D., O’Malley, P. M., Miech, R. A., Bachman, J. G., & Schulenberg, J. E. (2014).
Monitoring the Future national results on drug use: 1975-2013: Overview, key findings
on adolescent drug use. Ann Arbor, MI: Institute for Social Research, The University of
Michigan.
Kenny, D., Kashy, D., & Cook, W. (2006). Dyadic data analysis. New York, NY: The Guilford
Press.
Kia-Keating, M., Dowdy, E., Morgan, M. L., & Noam, G. G. (2011). Protecting and promoting:
An integrative conceptual model for healthy development of adolescents. Journal of
Adolescent Health, 48(3), 220-228. doi:10.1016/j.jadohealth.2010.08.006
Kuczmarski, R. J., Ogden, C. L., Guo, S. S., Grummer-Strawn, L. M., Flegal, K. M., Mei, Z., …
Johnson, C. L. (2002). 2000 CDC growth charts for the United States: Methods and
DYADIC HEALTH-RELATED TALK 41
development. National Center for Health Statistics. Vital and Health Statistics, 11(246),
1-190.
Lapinski, M. K., & Rimal, R. N. (2005). An explication of social norms. Communication Theory,
15(2), 127-147. doi:10.1111/j.1468-2885.2005.tb00329.x
León, J., Carmona, J., & García, P. (2010). Health-risk behaviors in adolescents as indicators of
unconventional lifestyles. Journal of Adolescence, 33(5), 663-671.
doi:10.1016/j.adolescence.20
MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power analysis and
determination of sample size for covariance structure modeling. Psychological Methods,
1(2), 130-149. doi:10.1037/1082-989X.1.2.130
McCambridge, J., McAlaney, J., & Rowe, R. (2011). Adult consequences of late adolescent
alcohol consumption: A systematic review of cohort studies. PLoS Medicine, 8(2),
e1000413. doi:10.1371/journal.pmed.1000413.
McLeod, K., White, V., Mullins, R., Davey, C., Wakefield, M., & Hill, D. (2008). How do
friends influence smoking uptake? Findings from qualitative interviews with identical
twins. Journal of Genetic Psychology, 169(2), 117-131. doi:10.3200/GNTP.169.2.117-
132
Meier, M. H., Caspi, A., Ambler, A., Harrington, H., Houts, R., Keefe, R. S., … Moffitt, T. E.
(2012). Persistent cannabis users show neuropsychological decline from childhood to
midlife. Proceedings of the National Academy of Sciences, 109(40), E2657-E2664.
doi:10.1073/pnas.1206820109
Metrik, J., Frissell, K. C., McCarthy, D. M., D’Amico, E. J., & Brown, S. A. (2003). Strategies
for reduction and cessation of alcohol use: Adolescent preferences. Alcoholism: Clinical
DYADIC HEALTH-RELATED TALK 42
and Experimental Research, 27(1), 74-80. doi:10.1097/01.ALC.0000046596.09529.03
Muthén, L. K., & Muthén, B. O. (2012). Mplus User’s Guide. Seventh Edition. Los Angeles, CA:
Muthén & Muthén.
Healthy People 2020: Priority objectives and core indicators for adolescent health. Retrieved
January 27, 2015, from http://nahic.ucsf.edu/resources/hp2020/ 2014
Newcomb, A. F., & Bagwell, C. L. (1995). Children’s friendship relations: A meta-analytic
review. Psychological Bulletin, 117(2), 306-347. doi:10.1037//0033-2909.117.2.306
Nichols, T. R., Birnel, S., Graber, J. A., Brooks-Gunn, J., & Botvin, G. J. (2010). Refusal skill
ability: An examination of adolescent perceptions of effectiveness. Journal of Primary
Prevention, 31(3), 127-137. doi:10.1007/s10935-010-0213-9
Nichter, M., Vuckovic, N., Quintero, G., & Ritenbaugh, C. (1997). Smoking experimentation
and initiation among adolescent girls: Qualitative and quantitative findings. Tobacco
Control, 6(4), 285-295. doi:10.1136/tc.6.4.285
Olsen, J. A., & Kenny, D. A. (2006). Structural equation modeling with interchangeable dyads.
Psychological Methods, 11(2), 127-141. doi:10.1037/1082-989X.11.2.127
Patterson, G. R., Dishion, T. J., & Yoerger, K. (2000). Adolescent growth in new forms of
problem behavior: Macro-and micro-peer dynamics. Prevention Science, 1(1), 3-13.
doi:10.1023/A:1010019915400
Peters, L. W., Wiefferink, C. H., Hoekstra, F., Buijs, G. J., Ten Dam, G. T., & Paulussen, T. G.
(2009). A review of similarities between domain-specific determinants of four health
behaviors among adolescents. Health Education Research, 24(2), 198-223.
doi:10.1093/her/cyn013
Peterson, J., Piehler, T., & Dishion, T. J. (2006). Topic Code v. 3.0 manual. Unpublished coding
DYADIC HEALTH-RELATED TALK 43
manual. Oregon Social Learning Center.
Peugh, J. L., DiLillo, D., & Panuzio, J. (2013). Analyzing mixed-dyadic data using structural
equation models. Structural Equation Modeling: A Multidisciplinary Journal, 20(2), 314-
337. doi:10.1080/10705511.2013.769395
Piehler, T. F., & Dishion, T. J. (2007). Interpersonal dynamics within adolescent friendships:
Dyadic mutuality, deviant talk, and patterns of antisocial behavior. Child Development,
78(5), 1611-1624. doi:10.1111/j.1467-8624.2007.01086.x
Prinstein, M. J., Boergers, J., & Spirito, A. (2001). Adolescents’ and their friends’ health-risk
behavior: Factors that alter or add to peer influence. Journal of Pediatric Psychology,
26(5), 287-298. doi:10.1093/jpepsy/26.5.287
Prinstein, M. J., & Wang, S. S. (2005). False consensus and adolescent peer contagion:
Examining discrepancies between perceptions and actual reported levels of friends’
deviant and health risk behaviors. Journal of Abnormal Child Psychology, 33(3), 293-
306. doi:10.1007/s10802-005-3566-4
Real, K., & Rimal, R. N. (2007). Friends talk to friends about drinking: Exploring the role of
peer communication in the theory of normative social behavior. Health Communication,
22(2), 169-180.
Rosengard, C., Adler, N. E., Gurvey, J. E., Dunlop, M. B. V., Tschann, J. M., Millstein, S. G., &
Ellen, J. M. (2001). Protective role of health values in adolescents’ future intentions to
use condoms. Journal of Adolescent Health, 29(3), 200-207. doi:10.1016/S1054-
139X(01)00216-6
R⊘ysamb, E., Rse, J., & Kraft, P. (1997). On the structure and dimensionality of health-related
behaviour in adolescents. Psychology & Health, 12(4), 437-452.
DYADIC HEALTH-RELATED TALK 44
doi:10.1080/08870449708406721
Simons-Morton, B. G., & Farhat, T. (2010). Recent findings on peer group influences on
adolescent smoking. Journal of Primary Prevention, 31(4), 191-208.
doi:10.1007/s10935-010-0220-x
Trost, M. R., Langan, E. J., & Kellar‐Guenther, Y. (1999). Not everyone listens when you “just
say no”: Drug resistance in relational context. Journal of Applied Communication
Research, 27(2), 120-138. doi:10.1080/00909889909365530
Turbin, M. S., Jessor, R., & Costa, F. M. (2000). Adolescent cigarette smoking: Health-related
behavior or normative transgression? Prevention Science, 1(3), 115-124.
doi:10.1023/A:1010094221568
United States Census Bureau. Selected economic characteristics. 2007-2011 American
Community Survey 5-Year Estimates. Retrieved 7/28/2013, from
www.factfinder.census.gov
Urberg, K. A., Luo, Q., Pilgrim, C., & Degirmencioglu, S. M. (2003). A two-stage model of peer
influence in adolescent substance use: Individual and relationship-specific differences in
susceptibility to influence. Addictive Behaviors, 28(7), 1243-1256. doi:10.1016/S0306-
4603(02)00256-3
Wiefferink, C. H., Peters, L., Hoekstra, F., Dam, G. T., Buijs, G. J., & Paulussen, T. G. (2006).
Clustering of health-related behaviors and their determinants: Possible consequences for
school health interventions. Prevention Science, 7(2), 127-149. doi:10.1007/s11121-005-
0021-2
Williams, P. G., Holmbeck, G. N., & Greenley, R. N. (2002). Adolescent health psychology.
Journal of Consulting and Clinical Psychology, 70(3), 828-842. doi:10.1037//0022-
DYADIC HEALTH-RELATED TALK 45
006X.70.3.828
Wolfe, D. A., Crooks, C. V., Chiodo, D., Hughes, R., & Ellis, W. (2012). Observations of
adolescent peer resistance skills following a classroom-based healthy relationship
program: A post-intervention comparison. Prevention Science, 13(2), 196-205.
doi:10.1007/s11121-011-0256-z
Wright, A. J., Nichols, T. R., Graber, J. A., Brooks-Gunn, J., & Botvin, G. J. (2004). It’s not
what you say, it’s how many different ways you can say it: Links between divergent peer
resistance skills and delinquency a year later. Journal of Adolescent Health, 35(5), 380-
391. doi:10.1016/j.jadohealth.2003.12.008
Yanovitzky, I., Stewart, L. P., & Lederman, L. C. (2006). Social distance, perceived drinking by
peers, and alcohol use by college students. Health Communication, 19(1), 1-10.
doi:10.1207/s15327027hc1901_1

DYADIC HEALTH-RELATED TALK 46
Table 1
Coded Talk Behaviors
Code Examples Inter-rater
reliability
Talk
behaviors

Positive
substance
talk
[Laughing approvingly] You got so drunk Friday night.
Marijuana helps me relax.
.91
Risk-positive
response
Yeah of course, definitely [let’s get drunk].
Just do it anyway. What is she gonna do [about your risk
behavior]?
.67
Cautious
health talk
I don’t drink alcohol. My family has a history of
alcoholism, a lot of issues on both sides. My parents are
not alcoholics.
When I talk to [other friend’s name], she’s like why aren’t
you hooking up with people? It’s not in my personality.
It’s not my objective, to go hook up with someone and get
a boyfriend.
.68
Health
promoting
response
Yeah drinking is bad, I don’t approve.
You have to set an example [for your brother who is
breaking the law]. That’s what an older brother does.
.64
Interpersonal
value talk
I’d like others to be intelligent so I can have an intelligent
conversation with them. That’s something I care about.
And also I’d like them to have similar interests with me.
I think personality definitely trumps hotness any day.
.78
Disgust talk The idea of hooking up with somebody I don’t really know
is gross.
Cigarettes are disgusting!
.84
Validity
behaviors

Engagement Youth is verbally and non-verbally responsive to friend .73
Faking Frequent sarcastic speech indicating inauthentic
statements
.92
Note. Inter-rater reliability was assessed using the intraclass correlation.

DYADIC HEALTH-RELATED TALK 47
Table 2
Descriptive Statistics of Talk Behavior and Health Indices
Continuous measures factor loading mean (SD) % any min–max intra-dyad correlation
Talk behavior
Dyadic risk promoting talk — .07 (.06)
a
88.3 0–.24 —
Positive substance talk .85
***
.29 (.27) 80.2 0–1.25 .70
***

Risk-positive responses .58
***
.10 (.14) 53.6 0–.67 .65
***

Dyadic health promoting talk — .34 (.16)
a
100.0 .01–.74 —
Cautious health talk .63
***
.40 (.21) 95.9 0–1.33 .55
***

Health promoting responses .80
***
.14 (.10) 88.7 0–.46 .57
***

Interpersonal value talk .58
***
.48 (.29) 93.2 0–1.25 .79
***

Disgust talk — .05 (.05)
a
73.9 0–.18 .53
***

Engagement — 2.39 (.35) 100.0 1.33–3.00
a
.70
***

Faking — .05 (.37) 4.5 0–5 .25
***

Individuals Dyads
Met criterion Both met criterion Neither met criterion Mixed
Categorical measures n (%) n (%) n (%) n (%)
Health & functioning
Binge drinking 55 (24.8) 15 (13.5) 71 (64.0) 25 (22.5)
Cigarettes 38 (17.1) 9 (8.1) 82 (73.9) 20 (18.0)
Marijuana 55 (24.8) 14 (12.6) 70 (63.1) 27 (24.3)
Pregnancy 20 (9.0) 4 (3.6) 95 (85.6) 12 (10.8)
Overweight or obese 72 (32.4) 16 (14.4) 55 (49.5) 40 (36.0)
Risky friends 16 (7.2) 2 (1.8) 97 (87.4) 12 (10.8)
Good grades 134 (60.4) 50 (45.0) 27 (24.3) 34 (30.6)
Prosocial friends 105 (47.3) 28 (25.2) 34 (30.6) 49 (44.1)
Note.
a
Dyadic mean score is presented. For correlations,
*
p < .05,
**
p < .01,
***
p < .001.
DYADIC HEALTH-RELATED TALK 48
Table 3
Associations Between Dyadic Talk Behavior and Current Health Domains
Binge drinking Cigarettes Marijuana Past pregnancy Overweight or Obese Risky friends Good grades Prosocial friends
Dyadic talk
Adj.
OR
(95%
CI)
Adj.
OR
(95%
CI)
Adj.
OR
(95%
CI)
Adj.
OR
(95%
CI)
Adj.
OR
(95%
CI)
Adj.
OR (95% CI)
Adj.
OR
(95%
CI)
Adj.
OR
(95%
CI)
Risk promoting
talk
3.34
***
(2.05-
5.46)
3.27
***
(2.17-
4.92)
4.36
***
(2.60-
7.31)
1.66
+
(.96-
2.86)
.94 (.64-
1.39)
1.97
*
(1.05-
3.70)
.91 (.65-
1.28)
1.01 (.75-
1.34)
× gender
.90 (.36-
2.27)
.73 (.35-
1.53)
1.02 (.37-
2.81)
.75 (0.26-
2.11)
.89 (.44-
1.83)
.52 (.16-
1.71)
1.02 (.55-
1.91)
1.78
*
(1.04-
3.06)
Gender-
specific slopes

Males — — — — — — — — — — — — — — 1.34
*
(1.00-
1.79)
Females — — — — — — — — — — — — — — .75

(.47-
1.21)
Health
promoting talk
.61
*
(.40-
.94)
.53
*
(.33-
.86)
.67 (.41-
1.08)
.57
+
(.30-
1.05)
.75 (.51-
1.10)
.56
+
(.30-
1.04)
1.53
*
(1.07-
2.19)
1.71
***
(1.27-
2.31)
× gender
.50 (.22-
1.14)
1.48 (.59-
3.70)
.78 (.32-
1.91)
.86 (.27-
2.72)
2.05
*
(1.04-
4.04)
.46 (.18-
1.16)
.85 (.43-
1.68)
.49
*
(.28-
.85)
Gender-
specific slopes

Males — — — — — — — — .77 (.68-
1.68)
— — — — .26
+
(.06-
1.04)
Females — — — — — — — — .52
*
(.30-
.92)
— — — — 2.44
***
(1.51-
3.94)
Disgust
talk
.67
+
(.43-
1.05)
.42
**
(.23-
.78)
.65
*
(.43-
.98)
.59 (.26-
1.36)
.74 (.50-
1.10)
.46
*
(.22-
.98)
1.08 (.71-
1.63)
1.22 (.85-
1.74)
× gender
1.30 (.52-
3.20)
1.27 (.36-
4.53)
1.15 (.50-
2.65)
.82 (.17-
3.91)
.63 (.30-
1.34)
.21
+
(.04-
1.07)
.71 (.32-
1.58)
.93 (.45-
1.93)
Gender-
specific slopes

Males — — — — — — — — — — .21
*
(.05-
.91)
— — — —
Females — — — — — — — — — — 1.00
+
(.56-
1.80)
— — — —
Note. Adjusted odds ratio estimates account for gender, age, race/ethnicity, and neighborhood poverty. Talk predictors were
standardized. For gender interactions, female is reference.
+
p < .09,
*
p < .05,
**
p < .01,
***
p < .001.
DYADIC HEALTH-RELATED TALK 49

Figure 1. Dyadic confirmatory factor analysis of talk behaviors. Standardized coefficients are
shown. PS = positive substance talk; RP = risk-positive responses; CH = cautious health talk; HP
= health promoting responses; IV = interpersonal value talk; DT = disgust talk. Not shown are
significant coding system-related interpersonal covariances between PS and RP (β = .18; p <
.01), and PS and HP (β = .31; p < .01), and intrapersonal covariance for PS and CH (β = -.21; p <
.01).
*
p < .05,
**
p < .01,
***
p < .001

DYADIC HEALTH-RELATED TALK 50
a. Most friends are prosocial

Risk promoting talk

b. Overweight or obese

Health promoting talk

c. Most friends are prosocial

Health promoting talk

d. Most friends are risky

Disgust talk

Figure 2. Significant gender interactions. Slope is significant at
+
p < .10,
*
p < .05,
**
p < .01,
***
p < .001.
*
*
***
+
*
+
Running Head: INTIMACY PROMOTING TALK 51

Emotional Disclosure and Support Between Close Adolescent Friends:
Does Intimacy Promoting Talk Relate to Sexual Risk Behavior?

Esti Iturralde
University of Southern California

INTIMACY PROMOTING TALK 52
Abstract
Some research has found that intimate friendship processes, such as friends’ support and
closeness, are related to lower sexual risk behavior among adolescents. These findings, however,
have been based on self-report questionnaires rather than observed behavior. The current study
used coded observations of close friends to examine talk behaviors that foster interpersonal
intimacy and to assess how these relate to youths’ sexual risk history. A sample of 222 ethnically
and socioeconomically diverse male and female adolescents (aged 15-20), consisting of 111
pairs of close, gender-matched friends, participated in a 30-minute structured conversation task.
Youths’ behavior was rated by observers for intimacy promoting talk (emotional disclosure,
perspective taking, and emotional support). Associations were tested between this dyadic talk
behavior and individuals’ reported past pregnancy involvement and a composite STI risk
variable (including history of STI, past-year unprotected sex, and number of sexual partners). In
multilevel analyses, dyadic intimacy promoting talk was associated with reduced likelihood of
pregnancy involvement and STI risk. Results persisted after adjusting for gender, age,
neighborhood poverty, race/ethnicity, and three other peer context variables: peer norms for
unprotected sex, the dyad partner’s risk level, and the dyad’s use of health promoting talk. The
effect of intimacy promoting talk was similar in absolute magnitude to the effect of peer norms, a
robust predictor in past studies of adolescent sexual risk behavior. This is the first study to use
behaviorally coded interactions to demonstrate an inverse association between adolescent sexual
risk and close friends’ talk that promotes intimacy. Findings add to growing evidence that
enhancing adolescents’ interpersonal skills may help reduce sexual risk behavior.
Keywords: adolescent sexual health, peer relationships, intimacy, social support, health
communication, behavioral observation, dyadic analysis
INTIMACY PROMOTING TALK 53
Emotional Disclosure and Support Between Close Adolescent Friends:
Does Intimacy Promoting Talk Relate to Sexual Risk Behavior?
Relative to other age groups, adolescents are particularly vulnerable to sexual health risks
such as sexually transmitted infections (STIs) and unintended pregnancy (Finer, 2010; Forhan et
al., 2009), health outcomes that present significant burdens both to individuals and society
(Casares, Lahiff, Eskenazi, & Halpern-Felsher, 2010; Trussell et al., 2013; Weinstock, Berman,
& Cates Jr, 2004). Therefore, considerable attention has been directed towards reducing
adolescents’ exposure to these risks, with the role of peers as a major area of focus for prevention
efforts (Simoni, Nelson, Franks, Yard, & Lehavot, 2011). Research has found that youth are
more likely to use condoms when they think their friends are also doing so (Henry, Deptula, &
Schoeny, 2012; Romer et al., 1994). Understandably, sexual health interventions have attempted
to harness peer influence processes. In addition to utilizing peers as health educators, prevention
programs have also targeted youths’ social-emotional competence with peers, such as by training
participants in healthy communications skills (Sieving et al., 2011; Tuttle, Bidwell-Cerone,
Campbell-Heider, Richeson, & Collins, 2000). Based on a positive youth development
framework (Gavin, Catalano, David-Ferdon, Gloppen, & Markham, 2010), this approach seeks
to capitalize on youth assets to offset negative social processes such as risky peer norms
(Kotchick, Shaffer, Miller, & Forehand, 2001). Despite these theoretical underpinnings, there has
been little focused investigation of how positive aspects of peer relationship functioning are
themselves associated with sexual health.
A small number of studies have suggested that the absence of intimate friendship puts
youth at risk for negative sexual health outcomes. Studies have found that pregnant teenagers are
less socially competent (Passino et al., 1993), have fewer friends (Alvarez, Burrows, Zvaighat, &
INTIMACY PROMOTING TALK 54
Muzzo, 1987), and experience less love (Adolph, Ramos, Linton, & Grimes, 1995) and
emotional closeness with friends (Pereira, Canavarro, Cardoso, & Mendonça, 2004) compared to
non-pregnant peers. In addition, adolescents with supportive and stable friendships have been
found to engage in less sexual risk behavior (Miller, Notaro, & Zimmerman, 2002; Ramiro,
Teva, Bermúdez, & Buela-Casal, 2013) and more consistent condom use over time (Elkington,
Bauermeister, & Zimmerman, 2011). Supportive friendships have also been found to amplify
resiliency factors (e.g., parent connectedness; Henrich, Brookmeyer, Shrier, & Shahar, 2006) and
buffer risk factors (e.g., uncontrollable stressful life events; Brady, Dolcini, Harper, & Pollack,
2009) in the prediction of sexual health outcomes. These positive effects may be due to the traits
of youths’ friends or the particular dynamics of their friendships, however, youth may also
benefit from having individual skills that promote intimacy with friends. Both intrapersonal
skills (e.g., emotional expression) and interpersonal skills (e.g., empathy) are associated with
reduced sexual health risk (Lando-King et al., in press). On the other hand, contrary findings also
exist, with peer support shown as predictive of increased sexual risk behavior (Benda & Corwyn,
1999; Salazar et al., 2007).
Measures used to assess intimate friendship processes in these studies are not consistent
and have relied on youths’ own report of these relationship qualities. Therefore, it is not known if
health effects are being driven by actual friendship characteristics or merely by youths’
perceptions of how their friendships function. Another problem is that youths’ perceptions of, for
example, how much they feel supported by friends, does not capture the specific behaviors
actually used by friends, thus limiting the utility for intervention design. There has not been
adequate focus in the literature on positive peer communication processes as they pertain to
sexual health behavior (Iturralde & Margolin, in preparation; Widman, Choukas-Bradley, Helms,
INTIMACY PROMOTING TALK 55
Golin, & Prinstein, 2014).
It is also not clear from the research how intimate friendship processes relate to other
peer context variables. Models that take into consideration both adaptive and risky dimensions of
peer influence offer a more comprehensive understanding of adolescent health phenomena (Kia-
Keating, Dowdy, Morgan, & Noam, 2011). A potential source of influence are descriptive peer
group norms, i.e., the perceived level of sexual risk behavior among peers. Descriptive norms are
a robust predictor of youths’ own risk level (Buhi & Goodson, 2007). Some have found evidence
of more influence from close friends’ norms in particular, compared to those of the wider peer
group (Ali & Dwyer, 2011). Also important may be communication processes that promote
alternatives to risk behavior. Some preliminary research has found that youth use a range of talk
behavior to promote healthy concepts with their close friends and that this is associated favorably
with a variety of health indices (Iturralde, 2015). The study described here examined not only
observed intimate communication, but also the contributions of peer group norms, risk behavior
of a close friend, and dyadic health promoting talk.
Current Study
The present study used a novel approach to investigate associations between intimate
friendship processes and youths’ sexual risk behavior. Youth of both genders were observed in
conversation with a close friend of the same gender, and their behavior was assessed using a
coding system to measure dimensions of intimacy promoting talk. Several talk behaviors were
selected on an a priori basis in view of past theoretical and empirical work, which has described
the importance of self-disclosure, perspective-taking, and empathic support in fostering an
intimate connection between friends in adolescence (Berndt, 1982; Buhrmester, 1990; Gottman
& Mettetal, 1986; Laurenceau, Barrett, & Pietromonaco, 1998; Paul & White, 1990). The current
INTIMACY PROMOTING TALK 56
study had three objectives: (1) to offer a proof of concept that a range of intimacy promoting talk
may be reliably measured among adolescent friends by using behavioral observation; (2) to
assess associations between intimacy promoting talk and likelihood of past pregnancy and STI
risk; and, (3) to compare the contribution of intimacy promoting talk with that of other peer
context variables. The three peer context variables were: peer group norms regarding unprotected
sex; past pregnancy involvement or STI risk of the dyad partner; and, health promoting talk
observed during the discussion task. Consistent with the existing literature, it was hypothesized
that intimacy promoting talk would be associated with reduced risk of past pregnancy and STI
risk.
Method
Participants
Adolescents in the current study were enrolled in a larger multi-wave longitudinal project
involving families recruited from a large urban community in the western United States. Each
study youth was instructed to recruit a close friend of the same gender and similar age. A total of
111 pairs of close adolescent friends (50 female and 61 male dyads; mean age = 17.63; SD =
1.32) participated in the present study. About 88% of participants rated the friend as one of my 5
closest or closer; 80.2% had known the friend for 3 or more years; 51.4% of dyads reported
being the same age whereas another 45.0% indicated 1 to 2 years of age difference. Youth were
diverse in race/ethnicity and SES. Of the total 222 participating youth, 31.5% were non-
Hispanic/Latino Caucasian, 31.5% indicated primarily Hispanic/Latino descent, 25.2% identified
with African-American or mixed African-American race/ethnicity, 5.0% identified as Asian or
Pacific Islander, and 6.8% were of another race/ethnicity. Youth varied considerably on a zip
code-based measure of neighborhood poverty; they ranged from having 0 to 46.1% of families in
INTIMACY PROMOTING TALK 57
their home neighborhood living under the federal poverty level (median = 8.5%). More details
about this sample can be found in Iturralde (2015).
Procedures
Data were collected during a lab visit attended by both members of the dyad. For those
youth under age 18, consent for participating in the study was obtained from parents before the
visit. Participants who were old enough to self-consent did so in the lab. Participants were
reassured throughout their visit that their responses to sexual behavior measures would be
aggregated with other data and would not be shared with family members or their accompanying
friend. To maximize privacy, youth were seated in separate rooms to complete computer-
administered questionnaires. Procedures were approved by the study university’s IRB.
Discussion task and behavioral coding system. Dyads participated in a discussion task
modeled on past research (Capaldi & Clark, 1998; Dishion, Andrews, & Crosby, 1995). Prior to
the discussion task, to help prime youth to disclose information of an emotional quality,
participants were separately presented with a list of close others (sibling, mom, dad, significant
other, etc.) and asked to indicate those with whom they have an unresolved issue or problem.
Shortly thereafter the friends were brought together to begin the interactive task, which was
structured around 5-minute topic discussions. The first six topics presented to youth were
analyzed in the current study; there were 3 subsequent topics not included here. Youth were left
alone to discuss each topic while being video-recorded. Topics were selected to be relevant to the
social lives of adolescents and to elicit a range of talk. In the first 2 topics, youth took turns
describing an unresolved problem with individuals they selected from their lists. The third topic
concerned attitudes about substance use. Topic 4 focused on short-term goals, topic 5 on
romantic partners, and topic 6 on activities of the youths’ peer group. In post-discussion
INTIMACY PROMOTING TALK 58
assessments, 68.9% of youth rated the discussion as moderately to very similar to past
discussions with the friend; 90.6% were able to express their point of view moderately to very
much; 95.5% said they were moderately to very honest during the discussion. (See also Iturralde,
2015.)
Behavior was assessed using the Friend Talk coding system (Iturralde, 2015). Using a
macro-coding approach, each coded behavior was assessed for overall frequency and intensity on
a 0 to 3 scale (none, some, a moderate amount, a lot). Scores were assigned once per 5-minute
topic interval. Each of the discussions was scored by 2 raters working independently. Scores
were averaged across the 6 topics to create a single discussion-level mean score per individual.
Inter-rater reliability was assessed by computing an intraclass correlation (ICC) across raters.
The 2 raters’ scores were then averaged together for use in analyses.
Measures
Intimacy promoting talk. Three codes were used. Given the strong inter-correlation of
talk behavior across partners (intra-dyad rs = .49–.62, ps < .001), scores for both friends were
averaged together to form a single dyadic intimacy promoting talk score. The first code,
emotional disclosure, assessed openness about experiences, thoughts, and feelings that could be
difficult or embarrassing. Emotional disclosure could include statements about vulnerable
feelings and related judgments, expressions of caring to the friend, and non-verbal cues such as
crying or frowning. The second code, taking peer’s perspective, was applied when the youth
made inferences about the friend’s motivations, feelings, beliefs, or experiences, for example by
voicing what the youth might be thinking. The final code, emotional support, captured
expressions of caring and empathy, such as making comforting sounds in response to
disclosures, non-verbal mirroring of vulnerable emotions, statements that reflect the emotion
INTIMACY PROMOTING TALK 59
underlying the friend’s disclosure, and humor that is supportive and friendly. Intimacy codes did
not apply to talk being used to promote or discourage risky behavior. The overall dyadic score
was internally consistent (Cronbach’s α = .74) and raters’ scores were reliably consistent with
one another (ICC = .78). Intimacy promoting talk was normally distributed. Sample quotations
illustrating the codes used to construct this talk measure are shown in Table 1.
Pregnancy involvement. A single yes or no item was used. Depending on gender,
participants were asked if they had ever been pregnant or gotten someone else pregnant.
Sexually transmitted infection (STI) risk. A composite score ranging from 0 to 3 was
computed based on how many of 3 items were endorsed by youth at the following criterion
levels: (1) any sexually transmitted infection in the past (yes or no); (2) 6 or more past-year
sexual partners; and, (3) more than 10 past-year instances of sex without using a condom. The
two latter items were adapted from the CDC Youth Risk Behavior Surveillance System (YRBS;
Eaton et al., 2012) and used 6-point scales, with the criterion level as the highest response. The
resulting distribution was not over-dispersed (mean = .30, SD
2
= .35); therefore, this variable
was appropriate for use in Poisson regression (Atkins, Baldwin, Zheng, Gallop, & Neighbors,
2013).
Sexual activity status. This dichotomous variable was scored as a 1 if participants
reported greater than zero lifetime sexual partners (intercourse, oral sex, or anal sex) on 1 item.
Demographic variables. Gender was effect coded (-.5 male, .5 female). Whole number
age was used as a covariate. Neighborhood poverty was the percentage of families in the youth’s
home neighborhood living under the federal poverty level, based on the zip code-indexed dataset
of the American Community Survey (U.S. Census Bureau, 2011); percentages were log-
transformed for correlation analyses to mitigate right skew. In regression analyses, race/ethnicity
INTIMACY PROMOTING TALK 60
was represented through weighted coding of three variables: Caucasian, Latino, and African-
American, with all other ethnicities treated as the reference group. Weighted effects coding
accounts for unequal group sizes, centers each effect variable around zero, and provides a
comparison between the given group and the mean across groups (Cohen, Cohen, West, &
Aiken, 2003).
Peers’ unprotected sex norms. One item was adapted from the Peer Behavior Inventory
(Prinstein, Boergers, & Spirito, 2001). Youth were asked how many of their friends engage in
unprotected sex, with responses made on a 0 to 4 scale (none, one, a few, more than half, almost
all). Scores were log-transformed for correlation analyses due to right skew.
Health promoting talk. Based on a prior factor analysis, 3 codes were averaged across
friends to create one dyadic score (Iturralde, 2015). Codes were cautious health talk (personal
beliefs that favored limiting risk behavior), health promoting responses (verbal or non-verbal
reinforcement of the friend’s healthy decision-making), and interpersonal value talk (statements
emphasizing valued, normative aspects of peer relationships). The rated behaviors were
internally consistent (α = .81), with strong intra-dyad correlations (rs = .55–.79, ps < .001), and
good inter-rater reliability (ICC = .81). Health promoting talk was normally distributed.
Analytic approach
Means, standard deviations, and categorical counts were calculated and stratified by
gender and sexual activity status. In addition to descriptive statistics, associations were assessed
between stratification/predictor variables. As data were a mix of dyad-level and individual-level
variables, p values for correlations were corrected to adjust for statistical non-independence
using a pairwise dataset appropriate for dyads that are not distinguishable by a structural variable
such as gender (see Griffin & Gonzalez, 1995). Two multilevel regression models were run to
INTIMACY PROMOTING TALK 61
test associations between sexual risk history and intimacy promoting talk, one using logistic
regression for the dichotomous pregnancy involvement variable and the other Poisson regression
for the STI risk composite variable. Descriptive statistics examined the whole sample, whereas
regression analyses included only the 132 individuals in 82 dyads who reported past sexual
activity. (Inclusion of the non-sexually active youth would contribute to excess zeroes, which can
bias model estimates; see Atkins et al., 2013). Regression analyses also adjusted for gender, age,
neighborhood poverty, and race/ethnicity, as well as the three peer context variables: peers’
unprotected sex norms, the dyad partner’s sexual risk behavior, and dyadic health promoting
talk. Continuous predictors were centered prior to analyses. The dyad partner variable matched
the given analysis, e.g., friend’s past pregnancy involvement to predict youth’s own past
pregnancy involvement. When the dyad partner was not sexually active, this variable was
included and coded as 0. Gender and race/ethnicity interactions with intimacy promoting talk
were also tested. Talk variables were standardized to facilitate interpretation. Adjusted odds
ratios for logistic regression (and incident rate ratios for Poisson regression) were calculated to
indicate the change in risk associated with a 1 unit increase for the given predictor. To compare
the effects of intimacy promoting talk with those of the other peer context variables, standardized
coefficients were computed and Wald χ
2
tests were used to assess differences. Given the
clustered nature of the data, regression models were run using sandwich-estimated robust
standard errors to correct for non-independence in Mplus 7 (Muthén & Muthén, 2012).
Results
Descriptive Statistics and Bivariate Associations
Means, standard deviations, and count statistics are presented in Table 2 and stratified by
gender and sexual activity status. Just over half of study youth reported past sex, and 20.8% of
INTIMACY PROMOTING TALK 62
sexually active females and 11.4% of sexually active males reported past pregnancy
involvement. One quarter of sexually active youth endorsed at least one STI risk item at a
criterion level. One hundred of the sexually active youth (about three-quarters) were matched
with a sexually active friend. Youth who were not Caucasian, Asian, or in the other category
were at higher risk of reporting past pregnancy: 30.8% and 14.0% of African-American and
Latino youth, respectively, compared to 2.78% of Caucasian youth and 0 youth from other
racial/ethnic categories. Differences in STI risk score were not apparent among the three main
racial/ethnic groups in the sample.
Associations among stratification and predictor variables were assessed for the whole
sample (see Table 3). For these analyses, race/ethnicity for the three main groups was dummy
coded (0 or 1). Sexually active youth were older (r = .24, p < 01), had more peers having
unprotected sex (r = .43, p < .001), and used less health promoting talk (r = -.23, p < .01). Older
youth had riskier peer norms (r = .31, p < .001) and used less health promoting talk (r = -.30, p <
.001). Females used more intimacy promoting talk (r = .36, p < .001) but otherwise did not differ
from males on variables. Caucasian youth tended to be younger (r = -.17, p < .05), experienced
less neighborhood poverty (r = -.20, p < .001), had fewer peers having unprotected sex (r = -.22,
p < .05), used more health promoting talk (r = .24, p < .01), and were less likely to be matched in
the study with a friend who reported past pregnancy involvement (r = -.25, p < .001). Latino
youth were more likely to be sexually active than not (r = .18, p < .05) and used less health
promoting talk (r = -.27, p < .01). African-American youth experienced more neighborhood
poverty (r = .24, p < .01) and riskier peer norms (r = .20, p < .05), but used more intimacy
promoting talk (r = .20, p < .05); they were more likely to be matched in the study with a friend
who reported past pregnancy involvement (r = .27, p < .001). Health promoting talk was
INTIMACY PROMOTING TALK 63
negatively correlated with neighborhood poverty (r = -.20, p < .05) and peers’ unprotected sex
norms (r = -.25, p < .01). Friend’s past pregnancy involvement was related negatively to health
promoting talk (r = -.17, p < .05) and positively to peers’ unprotected sex norms (r = .26, p <
.001). Not shown in Table 3, friend’s STI risk count was positively related in bivariate Poisson
regression models to age (OR = 1.26, 95% CI = 1.04-1.48, p < .05) and peers’ unprotected sex
norms (OR = 1.44, 95% CI = 1.03-1.85, p < .05).
Intimacy promoting talk was unrelated to sexual activity status (r = -.05, p = .56), peer
group norms (r = -.01, p = .93), or friend’s past pregnancy involvement (r = -.09, p = .27). It was
marginally related in a positive direction to health promoting talk (r = .17, p = .07). In a
bivariate Poisson regression model, it was marginally, negatively related to friend’s STI risk
count (OR = .66, 95% CI = .38-.93, p = 052; not shown). Bivariate associations did not notably
differ by gender.
Intimacy Promoting Talk and Pregnancy
Results of the logistic regression model are shown in Table 4. As hypothesized, intimacy
promoting talk was associated with lower likelihood of past pregnancy involvement; a 1SD
increase in this variable was related to 56% decreased odds (AOR = .44; 95% CI = .24-.80).
Peers’ unprotected sex norms were also a significant predictor; a 1-point change in the peer
norms score was associated with twice the odds of past pregnancy (AOR = 2.01; 95% CI = 1.17-
3.44). Neither health promoting talk nor the friend’s past pregnancy involvement were related to
youths’ own odds of past pregnancy. Gender and race/ethnicity interactions with intimacy
promoting talk were not significant and were dropped from the model.

INTIMACY PROMOTING TALK 64
Intimacy Promoting Talk and STI Risk
Results of the Poisson regression model are also shown in Table 4 and largely mirrored
those for pregnancy involvement. Consistent with hypotheses, intimacy promoting talk was
associated with lower STI risk scores (IRR = .65; 95% CI = .39-.92), meaning that a 1SD
increase in intimacy promoting talk was associated with a 35% lower STI risk score. Peers’
unprotected sex norms were positively related to STI risk (IRR = 1.52; 95% CI = 1.15-1.89). No
other significant main effects, including for the other peer context variables, were found. As with
the pregnancy involvement model, gender and race/ethnicity interactions were not significant
and were trimmed.
Intimacy Promoting Talk Versus Other Peer Context Variables
Besides intimacy promoting talk, the only other significant peer context variable in
models was peers’ unprotected sex norms as perceived by youth. Standardized coefficients
revealed that the two predictors were of similar magnitude, although in opposite directions. Post
hoc Wald tests found that the effect sizes for these two predictors were statistically equivalent for
both the past pregnancy involvement model (χ
2
(1) = .09, p = .76) and the STI risk model (χ
2
(1) =
.002, p = .96).
Discussion

The purpose of the current study was to investigate associations between intimate
friendship processes and adolescents’ sexual risk behavior by using behavioral observations of
teen’s actual conversations. This is the first study to examine this question using coded
observations of adolescents with their close friends. Through a novel coding system, we
measured, in a reliable fashion, youths’ level of emotional disclosure, empathic support, and
perspective taking with a close friend—dimensions that have theoretical importance with regards
INTIMACY PROMOTING TALK 65
to promoting intimacy in peer relationships (Berndt, 1982; Buhrmester, 1990; Paul & White,
1990). Furthermore, this intimacy promoting talk was found to be associated with less likelihood
of past pregnancy involvement and lower STI risk scores, thereby expanding upon past research
that was based only on self-report measures of intimate friendship processes. Intimacy promoting
talk was a more salient predictor than the sexual risk behavior of the accompanying friend in the
discussion or the health promoting talk used by the dyad; it was as important as risky peer norms,
which is a commonly measured predictor, in explaining adolescents’ risk behavior. Associations
were not found to differ by gender or race/ethnicity, however, intimacy promoting talk was
observed at higher levels among female youth, as would be expected from past research on
gender differences in relationship intimacy (Rose & Rudolph, 2006). Intimacy promoting talk
was also seen more often in pairs containing at least one African-American youth. The relatively
modest number of African-American participants in the current study precludes drawing strong
inferences, however, this association should be investigated in future research.
This study contributes to growing evidence that intimate friendship processes may protect
youth from negative sexual health outcomes (Elkington et al., 2011; Miller et al., 2002; Ramiro
et al., 2013). Results support the concept that social-emotional functioning should be targeted by
sexual health interventions, along with more traditional objectives such as increasing knowledge
about condoms and contraception. This study also highlights specific communication behaviors
that fall under the intimacy promotion umbrella. Expression of vulnerable emotion, taking a
peer’s perspective, and providing emotional and empathic support are skills that may prove
useful for youth to learn as part of sexual health interventions. The importance of fostering social
competence as a way to mitigate the effects of negative social determinants (such as risky peers)
is compatible with a positive youth development philosophy (Gavin et al., 2010).
INTIMACY PROMOTING TALK 66
The current study took into account not just the individual, but also his or her social
context, including the communication behavior observed with a close friend, the friend’s risk
level, and the perceived riskiness of his or her peer group. It is noteworthy that even though
intimacy promoting talk was a dyadic construct, which combined strongly intercorrelated
measures from both individuals, it was significantly associated with individual-level sexual
health variables. The reciprocal nature of this talk illustrates that close friends in adolescence are
often quite similar to one another, in part because adolescents befriend likeminded peers but also
because of processes of mutual influence (Brechwald & Prinstein, 2011). Given the importance
of the social context in explaining adolescent sexual risk behavior, it is important to take an
ecological, multilevel perspective when identifying mechanisms of action and designing
interventions, as this approach has a better chance of creating a lasting effect (DiClemente,
Salazar, & Crosby, 2007; Salazar et al., 2010).
Several limitations of the current study require mention. Sexual health measures were
collected retrospectively, which restricts our understanding of causal direction. Future studies
should assess the trajectory of intimate friendship processes and sexual health behavior over time
to establish the nature of these effects. An additional limitation was that peer discussions
involved only one member of the youth’s peer group and were not specifically focused on sexual
health. Rather, the task was designed to elicit conversation about a range of interpersonal
concerns as well as related risk behavior (substance use). The intimate friendship processes
measured here may not generalize to other relationships in the participants’ peer groups or to
other types of discussions, such as partner sexual communication.
It remains unclear what the mechanisms are that explain an inverse association between
intimate friendship processes and adolescents’ sexual risk behavior. Some have forwarded a
INTIMACY PROMOTING TALK 67
compensatory hypothesis, i.e., that youth who lack companionship from close friends seek
intimacy with romantic partners and may take risks with their sexual health to pursue
relationship security (Crockett, Bingham, Chopak, & Vicary, 1996; Pereira et al., 2004). Another
explanation concerns underlying emotion regulation processes. As seen in the current study,
intimacy promoting talk requires not just emotional self-awareness and self-expression but also a
degree of sophistication in order to recognize and discuss a friend’s thoughts and feelings. When
these abilities are absent, risk behavior may fill the void. Past work has conceptualized sexual
risk behavior as a maladaptive strategy for coping with negative emotions (Brady et al., 2009;
Cooper, Shapiro, & Powers, 1998; Ethier et al., 2006) and has proposed the importance of
developing skills in social-emotional intelligence (Sieving et al., 2011) and emotional self-
efficacy (Valois, Zullig, Kammermann, & Kershner, 2013) to reduce sexual risks.
A final possibility is that intimate friendship processes are a prerequisite for other kinds
of interpersonal communication that reduce sexual risk behavior. Increasing evidence suggests
that sexual communication with romantic partners (Widman, Noar, Choukas-Bradley, & Francis,
2014) and with peers (Iturralde & Margolin, in preparation) may help youth increase protective
sexual behaviors. Having the skills to establish a foundation of emotional intimacy may help
facilitate such discussion. Past work has suggested that comfort and self-efficacy (Halpern-
Felsher, Kropp, Boyer, Tschann, & Ellen, 2004; Lefkowitz & Espinosa-Hernandez, 2007) in
discussing sexual matters with peers is related to greater condom use. Future work should
specifically measure these possible explanatory mechanisms as a means of further informing
intervention strategy.

INTIMACY PROMOTING TALK 68
References
Adolph, C., Ramos, D. E., Linton, K. L. P., & Grimes, D. A. (1995). Pregnancy among Hispanic
Teenagers: Is good parental communication a deterrent? Contraception, 51(5), 303-306.
doi:10.1016/0010-7824(95)00081-K
Ali, M. M., & Dwyer, D. S. (2011). Estimating peer effects in sexual behavior among
adolescents. Journal of Adolescence, 34(1), 183-190.
doi:10.1016/j.adolescence.2009.12.008
Alvarez, M. L., Burrows, R., Zvaighat, A., & Muzzo, S. (1987). Sociocultural characteristics of
pregnant and nonpregnant adolescents of low socioeconomic status: A comparative study.
Adolescence, 22(85), 149-156.
Atkins, D. C., Baldwin, S. A., Zheng, C., Gallop, R. J., & Neighbors, C. (2013). A tutorial on
count regression and zero-altered count models for longitudinal substance use data.
Psychology of Addictive Behaviors, 27(1), 166-177. doi:10.1037/a0029508
Benda, B. B., & Corwyn, R. F. (1999). Developmental differences in theories of sexual behavior
among rural adolescents residing in AFDC families. Deviant Behavior, 20(4), 359-385.
Berndt, T. J. (1982). The features and effects of friendship in early adolescence. Child
Development, 53, 1447-1460. doi:10.2307/1130071
Brady, S. S., Dolcini, M. M., Harper, G. W., & Pollack, L. M. (2009). Supportive friendships
moderate the association between stressful life events and sexual risk taking among African
American adolescents. Health Psychology, 28(2), 238-248. doi:10.1037/a0013240
Brechwald, W. A., & Prinstein, M. J. (2011). Beyond homophily: A decade of advances in
understanding peer influence processes. Journal of Research on Adolescence, 21(1), 166-
179. doi:10.1111/j.1532-7795.2010.00721.x
INTIMACY PROMOTING TALK 69
Buhi, E. R., & Goodson, P. (2007). Predictors of adolescent sexual behavior and intention: A
theory-guided systematic review. Journal of Adolescent Health, 40(1), 4-21.
doi:10.1016/j.jadohealth.2006.09.027
Buhrmester, D. (1990). Intimacy of friendship, interpersonal competence, and adjustment during
preadolescence and adolescence. Child Development, 61(4), 1101-1111.
doi:10.1111/j.1467-8624.1990.tb02844.x
Capaldi, D. M., & Clark, S. (1998). Prospective family predictors of aggression toward female
partners for at-risk young men. Developmental Psychology, 34, 1175-1188.
doi:10.1037/0012-1649.34.6.1175
Casares, W. N., Lahiff, M., Eskenazi, B., & Halpern-Felsher, B. L. (2010). Unpredicted
trajectories: The relationship between race/ethnicity, pregnancy during adolescence, and
young women’s outcomes. Journal of Adolescent Health, 47(2), 143-150.
doi:10.1016/j.jadohealth.2010.01.013
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation
analysis for the behavioral sciences (3rd ed.). Mahwah, NJ: Erlbaum.
Cooper, M. L., Shapiro, C. M., & Powers, A. M. (1998). Motivations for sex and risky sexual
behavior among adolescents and young adults: A functional perspective. Journal of
Personality and Social Psychology, 75(6), 1528-1558. doi:10.1037//0022-3514.75.6.1528
Crockett, L. J., Bingham, C. R., Chopak, J. S., & Vicary, J. R. (1996). Timing of first sexual
intercourse: The role of social control, social learning, and problem behavior. Journal of
Youth and Adolescence, 25(1), 89-111. doi:10.1007/BF01537382
DiClemente, R. J., Salazar, L. F., & Crosby, R. A. (2007). A review of STD/HIV preventive
interventions for adolescents: Sustaining effects using an ecological approach. Journal of
INTIMACY PROMOTING TALK 70
Pediatric Psychology, 32(8), 888-906. doi:10.1093/jpepsy/jsm056
Dishion, T. J., Andrews, D. W., & Crosby, L. (1995). Antisocial boys and their friends in early
adolescence: Relationship characteristics, quality, and interactional process. Child
Development, 66(1), 139-151. doi:10.1111/j.1467-8624.1995.tb00861.x
Eaton, D. K., Kann, L., Kinchen, S., Shanklin, S., Flint, K. H., Hawkins, J., … Chyen, D. (2012).
Youth risk behavior surveillance—United States, 2011. Morbidity and Mortality Weekly
Report (MMWR), 61(4), 1-162.
Elkington, K. S., Bauermeister, J. A., & Zimmerman, M. A. (2011). Do parents and peers
matter? A prospective socio-ecological examination of substance use and sexual risk
among African American youth. Journal of Adolescence, 34(5), 1035-1047.
doi:10.1016/j.adolescence.2010.11.004
Ethier, K. A., Kershaw, T. S., Lewis, J. B., Milan, S., Niccolai, L. M., & Ickovics, J. R. (2006).
Self-esteem, emotional distress and sexual behavior among adolescent females: Inter-
relationships and temporal effects. Journal of Adolescent Health, 38(3), 268-274.
doi:10.1016/j.jadohealth.2004.12.010
Finer, L. B. (2010). Unintended pregnancy among U.S. adolescents: Accounting for sexual
activity. Journal of Adolescent Health, 47(3), 312-314.
doi:10.1016/j.jadohealth.2010.02.002
Forhan, S. E., Gottlieb, S. L., Sternberg, M. R., Xu, F., Datta, S. D., McQuillan, G. M., …
Markowitz, L. E. (2009). Prevalence of sexually transmitted infections among female
adolescents aged 14 to 19 in the United States. Pediatrics, 124(6), 1505-1512.
doi:10.1542/peds.2009-0674
Gavin, L. E., Catalano, R. F., David-Ferdon, C., Gloppen, K. M., & Markham, C. M. (2010). A
INTIMACY PROMOTING TALK 71
review of positive youth development programs that promote adolescent sexual and
reproductive health. Journal of Adolescent Health, 46(3 Suppl), S75-S91.
doi:10.1016/j.jadohealth.2009.11.215
Gottman, J. M., & Mettetal, G. (1986). Speculations about social and affective development:
Friendship and acquaintanceship through adolescence. In J. M. Gottman & J. G. Parker
(Eds.), Conversations of friends: Speculations on affective development (pp. 192-237).
New York: Cambridge University Press.
Griffin, D., & Gonzalez, R. (1995). Correlational analysis of dyad-level data in the exchangeable
case. Psychological Bulletin, 118(3), 430-439. doi:10.1037/0033-2909.118.3.430
Halpern-Felsher, B. L., Kropp, R. Y., Boyer, C. B., Tschann, J. M., & Ellen, J. M. (2004).
Adolescents’ self-efficacy to communicate about sex: Its role in condom attitudes,
commitment, and use. Adolescence, 39(155), 443-456.
Henrich, C. C., Brookmeyer, K. A., Shrier, L. A., & Shahar, G. (2006). Supportive relationships
and sexual risk behavior in adolescence: An ecological-transactional approach. Journal of
Pediatric Psychology, 31(3), 286-297. doi:10.1093/jpepsy/jsj024
Henry, D. B., Deptula, D. P., & Schoeny, M. E. (2012). Sexually transmitted infections and
unintended pregnancy: A longitudinal analysis of risk transmission through friends and
attitudes. Social Development, 21(1), 195-214. doi:10.1111/sode.2012.21.issue-1
Iturralde, E. (2015). Not just talk: Observed communication in adolescent friendship and its
implications for health risk behavior (Doctoral dissertation). University of Southern
California, Los Angeles, CA.
Iturralde, E., & Margolin, G. (in preparation). Adolescents’ communication with friends about
sex: A systematic review of sexual health correlates.
INTIMACY PROMOTING TALK 72
Kia-Keating, M., Dowdy, E., Morgan, M. L., & Noam, G. G. (2011). Protecting and promoting:
An integrative conceptual model for healthy development of adolescents. Journal of
Adolescent Health, 48(3), 220-228. doi:10.1016/j.jadohealth.2010.08.006
Kotchick, B. A., Shaffer, A., Miller, K. S., & Forehand, R. (2001). Adolescent sexual risk
behavior: A multi-system perspective. Clinical Psychology Review, 21(4), 493-519.
Lando-King, E. A., McRee, A. L., Gower, A. L., Shlafer, R. J., McMorris, B. J., Pettingell, S., &
Sieving, R. E. (in press). Relationships between social-emotional intelligence and sexual
risk behaviors in adolescent girls. Journal of Sex Research.
doi:10.1080/00224499.2014.976782
Laurenceau, J. P., Barrett, L. F., & Pietromonaco, P. R. (1998). Intimacy as an interpersonal
process: The importance of self-disclosure, partner disclosure, and perceived partner
responsiveness in interpersonal exchanges. Journal of Personality and Social Psychology,
74(5), 1238-1251. doi:10.1037/0022-3514.74.5.1238
Lefkowitz, E. S., & Espinosa-Hernandez, G. (2007). Sex-related communication with mothers
and close friends during the transition to university. Journal of Sex Research, 44(1), 17-27.
doi:10.1080/00224490709336789
Miller, A. L., Notaro, P. C., & Zimmerman, M. A. (2002). Stability and change in internal
working models of friendship: Associations with multiple domains of urban adolescent
functioning. Journal of Social and Personal Relationships, 19(2), 233-259.
doi:10.1177/0265407502192004
Muthén, L. K., & Muthén, B. O. (2012). Mplus User’s Guide. Seventh Edition. Los Angeles, CA:
Muthén & Muthén.
Passino, A. W., Whitman, T. L., Borkowski, J. G., Schellenbach, C. J., Maxwell, S. E., Keogh,
INTIMACY PROMOTING TALK 73
D., & Rellinger, E. (1993). Personal adjustment during pregnancy and adolescent
parenting. Adolescence, 28(109), 97-122.
Paul, E. L., & White, K. M. (1990). The development of intimate relationships in late
adolescence. Adolescence, 25(98), 375-400.
Pereira, A. I., Canavarro, M. C., Cardoso, M. F., & Mendonça, D. (2004). Relational factors of
vulnerability and protection for adolescent pregnancy: A cross-sectional comparative study
of Portuguese pregnant and nonpregnant adolescents of low socioeconomic status.
Adolescence, 40(159), 655-671.
Prinstein, M. J., Boergers, J., & Spirito, A. (2001). Adolescents’ and their friends’ health-risk
behavior: Factors that alter or add to peer influence. Journal of Pediatric Psychology,
26(5), 287-298. doi:10.1093/jpepsy/26.5.287
Ramiro, M. T., Teva, I., Bermúdez, M. P., & Buela-Casal, G. (2013). Social support, self-esteem
and depression: Relationship with risk for sexually transmitted infections/HIV
transmission. International Journal of Clinical and Health Psychology, 13(3), 181-188.
Romer, D., Black, M., Ricardo, I., Feigelman, S., Kaljee, L., Galbraith, J., … Stanton, B. (1994).
Social influences on the sexual behavior of youth at risk for HIV exposure. American
Journal of Public Health, 84(6), 977-985. doi:10.2105/AJPH.84.6.977
Rose, A. J., & Rudolph, K. D. (2006). A review of sex differences in peer relationship processes:
Potential trade-offs for the emotional and behavioral development of girls and boys.
Psychological Bulletin, 132(1), 98-131. doi:10.1037/0033-2909.132.1.98
Salazar, L. F., Bradley, E. L., Younge, S. N., Daluga, N. A., Crosby, R. A., Lang, D. L., &
DiClemente, R. J. (2010). Applying ecological perspectives to adolescent sexual health in
the United States: Rhetoric or reality? Health Education Research, 25(4), 552-562.
INTIMACY PROMOTING TALK 74
doi:10.1093/her/cyp065
Salazar, L. F., Crosby, R. A., Diclemente, R. J., Wingood, G. M., Rose, E., Sales, J. M., &
Caliendo, A. M. (2007). Personal, relational, and peer-level risk factors for laboratory
confirmed STD prevalence among low-income African American adolescent females.
Sexually Transmitted Diseases, 34(10), 761-766. doi:10.1097/01.olq.0000264496.94135.ac
Sieving, R. E., McMorris, B. J., Beckman, K. J., Pettingell, S. L., Secor-Turner, M., Kugler, K.,
… Bearinger, L. H. (2011). Prime Time: 12-month sexual health outcomes of a clinic-
based intervention to prevent pregnancy risk behaviors. Journal of Adolescent Health,
49(2), 172-179. doi:10.1016/j.jadohealth.2010.12.002
Simoni, J. M., Nelson, K. M., Franks, J. C., Yard, S. S., & Lehavot, K. (2011). Are peer
interventions for HIV efficacious? A systematic review. AIDS and Behavior, 15(8), 1589-
1595. doi:10.1007/s10461-011-9963-5
Trussell, J., Henry, N., Hassan, F., Prezioso, A., Law, A., & Filonenko, A. (2013). Burden of
unintended pregnancy in the United States: Potential savings with increased use of long-
acting reversible contraception. Contraception, 87(2), 154-161.
doi:10.1016/j.contraception.2012.07.016
Tuttle, J., Bidwell-Cerone, S., Campbell-Heider, N., Richeson, G., & Collins, S. (2000). Teen
Club: A nursing intervention for reducing risk-taking behavior and improving well-being in
female African American Adolescents. Journal of Pediatric Health Care, 14(3), 103-108.
doi:10.1067/mph.2000.102320
United States Census Bureau. Selected economic characteristics. 2007-2011 American
Community Survey 5-Year Estimates. Retrieved 7/28/2013, from
www.factfinder.census.gov
INTIMACY PROMOTING TALK 75
Valois, R. F., Zullig, K. J., Kammermann, S. K., & Kershner, S. (2013). Relationships between
adolescent sexual risk behaviors and emotional self-efficacy. American Journal of
Sexuality Education, 8(1-2), 36-55. doi:10.1080/15546128.2013.790224
Weinstock, H., Berman, S., & Cates Jr, W. (2004). Sexually transmitted diseases among
American youth: Incidence and prevalence estimates, 2000. Perspectives on Sexual and
Reproductive Health, 36(1), 6-10. doi:10.1363/3600604
Widman, L., Choukas-Bradley, S., Helms, S. W., Golin, C. E., & Prinstein, M. J. (2014). Sexual
communication between early adolescents and their dating partners, parents, and best
friends. Journal of Sex Research, 51(7), 731-741. doi:10.1080/00224499.2013.843148
Widman, L., Noar, S. M., Choukas-Bradley, S., & Francis, D. B. (2014). Adolescent sexual
health communication and condom use: A meta-analysis. Health Psychology, 33(10), 1113-
1124. doi:10.1037/hea0000112

INTIMACY PROMOTING TALK 76
Table 1
Coded Talk Behaviors
Talk variable Code Examples
Intimacy
promoting
talk
Emotional
disclosure
I wish I handled situations like that better but I can’t just turn off
the feelings inside.
I don’t want you to get hurt because I care about you.
Taking
peer’s
perspective
So you’re like, why am I in this situation?
You may not see that side of him because you are closer to him
than I am.
Emotional
support
That’s hard that your father has moved out.
Aw man, don’t say you’re not attractive. That’s not true.

INTIMACY PROMOTING TALK 77
Table 2
Descriptive Statistics
Females (n = 100) Males (n = 122)
Any past sex? Any past sex?
Yes (n = 53) No (n = 47) Yes (n = 79) No (n = 43)
Study variables mean (SD) mean (SD) mean (SD) mean (SD)
Continuous
Demographics
Age 17.82 (1.26) 17.32 (1.14) 17.99 (1.37) 17.16 (1.36)
Neighborhood poverty (%) 11.97 (9.67) 12.84 (9.39) 11.73 (9.25) 9.40 (8.58)

Peer context
Intimacy promoting talk
†
.63 (.22) .62 (.31) .45 (.21) .46 (.22)
Peers’ unprotected sex norms 1.30 (1.32) .37 (.73) 1.60 (1.22) .45 (1.01)
Health promoting talk
†
.33 (.15) .38 (.15) .30 (.14) .39 (.17)

n (%) n (%) n (%) n (%)
Categorical and counts
Race/ethnicity
Caucasian 14 (26.4) 15 (31.9) 22 (27.8) 19 (44.2)
Latino 18 (34.0) 13 (27.7) 32 (40.5) 7 (16.3)
African-American 20 (37.7) 11 (23.4) 19 (24.1) 6 (14.0)
Asian or Pacific Islander 1 (1.9) 5 (10.6) 1 (1.3) 3 (7.0)
Other 0 (0) 3 (6.4) 5 (6.3) 8 (18.6)

Peer context
Friend’s past pregnancy
involvement
9 (17.0) 2 (4.3) 9 (11.4) 0 (0)
Friend’s STI risk score
0 45 (84.9) 43 (91.5) 60 (75.9) 42 (97.7)
1 5 (9.4) 4 (8.5) 15 (19.0) 1 (2.3)
2 3 (5.7) 0 (0) 3 (3.8) 0 (0)
3 0 (0) 0 (0) 1 (1.3) 0 (0)

Sexual risk behavior
Past pregnancy involvement 11 (20.8) 9 (11.4)
STI risk score
0 41 (77.4) 59 (74.7)
1 9 (17.0) 16 (20.3)
2 3 (5.7) 3 (3.8)
3 0 (0) 1 (1.3)
Note.
†
Dyadic measure is presented.
INTIMACY PROMOTING TALK 78
Table 3
Zero Order Correlations Among Stratification/Predictor Variables
Variables 1 2 3 4 5 6 7 8 9 10 11
1. Age —
2. Being female
†

-.05 —
3. Neighborhood poverty .11 .08 —
4. Caucasian -.17
*
-.05 -.20
**
—
5. Latino .05 -.01 .08 -.46
***
—
6. African-American .08 .12 .24
**
-.39
***
-.39
***
—
7. Intimacy promoting talk
†

-.15 .36
***
-.12 .01 -.15 .20
*
—
8. Peers’ unprotected sex norms .31
***
-.15 .14 -.22
*
.06 .20
*
-.01 —
9. Health promoting talk
†

-.30
***
.08 -.20
*
.24
**
-.27
**
-.08 .17 -.25
**
—
10. Being sexually active .24
**
-.12 .05 -.11 .18
*
.11 -.05 .43
***
-.23
**
—
11. Friend’s past pregnancy .15 .13 .05 -.25
***
.04 .27
***
-.09 .26
***
-.17
*
.14 —
Note. N = 222 (full sample).
†
Dyad-level measure is presented. Unless otherwise indicated, variable is at the individual level.
Pearson’s r correlations were computed with a pairwise dataset with significance values adjusted to account for dyadic non-
independence and indistinguishability. Race/ethnicity was dummy-coded (0 or 1).
*
p < .05,
**
p < .01,
***
p < .001.
INTIMACY PROMOTING TALK 79
Table 4
Multilevel Logistic and Poisson Regression Models for Pregnancy and STI Risk
Pregnancy involvement
†
STI risk
‡

Predictors Std. Coef. AOR (95% C.I.) p value Std. Coef. IRR (95% C.I.) p value
Demographics
Age .06 1.14 (.84-1.55) .41 .47 1.35 (1.18-1.52) <.001
Female .22 3.84 (1.06-13.90) .04 .15 1.29 (.46-2.12) .44
Neighborhood
poverty

-.21 .93 (.87-1.00) .052 -.40 .97 (.91-1.02) .22
Caucasian -.36 .42 (.08-2.33) .32 -.43 .76 (.31-1.21) .36
Latino .19 1.40 (.58-3.37) .45 .30 1.15 (.64-1.66) .54
African-American .80 5.85 (1.99-17.20) .001 .26 1.16 (.46-1.87) .62
Peer context
Intimacy
promoting talk
-.27 .44 (.24-.80) .007 -.52 .65 (.39-.92) .04
Peers’ unprotected
sex norms
.28 2.01 (1.17-3.44) .011 .63 1.52 (1.15-1.89) .001
Friend’s sexual
risk
.04 1.38 (.30-6.39) .68 -.24 .71 (.34-1.08) .20
Health promoting
talk
-.12 .69 (.37-1.26) .22 -.02 .99 (.68-1.30) .93
Note. N = 132 who reported any past sex. Models used
†
logistic regression and
‡
Poisson
regression. Friend’s sexual risk was past pregnancy involvement (pregnancy model) and STI risk
count (STI risk model).

80
General Discussion and Conclusions
The two studies presented here examined individual-level health phenomena in the
context of close, dyadic relationships with peers, bridging several bodies of complementary
literature. Numerous theories of health decision making stress the importance of one’s beliefs in
predicting behavior (Ajzen, 1991; Bandura, 1986; Fishbein & Ajzen, 1975). The peer context is
implicated in these models, but in the form of perception—what, according to the individual, is
customary or desirable among his or her peers (Fishbein, 2008)? Communication appears to be a
missing link; it is assumed to underlie many of these perceptual processes but is seldom
measured directly (Bell & Cox, 2015; Kincaid, 2004; Lapinski & Rimal, 2005). Also of
relevance here, ecological theory emphasizes the social milieu in which individuals approach
health decisions (Bronfenbrenner, 1979). However, within most ecological approaches,
investigations do not measure face-to-face interactions directly.
The present study incorporated not only peer group norms from the point of view of each
dyad member, but also captured a communication process in which close friends discussed
personal beliefs and preferences, thus taking a multilevel perspective. In the present research,
study design allowed for the use of dyadic analytic techniques to model associations between
individuals’ health behavior and dyad-level interpersonal phenomena, while accounting for
contextual variables such as neighborhood poverty, peer norms, and the risk behavior of the co-
participating friend.
The methodology employed in these studies also drew from social learning theory
(Bandura, 1977), which emphasizes the importance of observed behavior and the way social
interactions reinforce learning and shape cognition. It is known from intervention research that
speaking aloud one’s motivations and confidence to adopt healthier practices is predictive of
81
positive behavior change (i.e., change talk; Miller & Rollnick, 2013), although this work refers
to interactions between patient and clinician not those between friends. Using a novel approach,
the present research offered a window into how youth might co-construct health-related
cognitions with a close friend.
Health-Related Talk
In the first study, a diverse array of talk behaviors were identified based on review of the
literature and refinement of a coding system to measure this behavior. A large majority of youth
were found to engage in each of these disparate forms of talk, including favorable talk about risk
behavior, but also articulation of cautious health beliefs, expressions of disgust for risk behavior,
and endorsements of normative interpersonal values. Youth were also found to engage actively
with what their friends said by encouraging health-protective ideas, disputing risk-oriented ones,
and less frequently, reinforcing risky ideas. Confirmatory factor analysis found evidence for two
underlying factors, risk promotion and health promotion, rather than the presence of a single
bipolar dimension. Independent raters and youth themselves found these discussions to be
authentic. Their candid quality was further supported by the largely consistent associations
between youths’ talk and their self-reported risk behavior and other health-related indices.
Given that the discussion task prompted youth to talk about their views on alcohol,
tobacco, and marijuana, it was unsurprising that their risk promoting talk was associated with
past-month use of these substances and affiliation with friends who engage in such behavior.
This finding, although not directly testing peer influence, is consistent with a deviancy training
phenomenon; youth with more risk behavior and more deviant friends engage to a greater level
in mutually reinforcing discussions promoting risky ideas (Dishion & Tipsord, 2011).
Of note, however, was that health promoting but not risk promoting talk was also
82
associated with other constructs, which unlike substance use, were not of particular focus in the
peer discussion task prompts. These included being overweight or obese, school success, and
affiliation with prosocial peers. The significance of health promoting talk across domains is
consistent with other research indicating that a health-oriented cognitive profile (positive
attitudes about health, concerns about health) and a desire to align behavior with normative
social norms (“conventionality”) is related to a broad array of positive health outcomes (León,
Carmona, & García, 2010; Wiefferink et al., 2006). In contrast, a risk-oriented profile has been
found to strongly predict involvement with problem behaviors specifically, and only weakly
predict other kinds of health behavior outcomes (Turbin, Jessor, & Costa, 2000). Therefore,
models that incorporate both protective and risk-oriented dimensions can offer a more
comprehensive picture of adolescent health (Kia-Keating, Dowdy, Morgan, & Noam, 2011).
Intimacy Promoting Talk
Peer contextual factors, such as having supportive friends or having friends who do not
use condoms, have been implicated in studies of sexual risk behavior (Henrich, Brookmeyer,
Shrier, & Shahar, 2006; Romer et al., 1994), although more research is needed to examine these
peer microsystem processes (Salazar et al., 2010). The second study presented here took a closer
look at intimate friendship processes, which were considered alongside other peer context
variables. An innovation forwarded by this study was the use of a behavioral definition of
intimate friendship processes, namely intimacy promoting talk. This definition drew upon
developmental and interpersonal models of friendship intimacy (Berndt, 1982; Buhrmester,
1990; Paul & White, 1990), as well as past behavioral observation conducted with adolescent
dyads (Gottman & Mettetal, 1986). As with health promoting talk, intimacy promoting talk was
a dyadic phenomenon. The degree to which youth disclosed emotional vulnerability to the friend
83
was correlated with friends’ use of perspective taking and emotional support, and vice versa.
Conceptually it is understandable why this process would be reciprocal and bidirectional. One
youth is more likely to divulge emotional information if he has learned from experience that his
friend tends to be emotionally supportive; likewise, an emotionally supportive response is
contingent on the friend sharing information that warrants such a response. This dyadic process
was inversely associated with individuals’ history of past pregnancy involvement and scores on a
STI risk composite variable.
In several ways, these findings expand upon past research demonstrating a protective role
for intimate friendship processes among youth vulnerable to sexual health risks (Elkington,
Bauermeister, & Zimmerman, 2011; Miller, Notaro, & Zimmerman, 2002; Ramiro, Teva,
Bermúdez, & Buela-Casal, 2013). Firstly, prior research has relied upon individuals’ ratings of
friendship characteristics, thus providing a vague understanding of what the relevant relationship
behaviors might be. Thus, the current study provides some specific communication processes
that could be explored as part of future research and intervention design. Secondly, other peer
context variables were examined alongside intimacy promoting talk, specifically peer norms for
unprotected sex, the co-participating friend’s own risk behavior, and youths’ degree of health
promoting talk as conceptualized in Study 1. Intimacy promoting talk was associated with sexual
health variables even after adjusting for peer norms, the friend’s sexual risk, health promoting
talk, age, gender, race/ethnicity, and neighborhood poverty.
It is remarkable that intimacy promoting talk, which was essentially improvised by the
youth during a discussion task, would be related to a highly consequential health issue such as
early pregnancy and STI risk. However, this study was designed to elicit intimate processes from
youth to the extent that this was a natural process in their friendship. Prior to the discussion task,
84
youth were individually instructed to consider close others with whom they had an unresolved
issue, and then they were reminded of their responses before a discussion of interpersonal
problems with their friend. Therefore, youth were primed to divulge information with emotional
qualities. This ability to identify one’s emotions and share them with a close other has been
identified as a protective skill in past sexual health research (Lando-King et al., in press). It was
also important in this study to examine intimacy promoting talk as a separate phenomenon from
other talk that was specifically health-related. Thus, raters were instructed not to score, for
example, emotional support for talk that was health-oriented. Indeed, intimacy promoting talk
was uncorrelated with the three other peer context variables, which were all explicitly health-
related (peer norms, friend’s sexual risk behavior, and health promoting talk). These findings
illustrate the multidimensional nature of peer communication as it pertains to adolescent health.
There appears to be a role for discussion of health-related beliefs and preferences, but other,
social-emotional qualities of the relationship may also be important.
Clinical Implications
The present research has implications for prevention efforts. Just as health behavior is
often conceptualized at the individual level, health interventions also tend to view success in
terms of individual-level changes in knowledge, attitudes, and behavior (Salazar et al., 2010).
The current research illustrates, however, that intervention at the relationship level could have
beneficial impacts on adolescents’ health. Recent health interventions, for example, have focused
on building social networks to disseminate health information (Wang, Brown, Shen, & Tucker,
2011), trained youth in improving relationship skills to reduce coercive peer processes (Wolfe,
Crooks, Chiodo, Hughes, & Ellis, 2012), and targeted partner sexual communication as an area
of change (Sales et al., 2012). With regards to how youth talk to friends, the current research
85
offers some possibilities of how peer communication could be shaped to improve adolescents’
own and their friends’ health behavior. The potential strategies suggested by this work are
diverse and include expressing personal health beliefs, promoting health concepts to friends,
orienting oneself towards more normative interests, engaging in empathic communication, and
even rejecting risk behavior based on personal preference.
Limitations and Future Directions
Several limitations of the current research must be highlighted, as they could be
addressed in future studies. A key problem within these studies was the cross-sectional level of
analysis. Past longitudinal research has suggested that peer context variables, such as social
norms or deviant communication, exert an effect on youths’ health behavior through pathways of
influence and selection, i.e., youth change their behavior to conform with friends’ own, however,
youth also select friends from the start with similar propensities (Brechwald & Prinstein, 2011).
Although many of the associations found here could be explained by influence and selection
processes, these mechanisms were not tested directly. Prospective or even experimental designs
(see Cohen & Prinstein, 2006) would be helpful in examining the role of influence and selection
in peers’ health-related and intimacy promoting talk. Also in service of this goal, it might be
valuable to study communication within other types and configurations of peer relationships
besides close friendships (e.g., acquaintances, social networks, romantic relationships). Talk
behavior among acquaintances, for example, might not have the strongly dyadic, reciprocal
quality seen among close friends, and might therefore allow for more observation of influence
processes as they unfold over time, such as would be tested through an actor-partner
interdependence model (Kenny, Kashy, & Cook, 2006).
Also with the goal of understanding mechanisms, future research on health-related and
86
intimacy promoting talk would benefit from assessing mediators and moderators of their
associations with health behavior. Future studies would benefit from integrating talk behavior
variables into models that include social cognitive constructs. Adolescents’ health perceptions,
including their attitudes, expectancies, sense of control, self-efficacy, and intentions, as well as
their understanding of peer norms, have been found to predict later health choices (Bandura,
1986; Fishbein, 2008). The role of peer communication among these constructs merits further
investigation. In a few studies, communication variables have been found to provide additional
explanatory power when combined with social cognitive measures (Dunlop, 2011; Hornik &
Yanovitzky, 2003; Hwang, 2012; Real & Rimal, 2007; van den Putte, Yzer, Southwell, de
Bruijn, & Willemsen, 2011). In addition, given the relevance of an ecological perspective in
understanding adolescent health behavior, it would be beneficial to integrate other contextual
variables such as family, school, or neighborhood environments, as past research has found
interactive effects between peer relationship variables and other ecological factors (Brady,
Dolcini, Harper, & Pollack, 2009; Henrich et al., 2006)
Conclusions
Overall, the studies presented here offer a novel perspective on how close adolescent
friends’ face-to-face communication has direct relevance to highly consequential health
outcomes, including substance use, sexual risk behavior, excess weight, and social determinants
of health. This research is unusual in its multilevel design and its integration of behavioral
observation methods with peer contextual variables. It is also the only set of studies to our
knowledge to use behavioral observation to examine health promoting and intimacy promoting
talk as they relate to adolescents’ health behavior. These studies offer specific descriptions of
behaviors that may be adaptive for youth and that may be integrated into interventions focused
87
on improving youths’ relationships and communication skills. Future research should investigate
the effects of close friends’ talk on health over time and investigate potential mediators and
moderators of these effects.

88
References for General Discussion
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision
Processes, 50(2), 179-211.
Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice Hall.
Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory.
Englewood Cliffs, NJ: Prentice Hall.
Bell, D. C., & Cox, M. L. (2015). Social norms: Do we love norms too much? Journal of Family
Theory & Review, 7(1), 28-46. doi:10.1111/jftr.12059
Berndt, T. J. (1982). The features and effects of friendship in early adolescence. Child
Development, 53, 1447-1460. doi:10.2307/1130071
Brady, S. S., Dolcini, M. M., Harper, G. W., & Pollack, L. M. (2009). Supportive friendships
moderate the association between stressful life events and sexual risk taking among African
American adolescents. Health Psychology, 28(2), 238-248. doi:10.1037/a0013240
Brechwald, W. A., & Prinstein, M. J. (2011). Beyond homophily: A decade of advances in
understanding peer influence processes. Journal of Research on Adolescence, 21(1), 166-
179. doi:10.1111/j.1532-7795.2010.00721.x
Bronfenbrenner, U. (1979). The ecology of human development. Cambridge, MA: Harvard
University Press.
Buhrmester, D. (1990). Intimacy of friendship, interpersonal competence, and adjustment during
preadolescence and adolescence. Child Development, 61(4), 1101-1111.
doi:10.1111/j.1467-8624.1990.tb02844.x
Cohen, G. L., & Prinstein, M. J. (2006). Peer contagion of aggression and health risk behavior
among adolescent males: An experimental investigation of effects on public conduct and
89
private attitudes. Child Development, 77(4), 967-983. doi:10.1111/j.1467-
8624.2006.00913.x
Dishion, T. J., & Tipsord, J. M. (2011). Peer contagion in child and adolescent social and
emotional development. Annual Review of Psychology, 62, 189-214.
doi:10.1146/annurev.psych.093008.100412
Dunlop, S. M. (2011). Talking “truth”: Predictors and consequences of conversations about a
youth antismoking campaign for smokers and nonsmokers. Journal of Health
Communication, 16(7), 708-725. doi:10.1080/10810730.2011.552000
Elkington, K. S., Bauermeister, J. A., & Zimmerman, M. A. (2011). Do parents and peers
matter? A prospective socio-ecological examination of substance use and sexual risk
among African American youth. Journal of Adolescence, 34(5), 1035-1047.
doi:10.1016/j.adolescence.2010.11.004
Fishbein, M. (2008). A reasoned action approach to health promotion. Medical Decision Making,
28(6), 834-844. doi:10.1177/0272989X08326092
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to
theory and research. Reading, MA: Addison-Wesley.
Gottman, J. M., & Mettetal, G. (1986). Speculations about social and affective development:
Friendship and acquaintanceship through adolescence. In J. M. Gottman & J. G. Parker
(Eds.), Conversations of friends: Speculations on affective development (pp. 192-237).
New York, NY: Cambridge University Press.
Henrich, C. C., Brookmeyer, K. A., Shrier, L. A., & Shahar, G. (2006). Supportive relationships
and sexual risk behavior in adolescence: An ecological-transactional approach. Journal of
Pediatric Psychology, 31(3), 286-297. doi:10.1093/jpepsy/jsj024
90
Hornik, R., & Yanovitzky, I. (2003). Using theory to design evaluations of communication
campaigns: The case of the National Youth Anti‐Drug Media Campaign. Communication
Theory, 13(2), 204-224. doi:10.1111/j.1468-2885.2003.tb00289.x
Hwang, Y. (2012). Social diffusion of campaign effects: Campaign-generated interpersonal
communication as a mediator of antitobacco campaign effects. Communication Research,
39(1), 120-141. doi:10.1177/0093650210389029
Kenny, D., Kashy, D., & Cook, W. (2006). Dyadic data analysis. New York, NY: The Guilford
Press.
Kia-Keating, M., Dowdy, E., Morgan, M. L., & Noam, G. G. (2011). Protecting and promoting:
An integrative conceptual model for healthy development of adolescents. Journal of
Adolescent Health, 48(3), 220-228. doi:10.1016/j.jadohealth.2010.08.006
Kincaid, D. L. (2004). From innovation to social norm: Bounded normative influence. Journal of
Health Communication, 9(S1), 37-57. doi:10.1080/10810730490271511
Lando-King, E. A., McRee, A. L., Gower, A. L., Shlafer, R. J., McMorris, B. J., Pettingell, S., &
Sieving, R. E. (in press). Relationships between social-emotional intelligence and sexual
risk behaviors in adolescent girls. Journal of Sex Research.
doi:10.1080/00224499.2014.976782
Lapinski, M. K., & Rimal, R. N. (2005). An explication of social norms. Communication Theory,
15(2), 127-147. doi:10.1111/j.1468-2885.2005.tb00329.x
León, J., Carmona, J., & García, P. (2010). Health-risk behaviors in adolescents as indicators of
unconventional lifestyles. Journal of Adolescence, 33(5), 663-671.
doi:10.1016/j.adolescence.20
Miller, A. L., Notaro, P. C., & Zimmerman, M. A. (2002). Stability and change in internal
91
working models of friendship: Associations with multiple domains of urban adolescent
functioning. Journal of Social and Personal Relationships, 19(2), 233-259.
doi:10.1177/0265407502192004
Miller, W. R., & Rollnick, S. (2013). Motivational interviewing: Helping people change (3rd
ed.). New York, NY: Guilford Press.
Paul, E. L., & White, K. M. (1990). The development of intimate relationships in late
adolescence. Adolescence, 25(98), 375-400.
Ramiro, M. T., Teva, I., Bermúdez, M. P., & Buela-Casal, G. (2013). Social support, self-esteem
and depression: Relationship with risk for sexually transmitted infections/HIV
transmission. International Journal of Clinical and Health Psychology, 13(3), 181-188.
Real, K., & Rimal, R. N. (2007). Friends talk to friends about drinking: Exploring the role of
peer communication in the theory of normative social behavior. Health Communication,
22(2), 169-180. doi:10.1080/10410230701454254
Romer, D., Black, M., Ricardo, I., Feigelman, S., Kaljee, L., Galbraith, J., … Stanton, B. (1994).
Social influences on the sexual behavior of youth at risk for HIV exposure. American
Journal of Public Health, 84(6), 977-985. doi:10.2105/AJPH.84.6.977
Salazar, L. F., Bradley, E. L., Younge, S. N., Daluga, N. A., Crosby, R. A., Lang, D. L., &
DiClemente, R. J. (2010). Applying ecological perspectives to adolescent sexual health in
the United States: Rhetoric or reality? Health Education Research, 25(4), 552-562.
doi:10.1093/her/cyp065
Sales, J. M., Lang, D. L., DiClemente, R. J., Latham, T. P., Wingood, G. M., Hardin, J. W., &
Rose, E. S. (2012). The mediating role of partner communication frequency on condom use
among African American adolescent females participating in an HIV prevention
92
intervention. Health Psychology, 31(1), 63-69. doi:10.1037/a0025073
Turbin, M. S., Jessor, R., & Costa, F. M. (2000). Adolescent cigarette smoking: Health-related
behavior or normative transgression? Prevention Science, 1(3), 115-124.
doi:10.1023/A:1010094221568
van den Putte, B., Yzer, M., Southwell, B. G., de Bruijn, G. J., & Willemsen, M. C. (2011).
Interpersonal communication as an indirect pathway for the effect of antismoking media
content on smoking cessation. Journal of Health Communication, 16(5), 470-485.
doi:10.1080/10810730.2010.546487
Wang, K., Brown, K., Shen, S. Y., & Tucker, J. (2011). Social network-based interventions to
promote condom use: A systematic review. AIDS and Behavior, 15(7), 1298-1308.
doi:10.1007/s10461-011-0020-1
Wiefferink, C. H., Peters, L., Hoekstra, F., Dam, G. T., Buijs, G. J., & Paulussen, T. G. (2006).
Clustering of health-related behaviors and their determinants: Possible consequences for
school health interventions. Prevention Science, 7(2), 127-149. doi:10.1007/s11121-005-
0021-2
Wolfe, D. A., Crooks, C. V., Chiodo, D., Hughes, R., & Ellis, W. (2012). Observations of
adolescent peer resistance skills following a classroom-based healthy relationship program:
A post-intervention comparison. Prevention Science, 13(2), 196-205. doi:10.1007/s11121-
011-0256-z

93
Appendix A.

Pre-discussion questionnaire

ID#_______%
%
Discussion(Topics(
Wave%5a%
%
We%all%have%people%in%our%lives%we%interact%with%on%a%regular%basis.%Sometimes%
we%wish%we%could%improve%our%interactions%with%these%people.%%
%
From%the%list%below,%put%a%check%mark%(√)%next%to%people%who%you%have%an%
unresolved(issue(or(problem%with.%%
%
_____Brother%or%Sister% % % % % _____Boyfriend%or%Girlfriend%
%
_____Mom%%%%%%%%%%%%%%% % % % % % _____Dad%
%
_____Friend%%%%%%%%%%% % % % % % _____Teammate%
%
_____Coach%%%%%%%%%%% % % % % % _____Boss%
%
_____Classmate%%%%%%%%%% % % % % % _____Teacher%
%
_____Someone%you%have%a%crush%on% % % % _____School%
principle/administrator%
%
_____Relative:%%Who?%____________________%%%%%%%%
%
_____Other:%%Who?%____________________%%%%%%%% % % % % %
%

94
Appendix B.
Peer Discussion Task Script%

95
Appendix B.
Peer Discussion Task Script (continued)

96
Appendix B.
Peer Discussion Task Script (continued)

97
Appendix C.
Post-Discussion questionnaire

ID #:

_____ - 3 (Youth)

_____ - 9 (Peer)

Please circle an answer for each of the questions.

1. HOW SIMILAR WAS THIS DISCUSSION TO OTHER DISCUSSIONS YOU HAVE HAD
WITH YOUR FRIEND?

Not at all
Similar
Slightly
Similar
Somewhat
Similar
Moderately
Similar
Very
Similar
0 1 2 3 4

2. HOW MUCH WERE YOU ABLE TO EXPRESS YOUR OWN POINT OF VIEW ON
THE TOPICS?

Not at all

A Little

Some
Moderate
Amount

A Lot
0 1 2 3 4

3. HOW OFTEN DO YOU HAVE DISCUSSIONS LIKE THIS WITH THIS FRIEND?

Never Once in a While Sometimes Often Very Often
0 1 2 3 4

4. HOW HONEST OR FRANK WERE YOU DURING THE DISCUSSION?

Not at all

A Little

Some
Moderate
Amount

Very
0 1 2 3 4

98
Appendix D.
Coding Manual Excerpts
II. General Procedure
• Discussions for individual topics all take 5 minutes each. The coding sheet has a place to record the starting
time for each topic, which begins when at least one of the two youth begins talking. Observations should not
be coded beyond 5 minutes past that starting time.
• Raters watch one youth at a time. Therefore, they watch each discussion at least twice, once to code for
their designated first youth, and a second time to code for their designated second youth. If necessary,
raters may rewind to rewatch difficult sections of the discussion.
• For every dyad, a coding log lists which of the two youth should be coded first for Topic 1. This “Youth 1” is
counterbalanced across dyads and alternated between coders. The target youth (T) is always the youth who
presents “the problem with a person” in Discussion 1. The invited peer (P) presents a problem in Discussion
2.
• After Topic 1, raters switch which youth they code first per topic. For example, if the rater codes Youth 1 first
for Topic 1 (and therefore Youth 2 second for Topic 1), then for Topic 2, the rater codes Youth 2 first,
followed by Youth 1. The following table illustrates this switching using an example:

Example of Switching First Youth to Code!

• On a single coding sheet, raters record a score (0-3) for each youth for each code for every discussion topic,
1 through 8, except...
o For one code, Faking, each youth is scored a 0 or 1 per topic.
o For two codes, Closeness and Likemindedness, a 0-3 score is assigned to the dyad, not to
each youth individually. These codes should be scored after the rater has finished scoring both
youth for the individual codes.

99
Appendix D.
Coding Manual Excerpts (continued)
III. Discussion Topics
1. Target youth problem
2. Invited peer problem
3. Substances
4. Goals
5. Dating
6. Friends
7. Change one thing about self
8. Planning a party
9. Sex and hooking up (Wave 5A only)
A. Discussion Topic Instructions !
This task is adapted from Dishion et al. (2002)
%
#1: Target youth problem
“Now I’d like the two of you to talk about a current problem with a person that [target youth] identified a few minutes
ago. [Target youth], you selected a few people on this list with whom you have an unresolved issue. Please talk about
why it is a problem and then if you’ve tried to solve it what you did and if it worked. Then talk with [invited peer] about
ways you might solve the problem and any ways that [invited peer] could help.”
#2: Invited peer problem
“Now I’d like the two of you to talk about a current problem with a person that [invited peer] identified a few minutes
ago. [Invited peer], you selected a few people on this list with whom you have an unresolved issue. Please talk about
why it is a problem and then if you’ve tried to solve it what you did and if it worked. Then talk with [target youth] about
ways you might solve the problem and any ways that [target youth] could help.”
#3: Substances
“For the next 5 minutes please talk about your beliefs about drinking alcohol, and using tobacco, marijuana, and other
drugs. Please talk about each one separately. If you think that use is appropriate for people your age, please say why
and in what settings it is appropriate to drink alcohol, and use tobacco, marijuana, and other drugs. Again, please talk
about each separately.”
#4: Goals
“For the next 5 minutes we would like each of you to talk about a major goal you have for the next year. We would
like you each to describe your goal, why it’s important and how you are going to get there. You can also react to each
other’s goals, or talk about how you can help each other reach your goals. Please make sure that each of you has an
opportunity to describe your goal.”
#5: Relationships
“This next activity is about dating. Please discuss some of the things you like and don’t like about people you might
date. Please discuss personality traits that you like or dislike, or things that they do. Think about the whole person,
not just one characteristic.”
#6: Friends
“This next topic is about the group of friends you spend time with. We’d like you to describe your friends, and the
kinds of things you like to do together. Then talk about what you like and don’t like about the friends you spend time
with. If you don’t hang out with a group, what group would you like to spend time with and why?”
#7: Change one thing about self
“The next topic is about something you would change about yourself. For the next five minutes, we would like each of
you to discuss one thing you would like to change about yourself.”
#8: Planning a Party
“For the last 5 minutes we’d like you to plan a party. This is a party you would have at one of your houses. Please talk
about who would be there, what you would do, and about how long it would last and anything else that you think is
important.”
#9: Sex and Hooking Up (Wave 5A only)
“The final topic for discussion today is sex. For the next 5 minutes, please talk about your beliefs about sex, hooking up, and
sexual responsibility for people your age.”
100
Appendix D.
Coding Manual Excerpts (continued)
IV. Coding Scale & Definitions
The following scoring guide applies to most codes. For scoring guides that are specific to individual codes, see Part
V.C.
%
In many cases, the definitions below (None, Some, A moderate amount, A lot) are enough to know what score is
appropriate for a given code. The additional Description information is provided to assist the rater in deciding more
difficult cases, such as when trying to decide between a 2 or a 3 on a given code.
%
Code Definition Description
(See below for definitions of terms used in this column)
0 None No hints, statements, or non-verbal indications of the behavior
1 Some 1-2 hints OR
1 statement or non-verbal cue of regular intensity
2 A moderate amount 3+ hints OR
2 statements of regular intensity OR
2+ non-verbal cues of regular intensity OR
1 statement of high intensity
3 A lot 3+ statements or non-verbal cues of regular intensity OR
2+ statements or non-verbal cues of high intensity OR
1+ statement or non-verbal cue of extremely high intensity
A. Definitions of Terms!
Hint
• Vague or ambiguous statements that might match the code but are difficult to interpret
• Statements that match the code in content but not affect, or in affect but not content
• Bound by ordinary grammatical constraints, such as sentences or sentence fragments. One speaker turn may
contain multiple hints.
• Non-verbal cues that are ambiguous or contain incongruent affect. Bound by the behavior, or if continuous, by
speaker turn. Could be coded while the youth is speaking or while listening.

Examples
“I am un-serious about everything.” (Emotional Disclosure)
“It makes her feel less bad if I do it too.” (Perspective Taking of Others)
Statement
• A verbal remark that matches the code
• Bound by ordinary grammatical constraints, such as sentences or sentence fragments
• One speaker turn may contain multiple statements

Examples
“Now you’ll get the distance you need and you can focus on yourself.” (Value Finding, Regular Intensity)
“I wish I handled situations like that better but I can’t just turn off the feelings inside.” (Emotional Disclosure,
Regular Intensity)
“All our best times have been when we’ve been wasted.” (Positive Talk: Substances, High Intensity)
“I got my girlfriend pregnant but she had an abortion.” (Emotional Disclosure, Risky Talk: Sex, High
Intensity)
101
Appendix D.
Coding Manual Excerpts (continued)
* Extremely high intensity describes communication behavior that is very low base rate given the lab setting,
the topic, and the fact that the youth is talking to a friend.
%
Non-Verbal Cue
• Non-verbal behavior that matches the code
• Bound by the behavior, or if continuous, by speaker turn
• Could be coded while the youth is speaking or listening

Examples
laughing, smiling, frowning, nodding, shaking of head, eye rolling (regular intensity)
shouting, crying (high intensity)
leaving the room in anger (extremely high intensity)
V. The Codes
C. Code Definitions With Examples
%
GENERAL INTERACTION BEHAVIORS
%
Emotional Disclosure
• Openness about experiences, thoughts, and feelings that express vulnerability
• Could include use of feeling words (e.g. sad) and related judgments (e.g. “It’s hard for me”)
• Could involve candid discussion of experiences, beliefs, or attributions that seem to cause the youth
embarrassment, sadness, or other negative emotions
• Could contain expressions of love, affection or closeness
• Could consist of non-verbal cues such as crying or frowning

Examples
“I’m super shy.”
“The whole future thing scares the shit out of me.”
“I can’t make the first move.”
“I don’t appreciate her making me feel inadequate.”
“Having a lot of people around me makes me feel lonely.”
“That made me bond with my dad a little more.”
“I’ve been missing having someone being there. I feel like I made a mistake with my ex-girlfriend and I got
really close to her.”
“I wish I could be that close to my parents.”
“I wish I handled situations like that better but I can’t just turn off the feelings inside.”
“I don’t like that he tries to put me down in front of other people.”
“I love them so much. I’m so sad. I cried after Denny’s.”
“I’m un-serious about everything.”
102
Appendix D.
Coding Manual Excerpts (continued)
Youth makes a sad face and looks down at the floor when voicing that he has never dated before.
Scoring
See Part IV, Coding Scale & Definitions
%
Faking (0 or 1)
• This should be scored as a 1 only if the youth’s participation in the discussion topic appeared significantly
compromised by the following features
• Stiffness, unnatural speech, artificiality
• Interacting with the camera or performing for an imaginary audience instead of talking to the peer

Scoring
Score as a 0 or 1. The score is based on the tone of the whole discussion, not just that of a few statements.
A 1 signifies that the youth was so awkward in the task that you question the authenticity and validity of what
he or she said.
%
POSITIVE & NEGATIVE SUPPORT BEHAVIORS
%
These codes are applied to youth’s responses to his/her friend in the discussion. Responses could also be made in
response to a response.
%
Engagement
• Communication behaviors that indicate a high degree of interest in what the peer is saying
• Could include verbal and non-verbal responses indicating that the youth is paying attention
• More engagement can be assumed from higher energy and arousal, which may be indicated by louder voice
tone, laughing, more frequent movement, sitting up straight, etc.
• More engagement can be assumed when the dyad appears to run out of time for the topic, as though they would
continue discussing the topic intently if we gave them more time.

Examples
Acknowledging and responding relevantly to comments
Use of nonspecific facilitators to keep interaction going (head nodding, uh-huh, yeah, smiling at friend, eye
contact, etc.)
Behaviors can also be frowning, eye-brow raise, and other seemingly negative behaviors
Finishing peer’s sentences, and other friendly interruptions that denote interest in what the friend is saying
Scoring
0 Apparent low interest and low energy. Low use of engagement behavior.
1 The youth is going through the motions but does not appear that interested. The youth appears to
pay attention to some of the friend’s statements, but not more than 25%.
2 Moderate energy and frequency of engagement behavior, e.g. listening but not affirming or
responding to every
statement made by the peer. The youth responds to 25-75% of statements.
3 High energy and high frequency of engagement behavior, e.g. responses to 75+% of statements.
%
Emotional Support
• Could include expressions of empathy, sympathy, and caring
• Making “empathic sounds” such as awwwww, etc.
• Non-verbal mirroring of vulnerable emotion such as sadness

103
Appendix D.
Coding Manual Excerpts (continued)
• Statements that capture the emotion of the peer but are not quite as cognitively articulated as perspective
taking (e.g. “That’s hard”)
• Could include a statement of understanding based on the youth’s own experience
• Could involve reflecting the emotional content of the peer’s disclosure
• Emotional support in response to peer’s Negative View of Self is counted under this code.
• Could contain humor that is supportive and friendly

Examples
[In response to youth saying that she will miss her boyfriend when she moves away] “Do you love him?”
[In response to peer’s Negative View of Self comment “I need a nose job”] “I don’t think you do, your nose is
fine, but if it makes you feel better then do it.”
[In response to youth saying that he’s not very attractive] “Aw man, don’t say that. That’s not true.”
[In response to youth saying that he can’t make the first move in dating] “Yeah, I can’t do that either.”
Scoring
See Part IV, Coding Scale & Definitions
%
Taking Peer’s Perspective
• Imagining or wondering about the motivations, feelings, beliefs, or experiences of peer
• Anticipating what the peer might think, say, or do
• Thinking about why the peer behaves the way he/she does
• Can be related to own disclosure or in response to friend’s account

Examples
“So you’re like, why am I here now?”
%
“You may not see that side of him because you are closer to him than I am”
%
“Do you think you act that way because of your lack of caring in the relationship?”
Scoring
See Part IV, Coding Scale & Definitions
%
TALK ABOUT RISK
%
The following codes are inspired by and adapted from the Conversation Topic Code (CTC; Poe et al., 1990), which
labels discussion content as deviant, prosocial, and normative.
%
Riskier Talk
%
Positive Talk About Substances
• This code should be applied to any remarks that indicate positive feelings or beliefs about a substance,
interest in having a substance, or past experience with a substance, even if you think the level of use is
typical for someone that age
• Apply this code even when the positive beliefs espoused by the youth are defensible based on current
popular understanding (e.g. “A glass of wine per day is good for you.”)
• Apply this code to statements that appear to be rationalizations for why risk behavior is acceptable given
certain conditions, but that do not offer well-developed details about self-restraint (e.g. “It’s OK to drink if you
are with your friends”). See Cautious Health Talk for more details.
104
Appendix D.
Coding Manual Excerpts (continued)
• Also apply to remarks about other people’s substance use ONLY if the remarks also include personal beliefs
about the benefits, justifications, or appeal of those substances (e.g. “My brother’s grades improved when
he became a pothead, so it can’t be that bad for him.”)
• Do not apply if the remarks fit better with Live and Let Live Talk (e.g. “If my brother wants to smoke weed I
am not going to stop him.”)
• Can include jokes that indicate positive feelings or beliefs about substance use
• Any positive remarks or past experience (not jokingly) involving illegal drug use BESIDES alcohol, tobacco,
and marijuana should be considered high intensity and should be coded accordingly.

Examples
“We need to have a lot of booze at the party.”
[Laughing approvingly] “You got so drunk Friday night.”
“Drinking is not a big deal because I can handle it.”
“Marijuana helps me relax.”
“It’s OK to drink if you are with your friends.”
“I am really good at knowing when to stop drinking.”
“It’s OK to drink as long as it doesn't endanger your health and as long as you can control it.”
“I think we should go skiing.” [Pretends to snort cocaine] (High Intensity)
Scoring
See Part IV, Coding Scale & Definitions
%
Responses to Riskier Talk

The following codes should be applied whenever a peer responds to a youth’s risky talk, including Live and Let Live
Talk.

Positive Response to Riskier Talk
• This code should be applied to any response to risky talk, including Live and Let Live Talk, that appears
favorable, supportive, or encouraging of risky behavior
• Can consist solely of positive affect cues such as smiling or laughing
• This code should not be applied with responses to Hooking Up Talk. If a youth responds to Hooking Up Talk
with more Hooking Up Talk (or Risky Talk About Sex), then the correct code to apply is Hooking Up Talk (or
Risky Talk About Sex).

Examples
[In response to peer saying, “We have to get wasted at the party”] “Yeah of course, definitely.”
Smiling in response to peer’s comment about bumming cigarettes from friends
[In response to peer saying that his mom doesn’t approve of his drinking] “Just do it anyway what is she
gonna do?”

Scoring
See Part IV, Coding Scale & Definitions

Negative Response to Riskier Talk
• This code should be applied to any response to Riskier Talk, including Live and Let Live Talk, that indicates
disagreement, disapproval, or discouragement
105
Appendix D.
Coding Manual Excerpts (continued)
• Can consist solely of negative affect cues such as shaking one’s head no
• This code should not be applied to negative responses to Hooking Up Talk UNLESS they address risk issues
(e.g. “Hooking up when you are really drunk can get you into a bad situation.”)

Examples
[In response to statement that peer occasionally smokes marijuana] “I don’t approve of marijuana, I would
never smoke.”
[In response to peer justifying his brother’s sale of marijuana by saying that he does not make any money
from it] “But he is still selling it and he might still get caught and get in trouble.”
[In response to peer saying that he wants alcohol at the party they are planning] “‘Why? None of us really
drink so there’s no point.”
[In response to peer saying that he does not want to tell his younger brother to stop selling drugs because
he does not want his brother to stop liking him] “You have to realize that that’s what an older brother does.
You need to set an example and even if he dislikes you for it now, he will eventually realize you did him a
favor and appreciate it.”
Scoring
See Part IV, Coding Scale & Definitions

Low Risk Talk

Cautious Health Talk
• This code should be applied when the youth expresses views of risk behavior that minimize harm and
incorporate personal beliefs about values, identity, health, safety, or rights of the self or others
• May include personal limits that reduce risk exposure and are specific in nature, such as “I never have more
than one drink. It’s always with a small group of friends and we always have a designated driver.”
• May include an admission of past experimentation with alcohol, tobacco, or marijuana (not other illegal
drugs) if the experimentation was brief and the youth concluded that the substance did not conform with
beliefs about values, identity, health, safety, etc. (e.g. “I’ve tried marijuana a few times, but I don’t do it
anymore because I don’t want to become an idiot pothead”)
• The views should be well-developed and indicate self-restraint, not simply a rationalization for why risk
behavior is acceptable given certain conditions.
o For example, DO NOT apply this code to statements such as:
“It’s OK to drink if you are with your friends”
“Marijuana is OK if you need it to relax”
“I am really good at knowing when to stop drinking”
“It’s OK to drink as long as it doesn't endanger your health and as long as you can control it.”
“My goal for the next year is to not get pregnant”
o These statements are all examples of Positive Talk About Substances (see above).
• The views should not merely be expressions of disgust for a substance (see Disgust Talk, below)
• This code should be applied if the youth expresses any approval for condom or contraception use, or STI
testing
• This code should be applied if the youth indicates avoidance of risky peers, such as excluding risky guests
when planning the party
• This code should be applied if the youth suggests non-risky activities for guests when planning the party
(e.g. baking cupcakes, watching a movie) at the exclusion of risk behavior
106
Appendix D.
Coding Manual Excerpts (continued)
o If the youth NEVER suggests riskier activities such as substance use or hooking up when
planning the party, this should be considered High Intensity and coded accordingly
• If the view appears dogmatic in nature, i.e. invokes a legal or religious restriction without much integration of
personal beliefs (e.g. “Obviously we shouldn’t drink because we are younger than 21,” said in a sarcastic
tone), this should be considered a Hint and coded accordingly

Examples
“I don’t want smoke in my body.” (Hint)
“I don’t do marijuana because I’m scared that I would get paranoid.”
“I don’t wanna date guys who do drugs and drink often.” (Also code as Interpersonal Value Talk)

“I don’t like parties.” (Hint)
“I don’t drink and drive.”
“We don’t drink every week like most college students do. We go watch movies and just go to dinner.”
“We don’t want the party to be a drinkfest.”
“I don’t like drugs. There’s no proper social setting in which you can do drugs.”
“I won’t drink by myself, I won’t drink at home after work.” (Hint)
“I don’t approve of drugs or people who do drugs. I haven’t tried drugs and don’t plan to in the future.” (High
Intensity)
“When I talk to Laura she’s like why aren’t you [hooking up with people]? It’s not in my personality. It’s not
my objective, to go hook up with someone and get a boyfriend.” (Hint)
“I can’t drink because I have to drive everybody’s drunk asses home.”
Scoring
See Part IV, Coding Scale & Definitions

Disgust Talk
• This code should be applied when the youth expresses negative views of risk behavior that are largely based on
looks, taste, or other sensory information

Examples
“Cigarettes are disgusting!”
“Wine tastes weird.”
“I don’t drink beer because it’s not ladylike.”
“Alcohol is nasty.”
“The idea of hooking up with somebody I don’t really know is gross.”
Scoring
See Part IV, Coding Scale & Definitions

107
Appendix D.
Coding Manual Excerpts (continued)
Responses to Low Risk Talk
The following codes should be applied whenever a peer responds to a youth’s Low Risk or Change Talk about risk.

Positive Response to Low Risk Talk
• This code should be applied to any response to Low Risk or Change Talk that appears to agree, approve,
support, or encourage avoidance or reduction of risk
• If the remark is in response to change talk concerning non-risk behavior (e.g. going on a diet, studying more)
this code does not apply. Code instead as Emotional Support, Instrumental Support, or another non-risk
code.
• Can consist solely of positive affect cues such as smiling or nodding.

Examples
[In response to friend who said that he did not drink at all as a freshman] “Wow that’s impressive”
[In response to friend who said that he does not approve of people who use drugs] “It’s alright. You still love
me.”

Scoring
See Part IV, Coding Scale & Definitions

Negative Response to Low Risk Talk
• This code should be applied to any response to Low Risk or Change Talk that indicates disagreement,
disapproval, or discouragement of risk avoidance or reduction
• If the remark is in response to change talk concerning non-risk behavior (e.g. going on a diet, studying more) this
code does not apply.
• Can consist solely of negative affect cues such as shaking one’s head no

Examples
[In response to peer saying that she does not want alcohol at the party] “Well other people probably want to
drink.”
[In response to peer saying, “I think anything tobacco related is stupid”] “I don’t care as much... I just don’t
feel like there’s any problem.”
[In response to peer saying, “I think marijuana should only be for adults. I like the law 21 and over.”] I think
that’s retarded; I think the drinking age as well as the marijuana age should be lowered to 16 years of
age.”
Scoring
See Part IV, Coding Scale & Definitions

Interpersonal Attitudes & Future Expectations

Interpersonal Value Talk
• This code should be applied when the youth praises friends, dating partners, or other peers for personality- or
relationship-based characteristics such as behavior patterns, personal style, shared interests, shared
understanding, or similar goals
• This code should also be applied when the youth states a desire for friends or partners with such characteristics

Examples
“I like guys when they have goals, when they understand me” (Also code as Future Orientation)
“I like my friends because they are smart and funny”
“ I wouldn’t date a guy who does drugs or drinks a lot” (Also code as Personalized Low Risk)
108
Appendix D.
Coding Manual Excerpts (continued)
“ I love that me and my boyfriend can have long moments of silence and that I can be myself with him. I’m
very comfortable around him”
“We have fun together. He has my sense of humor.”
“I want my girlfriend to be successful and driven in life, and someone that puts up with me.” (Also code as
Future Orientation)
“It’s important that I have a good relationship with his family”
“It’s not just about a sexual or physical attraction, but also about trust, and it’s important that I am
comfortable with my girl and am able to be myself around her”
“It’s important to find someone you really connect with”
“I need someone who’s equally invested in the relationship as I am”
“ I need someone who’s not necessary supportive but understanding of the choices I make in my life.. and
it’s all about give and take in a relationship. You can’t just take or just give”

Scoring
See Part IV, Coding Scale & Definitions

%
%

109
Appendix E.
Coder Sheet
WA V E 5 A DYA DI C DI S CUS S I ON CODI NG (page 1 of 2)

ID# SEX: DATE: CODER:

* Highlighted codes are included in current studies
110
Appendix F.
Substance Use Items

111
Appendix F.
Substance Use Items (continued)

112

Appendix G.
Peer Norm Items

* Used for risky friends measure ‡ Used for prosocial friends measure
‡%
‡%
‡%
‡%
‡%
*%
*%
*%
*%
*%
*%
113
Appendix G.
Peer Norm Items (continued)

† Used for unprotected sex norms measure * Used for risky friends measure
‡ Used for prosocial friends measure

‡%
‡%
‡%
‡%
‡%
‡%
‡%
†%
*%
114
Appendix H.
Grades Item

115
Appendix I.
Sexual Risk Behavior Items

116
Appendix I.
Sexual Risk Behavior Items (continued)
For females:

For males:

117
Appendix J.
Quantitative Approach to Dyadic Data
The present dataset posed a challenge in that individual participants’ data were clustered
within dyads, violating the independence assumption inherent to most statistical analyses
(Kenny, Kashy, & Cook, 2006). Furthermore, the dyads themselves were theoretically
“indistinguishable” in the sense that dyad members could not be ordered systematically by a
structural variable such as gender, age, or role (Griffin & Gonzalez, 1995). Indistinguishability
poses a problem in correlational analysis; estimates are biased because the ordering of one dyad
member as first and the other as second is arbitrary. Therefore, modeling techniques were
selected to handle individual-level data that is clustered within indistinguishable dyads, as
appropriate to the analysis.
First, within dyadic CFA, a dyad-level structure was used, with each dyad member’s data
entered as separate variables. Similar to the hypothetical example in Table A1 below, the scores
for both members of a given dyad were entered on the same line (e.g., M
AD
and M
AE
). As
explained in more detail in the next section, dyadic non-independence was resolved through the
addition of special paths and constraints, whereas indistinguishability was addressed through an
adjustment to the model’s chi-squared statistic (Olsen & Kenny, 2006; Peugh, DiLillo, &
Panuzio, 2013).
Meanwhile, correlation coefficients were computed using a pairwise data structure. As
shown in the hypothetical pairwise structure in Table A2 below, the scores for a given dyad were
repeated but with the dyad members’ scores reversed. This “checkerboard pattern” results in an
unbiased Pearson’s r correlation, although the double listing of data requires recalculation of p

118
Appendix J.
Quantitative Approach to Dyadic Data (continued)

values using an adjusted sample size, which itself takes into account nonindependence using the
intra-dyad correlation (for more information, see Griffin & Gonzalez, 1995).
Table A
Dyadic Data Structures
1. Dyadic CFA structure
Dyad ID 1
st
Youth ID 2
nd
Youth ID Measure M1 Score
for 1
st
Youth
Measure M1 Score
for 2
nd
Youth
…
A D E M
AD
M
AE
…
B D E M
BD
M
BE
…
C D E M
CD
M
CE
…
… … … … … …

2. Pairwise structure

Dyad ID 1
st
Youth ID 2
nd
Youth ID Measure M1 Score
for 1
st
Youth
Measure M1 Score
for 2
nd
Youth
…
A D E M
AD
M
AE
…
A E D M
AE
M
AD
…
B D E M
BD
M
BE
…
B E D M
BE
M
BD
…
C D E M
CD
M
CE
…
C E D M
CE
M
CD
…
… … … … … …
Note. M1
BD
represents the score on measure M1 for youth D in Dyad B.

For the logistic analyses, dyadic non-independence was handled by computing robust
standard errors using sandwich estimation (Muthén & Muthén, 2012). Indistinguishability was
not at issue in these analyses because intra-dyad associations were not examined.

119
Appendix J.
Quantitative Approach to Dyadic Data (continued)

Dyadic Confirmatory Factor Analysis. A dyadic CFA extends conventional CFA
through a number of adjustments. As shown in Figure A below, two dyadic latent factors were
expected (risk promotion and health promotion), each with indicators corresponding to both
members of the dyad. Due to dyadic indistinguishability, parameters that were common to dyads
were constrained to be equal, including intercepts (a, b, c, d), factor loadings (f, g, h, i), and error
variances (j, k, l, m). In addition, dyadic nonindependence required that errors be allowed to
covary (n, o, p, q); these paths represent residual intraclass covariances, i.e., the degree of
nonindependence in the indicators after accounting for the latent factor variance. The two dyadic
latent factors have a covariance r; means were set to 0 and variances to 1.

Figure A. Dyadic confirmatory factor analysis of talk behaviors. M1
D
and M1
E
are the scores on
talk behavior measure M1 corresponding to Youth D and Youth E, respectively.
120
Appendix J.
Quantitative Approach to Dyadic Data (continued)

As with conventional CFA, models were found to fit if standardized factor loadings were
sufficiently large (absolute value greater than .50) and fit indices met acceptable thresholds. If
some indicators did not load strongly on the dyadic factors, they were maintained in the model as
separate, observable variable pairs and allowed to correlate across dyad and with the dyadic
factors.
Additional paths were estimated to account for other sources of shared variance.
Specifically, shared intrapersonal variance would be expected in dyadic factor models such as
these (Peugh et al., 2013), and therefore within-person residual covariance paths were added
across factors and constrained to equality within dyad; examples of such paths are labeled s in
Figure 1. Furthermore, some additional residual covariance would be expected due to
dependencies in the coding system itself. For example, raters would be more likely to score one
youth’s risk-positive responses if her friend were observed using positive substance talk.
Therefore, cross-person paths were estimated between the residuals of these interdependent
codes and constrained to equality, as demonstrated by paths labeled t in Figure 1.
Finally, to correct the additional misfit contributed by the arbitrary ordering of
indistinguishable dyad members, chi-squared and related statistics were adjusted using
procedures outlined in Peugh et al. (2013), which required the estimation of a saturated and a
null model. The saturated model itself functions as an empirical test of the indistinguishability
assumption (Kenny et al., 2006). In the current study, this feature allowed us to test the
possibility that dyad members could be distinguished on the basis of recruitment differences. To
121
Appendix J.
Quantitative Approach to Dyadic Data (continued)

allow for this test, the dataset was intentionally structured so that Youth 1 would be the original
study youth and Youth 2 would be the invited friend. In the model tested here, the saturated
model fit well, indicating that the dyad members were not distinguishable based on recruitment
differences.

References
Griffin, D., & Gonzalez, R. (1995). Correlational analysis of dyad-level data in the exchangeable
case. Psychological Bulletin, 118(3), 430. doi:10.1037/0033-2909.118.3.430
Kenny, D., Kashy, D., & Cook, W. (2006). Dyadic data analysis. New York: The Guilford Press.
Muthén, L. K., & Muthén, B. O. (2012). Mplus User’s Guide. Seventh Edition. Muthén &
Muthén.
Olsen, J. A., & Kenny, D. A. (2006). Structural equation modeling with interchangeable dyads.
Psychological Methods, 11(2), 127-141. doi:10.1037/1082-989X.11.2.127
Peugh, J. L., DiLillo, D., & Panuzio, J. (2013). Analyzing Mixed-Dyadic Data Using Structural
Equation Models. Structural Equation Modeling: A Multidisciplinary Journal, 20(2), 314-
337. doi:10.1080/10705511.2013.769395

122
Appendix K.
CFA Mplus Code
TITLE: CFA - NULL Model

DATA:
FILE IS "Dyadic CFA 111.csv" ;
FORMAT IS free ;

VARIABLE:

NAMES ARE
FID resp1 resp2 sexeff Age1 Age2 nepov1 nepov2
whiteff1 whiteff2 Hispeff1 Hispeff2 Afeff1 Afeff2
Lpossub1 Lpossub2 Lriskre1 Lriskre2
cautiou1 cautiou2 hlthpro1 hlthpro2
interpe1 interpe2 Ldisgus1 Ldisgus2
;

USEVARIABLES ARE
FID
Lpossub1 Lpossub2 ! Lpossub - Positive substance talk (log transformed)
Lriskre1 Lriskre2 ! Lriskre - Risk-positive responses (log transformed)
cautiou1 cautiou2 ! cautiou - Cautious health talk
hlthpro1 hlthpro2 ! hlthpro - Health-promoting responses
interpe1 interpe2 ! interpe - Interpersonal value talk
Ldisgus1 Ldisgus2 ! Ldisgus - Disgust talk (log transformed)
;

IDVARIABLE IS FID ; ! Family ID

MISSING ARE all (-999) ;

MODEL:

! Equality of Intercepts

[Lpossub1 Lpossub2] (100) ;
[Lriskre1 Lriskre2] (101) ;
[cautiou1 cautiou2] (102) ;
[hlthpro1 hlthpro2] (103) ;
[interpe1 interpe2] (104) ;
[Ldisgus1 Ldisgus2] (105) ;

123
Appendix K.
CFA Mplus Code (continued)
! Equality of Variances

Lpossub1 Lpossub2 (200) ;
Lriskre1 Lriskre2 (201) ;
cautiou1 cautiou2 (202) ;
hlthpro1 hlthpro2 (203) ;
interpe1 interpe2 (204) ;
Ldisgus1 Ldisgus2 (205) ;

! All Possible Correlations/Covariances Fixed to Zero

Lpossub1 WITH Lpossub2@0
Lriskre1@0 Lriskre2@0
cautiou1@0 cautiou2@0
hlthpro1@0 hlthpro2@0
interpe1@0 interpe2@0
Ldisgus1@0 Ldisgus2@0
;

Lpossub2 WITH Lriskre1@0 Lriskre2@0
cautiou1@0 cautiou2@0
hlthpro1@0 hlthpro2@0
interpe1@0 interpe2@0
Ldisgus1@0 Ldisgus2@0
;

Lriskre1 WITH Lriskre2@0
cautiou1@0 cautiou2@0
hlthpro1@0 hlthpro2@0
interpe1@0 interpe2@0
Ldisgus1@0 Ldisgus2@0
;

Lriskre2 WITH cautiou1@0 cautiou2@0
hlthpro1@0 hlthpro2@0
interpe1@0 interpe2@0
Ldisgus1@0 Ldisgus2@0
;

124
Appendix K.
CFA Mplus Code (continued)
cautiou1 WITH cautiou2@0
hlthpro1@0 hlthpro2@0
interpe1@0 interpe2@0
Ldisgus1@0 Ldisgus2@0
;

cautiou2 WITH hlthpro1@0 hlthpro2@0
interpe1@0 interpe2@0
Ldisgus1@0 Ldisgus2@0
;

hlthpro1 WITH hlthpro2@0
interpe1@0 interpe2@0
Ldisgus1@0 Ldisgus2@0
;

hlthpro2 WITH interpe1@0 interpe2@0
Ldisgus1@0 Ldisgus2@0
;

interpe1 WITH interpe2@0
Ldisgus1@0 Ldisgus2@0
;

interpe2 WITH Ldisgus1@0 Ldisgus2@0
;

Ldisgus1 WITH Ldisgus2@0
;

OUTPUT:
standardized sampstat ;

125
Appendix K.
CFA Mplus Code (continued)
TITLE: CFA - SATURATED Model

DATA:
FILE IS "Dyadic CFA 111.csv" ;
FORMAT IS free ;

VARIABLE:

NAMES ARE
FID resp1 resp2 sexeff Age1 Age2 nepov1 nepov2
whiteff1 whiteff2 Hispeff1 Hispeff2 Afeff1 Afeff2
Lpossub1 Lpossub2 Lriskre1 Lriskre2
cautiou1 cautiou2 hlthpro1 hlthpro2
interpe1 interpe2 Ldisgus1 Ldisgus2
;

USEVARIABLES ARE
FID
Lpossub1 Lpossub2 ! Lpossub - Positive substance talk (log transformed)
Lriskre1 Lriskre2 ! Lriskre - Risk-positive responses (log transformed)
cautiou1 cautiou2 ! cautiou - Cautious health talk
hlthpro1 hlthpro2 ! hlthpro - Health-promoting responses
interpe1 interpe2 ! interpe - Interpersonal value talk
Ldisgus1 Ldisgus2 ! Ldisgus - Disgust talk (log transformed)
;

IDVARIABLE IS FID ; ! Family ID

MISSING ARE all (-999) ;
MODEL:

! Equality of Intercepts

[Lpossub1 Lpossub2] (100) ;
[Lriskre1 Lriskre2] (101) ;
[cautiou1 cautiou2] (102) ;
[hlthpro1 hlthpro2] (103) ;
[interpe1 interpe2] (104) ;
[Ldisgus1 Ldisgus2] (105) ;

126
Appendix K.
CFA Mplus Code (continued)
! Equality of Variances

Lpossub1 Lpossub2 (200) ;
Lriskre1 Lriskre2 (201) ;
cautiou1 cautiou2 (202) ;
hlthpro1 hlthpro2 (203) ;
interpe1 interpe2 (204) ;
Ldisgus1 Ldisgus2 (205) ;

! Intra-personal Dyadic Dependence

Lpossub1 WITH Lriskre1 (1);
Lpossub2 WITH Lriskre2 (1);
Lpossub1 WITH cautiou1 (2);
Lpossub2 WITH cautiou2 (2);
Lpossub1 WITH hlthpro1 (3);
Lpossub2 WITH hlthpro2 (3);
Lpossub1 WITH interpe1 (4);
Lpossub2 WITH interpe2 (4);
Lpossub1 WITH Ldisgus1 (5);
Lpossub2 WITH Ldisgus2 (5);

Lriskre1 WITH cautiou1 (6);
Lriskre2 WITH cautiou2 (6);
Lriskre1 WITH hlthpro1 (7);
Lriskre2 WITH hlthpro2 (7);
Lriskre1 WITH interpe1 (8);
Lriskre2 WITH interpe2 (8);
Lriskre1 WITH Ldisgus1 (9);
Lriskre2 WITH Ldisgus2 (9);

cautiou1 WITH hlthpro1 (10);
cautiou2 WITH hlthpro2 (10);
cautiou1 WITH interpe1 (11);
cautiou2 WITH interpe2 (11);
cautiou1 WITH Ldisgus1 (12);
cautiou2 WITH Ldisgus2 (12);

hlthpro1 WITH interpe1 (13);
hlthpro2 WITH interpe2 (13);
hlthpro1 WITH Ldisgus1 (14);
hlthpro2 WITH Ldisgus2 (14);
127

Appendix K.
CFA Mplus Code (continued)
interpe1 WITH Ldisgus1 (15);
interpe2 WITH Ldisgus2 (15);

! Inter-personal Dyadic Dependence;

Lpossub1 WITH Lriskre2 (31);
Lpossub2 WITH Lriskre1 (31);
Lpossub1 WITH cautiou2 (32);
Lpossub2 WITH cautiou1 (32);
Lpossub1 WITH hlthpro2 (33);
Lpossub2 WITH hlthpro1 (33);
Lpossub1 WITH interpe2 (34);
Lpossub2 WITH interpe1 (34);
Lpossub1 WITH Ldisgus2 (35);
Lpossub2 WITH Ldisgus1 (35);

Lriskre1 WITH cautiou2 (36);
Lriskre2 WITH cautiou1 (36);
Lriskre1 WITH hlthpro2 (37);
Lriskre2 WITH hlthpro1 (37);
Lriskre1 WITH interpe2 (38);
Lriskre2 WITH interpe1 (38);
Lriskre1 WITH Ldisgus2 (39);
Lriskre2 WITH Ldisgus1 (39);

cautiou1 WITH hlthpro2 (40);
cautiou2 WITH hlthpro1 (40);
cautiou1 WITH interpe2 (41);
cautiou2 WITH interpe1 (41);
cautiou1 WITH Ldisgus2 (42);
cautiou2 WITH Ldisgus1 (42);

hlthpro1 WITH interpe2 (43);
hlthpro2 WITH interpe1 (43);
hlthpro1 WITH Ldisgus2 (44);
hlthpro2 WITH Ldisgus1 (44);

interpe1 WITH Ldisgus2 (45);
interpe2 WITH Ldisgus1 (45);

128
Appendix K.
CFA Mplus Code (continued)
! Allow Correlations/Covariances to Be Freely Estimated

Lpossub1 WITH Lpossub2 ;
Lriskre1 WITH Lriskre2 ;
cautiou1 WITH cautiou2 ;
hlthpro1 WITH hlthpro2 ;
interpe1 WITH interpe2 ;
Ldisgus1 WITH Ldisgus2 ;

OUTPUT:
standardized res ;

129
Appendix K.
CFA Mplus Code (continued)
TITLE: CFA Model 2
DATA:
FILE IS "Dyadic CFA 111.csv" ;
FORMAT IS free ;

VARIABLE:

NAMES ARE
FID resp1 resp2 sexeff Age1 Age2 nepov1 nepov2
whiteff1 whiteff2 Hispeff1 Hispeff2 Afeff1 Afeff2
Lpossub1 Lpossub2 Lriskre1 Lriskre2
cautiou1 cautiou2 hlthpro1 hlthpro2
interpe1 interpe2 Ldisgus1 Ldisgus2
;

USEVARIABLES ARE
FID
Lpossub1 Lpossub2 ! Lpossub - Positive substance talk (log transformed)
Lriskre1 Lriskre2 ! Lriskre - Risk-positive responses (log transformed)
cautiou1 cautiou2 ! cautiou - Cautious health talk
hlthpro1 hlthpro2 ! hlthpro - Health-promoting responses
interpe1 interpe2 ! interpe - Interpersonal value talk
Ldisgus1 Ldisgus2 ! Ldisgus - Disgust talk (log transformed)
;

IDVARIABLE IS FID ; ! Family ID

MISSING ARE all (-999) ;

MODEL:

DRP BY ! Dyadic Risk Promotion
Lpossub1* (1)
Lriskre1 (2)
Lpossub2 (1)
Lriskre2 (2) ;
[DRP@0] ;
DRP@1 ;

130
Appendix K.
CFA Mplus Code (continued)
DHP BY ! DHP - Dyadic Health Promotion
cautiou1* (3)
hlthpro1 (4)
interpe1 (5)
cautiou2 (3)
hlthpro2 (4)
interpe2 (5) ;
[DHP@0] ;
DHP@1 ;

DRP WITH DHP ;

DRP WITH Ldisgus1 (10) ;
DRP WITH Ldisgus2 (10) ;

DHP WITH Ldisgus1 (11) ;
DHP WITH Ldisgus2 (11) ;

! Correlated errors, i.e., residual intraclass covariances

Lriskre1 WITH Lriskre2 ;
cautiou1 WITH cautiou2 ;
hlthpro1 WITH hlthpro2 ;
interpe1 WITH interpe2 ;
Ldisgus1 WITH Ldisgus2 ;
Lpossub1 WITH Lpossub2 @0;

! The last of these covariances was set to 0 because of Psi Not Positive Definite error

! Shared intrapersonal variance
Lpossub1 WITH cautiou1 (50) ;
Lpossub2 WITH cautiou2 (50) ;

! Residual covariance due to coding system dependencies

Lriskre1 WITH Lpossub2 (100) ;
Lriskre2 WITH Lpossub1 (100) ;
hlthpro1 WITH Lpossub2 (101) ;
hlthpro2 WITH Lpossub1 (101) ;
hlthpro1 WITH cautiou2 (102) ;
hlthpro2 WITH cautiou1 (102) ;

131
Appendix K.
CFA Mplus Code (continued)
! Equality of Intercepts

[Lpossub1 Lpossub2] (200) ;
[Lriskre1 Lriskre2] (201) ;
[cautiou1 cautiou2] (202) ;
[Ldisgus1 Ldisgus2] (203) ;
[hlthpro1 hlthpro2] (204) ;
[interpe1 interpe2] (205) ;

! Equality of Residuals

Lpossub1 Lpossub2 (300) ;
Lriskre1 Lriskre2 (301) ;
cautiou1 cautiou2 (302) ;
Ldisgus1 Ldisgus2 (303) ;
hlthpro1 hlthpro2 (304) ;
interpe1 interpe2 (305) ;

OUTPUT:
standardized sampstat ;

%

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Emotion regulation as a mechanism linking parents’ marital aggression to adolescent behavioral problems: a longitudinal analysis

PDF

Using observed peer discussions to understand adolescent depressive symptoms and interpersonal interactions

PDF

Parental military service and adolescent mental and behavioral health: the role of adolescent-civilian parent interactions

PDF

Actual and perceived social reinforcements of weight-related cognitions and behaviors in adolescent peer groups

PDF

Social status, perceived social reputations, and perceived dyadic relationships in early adolescence

PDF

Risky behaviors, interpersonal conflict, and their relation to fluctuations in adolescents’ diurnal HPA rhythms

PDF

Family aggression, prosocial friends, and the risk of dating and friend victimization in late adolescence

PDF

Impact of change in sexual identity on mental health risks among sexual minority adolescents

PDF

Examining the longitudinal relationships between community violence exposure and aggressive behavior among a sample of maltreated and non-maltreated adolescents

PDF

Perspective taking behavior and social outcomes in late adolescence

PDF

The link between maternal depression and adolescent daughters' risk behavior: the mediating and moderating role of family

PDF

Sex talks: an examination of young Black women's social networks, sexual health communication, and HIV prevention behaviors

PDF

Anxiety symptoms and nicotine use among adolescents and young adults

PDF

Neighborhood context and adolescent mental health: development and mechanisms

PDF

The role of social support in the relationship between adverse childhood experiences and addictive behaviors across adolescence and young adulthood

PDF

A multiple systems approach to examining physiological stress and its association with internalizing disorders in adolescence

PDF

Examining the associations of respiratory problems with psychological functioning and the moderating role of engagement in pleasurable activities during late adolescence

PDF

Family aggression exposure and community violence exposure associated with brain volume in late adolescence: a comparison of automated versus manual segmentation

PDF

Adolescent conduct problems and substance use: an examination of the risk pathway across the transition to high school

PDF

A network analysis of online and offline social influence processes in relation to adolescent smoking and alcohol use

Asset Metadata

Creator Iturralde, Estibaliz (author)

Core Title Not just talk: observed communication in adolescent friendship and its implications for health risk behavior

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Psychology

Publication Date 04/21/2015

Defense Date 03/09/2015

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

Tag adolescent risk behavior,adolescent sexual health,behavioral observation,deviancy training,dyadic analysis,health communication,intimacy,OAI-PMH Harvest,peer relationships,positive youth development,social learning,social support

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Margolin, Gayla (committee chair), Cederbaum, Julie A. (committee member), Lopez, Steven R. (committee member), Manis, Franklin R. (committee member), Narayanan, Shrikanth S. (committee member)

Creator Email esti.m.iturralde@gmail.com,iturrald@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-555959

Unique identifier UC11300037

Identifier etd-IturraldeE-3354.pdf (filename),usctheses-c3-555959 (legacy record id)

Legacy Identifier etd-IturraldeE-3354.pdf

Dmrecord 555959

Document Type Dissertation

Format application/pdf (imt)

Rights Iturralde, Estibaliz

Type texts

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

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA