Week 9 Discussion

Week9ArticlebyEpstein.pdf

Home >Psychology homework help >Week 9 Discussion

The Effect of General and Drug-Specific Family Environments on Comorbid and Drug-Specific Problem Behavior:

A Longitudinal Examination

Marina Epstein, Karl G. Hill, Jennifer A. Bailey, and J. David Hawkins University of Washington

Previous research has shown that the development of alcohol and tobacco dependence is linked and that both are influenced by environmental and intrapersonal factors, many of which likely interact over the life course. The present study examines the effects of general and alcohol- and tobacco-specific environmental influences in the family of origin (ages 10 –18) and family of cohabitation (ages 27–30) on problem behavior and alcohol- and tobacco-specific outcomes at age 33. General environmental factors include family management, conflict, bonding, and involvement. Alcohol environment includes parental alcohol use, parents’ attitudes toward alcohol, and children’s involvement in family drinking. Tobacco-specific environment is assessed analogously. Additionally, analyses include the effects of childhood behavioral disinhibition, initial behavior problems, and age 18 substance use. Analyses were based on 469 participants drawn from the Seattle Social Development Project (SSDP) sample. Results indicated that (a) environmental factors within the family of origin and the family of cohabitation are both important predictors of problem behavior at age 33; (b) family of cohabitation influences partially mediate the effects of family of origin environments; (c) considerable continuity exists between adolescent and adult general and tobacco (but not alcohol) environments; age 18 alcohol and tobacco use partially mediates these relationships; and (d) childhood behavioral disinhibition contributed to age 33 outcomes, over and above the effects of family of cohabitation mediators. Implications for preventive interventions are discussed.

Keywords: family environment, behavioral disinhibition, romantic partner, adolescent alcohol and tobacco use, comorbid problem behavior

Supplemental materials: http://dx.doi.org/10.1037/a0029309.supp

Along with other risk-taking behaviors, alcohol and tobacco use increases and peaks during adolescence and young adulthood, with 50% of all young adults reporting binge drinking in the past month and over two thirds reporting lifetime smoking (Bachman et al., 2002; Johnston, O’Malley, Bachman, & Schulenberg, 2011; Sub-

stance Abuse and Mental Health Services Administration [SAM- HSA], 2010). The majority of adolescents reduce the frequency of their alcohol use, and many quit smoking by their mid-20s when they begin to take on adult roles (Chassin, Pitts, & Prost, 2002; Maggs & Schulenberg, 2004). Consequently, by their 30s, only 40% of Americans report past-year tobacco use, and one third report past-month binge drinking (SAMHSA, 2010). However, the group of young adults who are chronic or persistent users are of significance in addiction research because this group may have already developed or are at risk for developing abuse and depen- dence disorders (Chassin, Pitts, & Prost, 2002; Schulenberg, O’Malley, Bachman, Wadsworth, & Johnston, 1996).

Substance abuse and dependence are generally believed to be influenced by a combination of environmental and individual risk factors (Kreek, Nielsen, Butelman, & LaForge, 2005; Rutter, Mof- fitt, & Caspi, 2006). The same risk factors have also been impli- cated in other problem behaviors that frequently co-occur with alcohol and tobacco use, such as illicit drug use, risky sex, and criminal activity (Jackson, Sher, & Schulenberg, 2005; McGee & Newcomb, 1992; Young, Rhee, Stallings, Corley, & Hewitt, 2006). Among these risk and protective factors, the effects of family experiences have been particularly well documented (Hawkins, Catalano, & Miller, 1992; Hill, Hawkins, Catalano, Abbott, & Guo, 2005; Hops, Tildesley, Lichtenstein, Ary, &

This article was published Online First July 16, 2012. Marina Epstein, Karl G. Hill, Jennifer A. Bailey, and J. David Hawkins,

Social Development Research Group, School of Social Work, University of Washington.

Funding for this study was provided by National Institute on Drug Abuse Grants R01DA009679 and R01DA024411, National Institute on Alcohol Abuse and Alcoholism Grant R01AA016960, and by Robert Wood John- son Foundation Grant 21548. The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies. The authors gratefully acknowledge Seattle Social Development Project panel participants for their continued contri- bution to the longitudinal study. We also acknowledge the Social Devel- opment Research Group (SDRG) Survey Research Division for their hard work maintaining high panel retention and the SDRG editorial and admin- istrative staff for their editorial and project support.

Correspondence concerning this article should be addressed to Marina Epstein, Social Development Research Group, University of Washington, 9725 3rd Avenue, NE, Suite 401, Seattle, WA 98115. E-mail: marinaep@ uw.edu

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

1151

http://dx.doi.org/10.1037/a0029309.supp

Sherman, 1990). Studies have also found that, as adolescents leave parental homes, families created with romantic partners and spouses (referred to here as family of cohabitation) become in- creasingly influential. The quality of partnered relationships has been linked to problem behavior, and studies have shown a con- cordance between cohabitating partners’ level of substance use (for review, see Rhule-Louie & McMahon, 2007). In the present study, we examine the effects of environmental influences in the family of origin and family of cohabitation on the development of alcohol- and tobacco-related problems and other comorbid behav- iors such as illicit drug use, sexual risk behavior, and crime.

Predictors of Problem Behavior: Family Environments

Within the family domain, general family functioning and alcohol- and tobacco-specific influences have been identified as important predictors of problem behavior. Moffitt has argued that the strongest predictors of adult deviance can be traced to early childhood (Moffitt, 1993a, 2003), and studies have found that early risk factors in the family, such as parental substance use, low parental monitoring, and family conflict, predict later substance abuse, high-risk sexual behavior, and involvement in crime (e.g., Chassin, Presson, Rose, Sherman, & Prost, 2002; Engels, Ver- mulst, Dubas, Bot, & Gerris, 2005; Moffitt & Caspi, 2001). On the other hand, protective factors such as consistent family manage- ment and bonding can act as buffers against risk exposure and are associated with more positive outcomes (Galaif, Stein, Newcomb, & Bernstein, 2001; Guo, Hawkins, Hill, & Abbott, 2001; Ryan, Jorm, & Lubman, 2010).

Family influences remain important contributors to problem behavior throughout development, although the family composi- tion changes as children move away from parental homes and establish their own families. Relationship quality with an intimate partner, such as attachment, involvement, and support, consistently play a protective role against problem behavior (Laub, Nagin, & Sampson, 1998; Simons, Stewart, Gordon, Conger, & Elder, 2002). At the same time, studies routinely find partner intercorre- lations of .30 –.40 for alcohol use and smoking (Rhule-Louie & McMahon, 2007). For example, Kuo et al. (2007) found consid- erable spousal concordance of lifetime smoking (rs � .19 –.48) in three generations of Australian adults.

Some researchers have made distinctions between general en- vironmental factors that predict problem behavior in general and those risks that are unique to a specific drug (e.g., Andrews, Hops, & Duncan, 1997; Hill et al., 2005). In this work, on the one hand, we define general family environment as overall family function- ing that is not related to substance use, such as parental monitor- ing, family conflict, and parental warmth. On the other hand, alcohol family environment or tobacco family environment each refer to influences within the family that are specifically associated with alcohol or tobacco, including parental use of alcohol or tobacco, attitudes regarding each substance, and access to those substances in the home. A large body of literature has shown that positive general family environment plays a protective role in children’s lives, including lowering the risk of aggression and delinquency (e.g., Loeber & Dishion, 1983; Newcomb & Loeb, 1999). However, tobacco (Bricker et al., 2006; Engels, Knibbe, de Vries, Drop, & van Breukelen, 1999) and alcohol (Johnson & Leff, 1999; Merline, Jager, & Schulenberg, 2008) environments have

each been shown to be significant risk factors for later tobacco and alcohol dependence, respectively. Bailey and colleagues (Bailey, Hill, Meacham, Young, & Hawkins, 2011) found that general family environment during adolescence was uniquely associated with comorbid problem behavior in young adulthood but that drug-specific family factors such as parent smoking and drinking were uniquely linked to problematic use of tobacco and alcohol, respectively, and did not predict problem behavior in general.

Developmental Continuity in Family Environment: The Social Development Model

Life course models in the development of addiction suggest that early family experiences can have a large impact on future behav- ior, including intergenerational continuity in drug use and other antisocial actions. The theory guiding the present study, the Social Development Model (SDM; Catalano & Hawkins, 1996; Hawkins & Weis, 1985), explains such continuity in terms of opportunities for involvement, rewards, skills, bonding, and beliefs fostered within the family that set children on either a prosocial or an antisocial path. The SDM has successfully predicted tobacco and alcohol use among adolescents and emerging adults (e.g., Fleming, Kim, Harachi, & Catalano, 2002; Hill et al., 2005). The SDM also incorporates developmental submodels that specify the different socialization forces and different positive and negative outcomes salient for each developmental stage (Catalano & Hawkins, 1996).

As individuals transition into adulthood and marry or partner, families of origin are joined—and, for many, replaced— by fam- ilies of cohabitation (Bachman et al., 2002; Schulenberg, Bryant, & O’Malley, 2004). As children move toward establishing their own families, they are hypothesized to use the skills and practices they learn in the family of origin in their own families. Romantic partners thus become important sources of influence as the risk and protective factors previously associated with family of origin are then transferred to corresponding risk (e.g., conflict) and protective (e.g., involvement, bonding) factors in the adult family (Catalano & Hawkins, 1996). Studies examining intergenerational continuity of pro- and antisocial behavior have found that a positive envi- ronment in the family of origin is carried on both through the choice of partner and the subsequent partnered family environment (Fischer, Fitzpatrick, & Cleveland, 2007; Harter, 2000; Leveridge, Stoltenberg, & Beesley, 2005). For example, Donnellan, Larsen- Rife, and Conger (2005) found that youth who experienced posi- tive interactions in the family of origin had more positive and stable romantic relationships later in life.

A number of studies have also examined the apparent continuity in alcohol and tobacco environment that is evident when children of substance abusers partner with substance-abusing others (for review, see Harter, 2000; Johnson & Leff, 1999). This link may be mediated by children’s own substance abuse prior to partnering (e.g., Bailey, Hill, Oesterle, & Hawkins, 2006; Latendresse et al., 2008). The extensive research on children of alcoholics shows that having alcoholic parents is a risk factor for choosing to marry a substance abuser (e.g., Olmsted, Crowell, & Waters, 2003). There is less direct evidence that children of smokers choose smoking partners, yet children of smokers have been shown to associate with smoking peers (e.g., Engels, Vitaro, Den Exter Blokland, de Kemp, & Scholte, 2004) who are likely to make up the social pool from which one’s romantic partner is drawn. Furthermore, the

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

1152 EPSTEIN, HILL, BAILEY, AND HAWKINS

concordance between parent and child smoking (Engels et al., 2004; Taylor, Conard, O’Byrne, Haddock, & Poston, 2004) and high spousal smoking concordance (Rhule-Louie & McMahon, 2007) both suggest that such continuity exists.

Individual Vulnerability

Another body of literature has documented the predictive role that individual vulnerabilities play in the development of drug use and other problem behaviors. In particular, a cluster of highly heritable personality traits characterized by sensation seeking, risk taking, and other externalizing behaviors, referred to as behavioral disinhibition (BD; Iacono, Carlson, Taylor, Elkins, & McGue, 1999; Iacono, Malone, & McGue, 2008), has been shown to predict initiating and escalating substance use in adolescence (Brook, Ning, & Brook, 2006; Hill, White, Chung, Hawkins, & Catalano, 2000; Neighbors, Kempton, & Forehand, 1992) and substance abuse and dependence in adulthood (Hu, Davies, & Kandel, 2006; Jackson & Sher, 2005; Tucker, Ellickson, & Klein, 2003).

In addition to being a predictor of alcohol and tobacco use, BD has been linked with other problem behaviors, making it a general rather than a substance-specific vulnerability. Behavioral geneti- cists have found evidence supporting a genetic liability that is common to both BD and substance abuse (e.g., Button et al., 2007; Iacono et al., 1999, 2008), suggesting that BD may be an indicator, or endophenotypic marker, of vulnerability to antisocial behavior. This common genetic liability also has been hypothesized to explain the high degree of comorbidity between substance use and other problem behavior, such as involvement in crime and sexual risk taking (McGee & Newcomb, 1992; McGue, Iacono, & Krueger, 2006; Young et al., 2006). Because of this comorbidity, it is difficult to separate predictors of general externalizing behav- ior from factors that predict substance-specific addiction (Conway, Compton, & Miller, 2006), making it difficult to establish, for example, whether a particular gene is associated with involvement in many types of problem behavior or with only drug-specific behavior.

The Present Study

Although extensive research has focused on environmental risks during adolescence and adulthood, less is known about the relation between adolescent and adult family environments in predicting problem behavior in adulthood. Also, little is known about the ways that family environments interact with individual vulnerabil- ities, such as BD. The goal of this study was to build a model of adult comorbid problem behaviors and noncomorbid alcohol and tobacco problems that identifies the effects of family environmen- tal and individual characteristics from adolescence to adulthood. We consider family environments as a sequence of shifting con- texts from family of origin in adolescence to family of cohabitation in young adulthood, and distinguish general and alcohol- and tobacco-specific family factors as predictors. We also distinguish predictors of comorbid problem behavior from predictors of to- bacco and alcohol problems that occur without involvement in other forms of problem behavior. The study is guided by four hypotheses (see online Supplemental Materials, Appendix 1):

Hypothesis 1: General, alcohol-specific, and tobacco-specific environmental factors in the family of origin predict age 33 comorbid problem behavior, alcohol abuse and dependence, and tobacco dependence, respectively

Bailey et al. (2011) found that general adolescent family envi- ronment predicted age 24 comorbid problem behavior, whereas adolescent family tobacco-specific and alcohol-specific environ- ments predicted age 24 alcohol and tobacco use, respectively. We hypothesized that these relationships persist through age 33. In extending the work of Bailey et al., we believe it is important to examine a range of distal outcomes related to social environments to better understand the potentially long-lasting influence that early experiences may have on later problem behavior. Further- more, we sought to extend the model proposed in the Bailey et al. study to a later age when alcohol and tobacco misuse and other problem behavior are no longer part of a normative trend (Schu- lenberg & Maggs, 2002).

Hypothesis 2: BD assessed during adolescence predicts co- morbid problem behavior at age 33.

We hypothesized that BD predicts comorbid problem behavior at age 33 but not alcohol- or tobacco-specific outcomes. In this prediction, we relied on previous research by behavioral geneti- cists that has demonstrated that a heritable latent vulnerability toward general problem behavior is manifested through BD (e.g., Button et al., 2007; Iacono et al., 2008). We also hypothesized that BD moderates the protective effect of adolescent family environ- ment on comorbid problem behavior at age 33 (Hill et al., 2010). We included a baseline measure of behavior problems (delin- quency at age 10) that we expect to be highly related to BD because of their underlying common cause. Finally, we hypothe- sized that early delinquent acts, such as stealing and fighting, might be associated with both comorbid behavior problems and criminal behavior in adulthood at age 33.

Hypothesis 3: General and alcohol- and tobacco-specific en- vironments in the family of origin (ages 10 –18) predict general and alcohol- and tobacco-specific environments in the adult family of cohabitation (ages 27–30).

Consistent with the life course view of the SDM (Catalano & Hawkins, 1996), we expected to find continuity of environmental influences, such that the general family environment and alcohol- and tobacco-specific family environmental factors in the family of origin are positively associated with their respective family envi- ronment counterparts in the family of cohabitation. We hypothe- sized that skills such as conflict management, which are modeled and learned in the family of origin, are likely to be applied in one’s relationship with a romantic partner. However, early exposure to alcohol and tobacco use may predispose participants toward choice of an intimate partner who engages in drinking or smoking behav- ior. We tested whether participants’ alcohol and tobacco use at age 18 mediated these pathways.

Hypothesis 4: General and alcohol- and tobacco-specific en- vironments in the family of cohabitation partially mediate the relation between family of origin environments and adult problem behaviors.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

1153GENERAL AND DRUG-SPECIFIC ENVIRONMENTS

Following research suggesting lasting effects of both childhood and adult family influences on problem behavior, we hypothesized that adolescent social influences will emerge as distinct predictors from adult factors in predicting age 33 outcomes. We also ex- pected that these influences will persist over and above the asso- ciation between adolescent substance use and adult substance use problems. Specifically, we hypothesized environmental factors in the family of cohabitation to partially mediate the effect of early family influences, such that family of origin environments would have both direct and indirect effects on age 33 problem behavior. We did not expect to see any change in the effect of BD on outcomes with the addition of family of cohabitation environmen- tal factors in analyses, because BD has been found to develop early and remain a life course-consistent trait (Cloninger, Sigvardsson, & Bohman, 1996; Iacono et al., 1999; Moffitt, 1993b).

Method

Participants and Procedure

Data for this study were drawn from the Seattle Social Devel- opment Project, a longitudinal study of 808 youth (412 male) recruited in 1985 from elementary schools serving a mixture of neighborhoods including neighborhoods with high rates of crime (Hawkins, Kosterman, Catalano, Hill, & Abbott, 2005). Almost half of the original sample (46%) came from families with a family income under $20,000 per year, and 52% participated in the National School Lunch/School Breakfast program during at least 1 year between fifth and seventh grade. Face-to-face interviews were conducted with participants at ages 10, 11, 12, 13, 14, 15, 16, and 18, and questionnaire data from parents were also collected annu- ally at ages 10 through 16. Follow-up interviews were then ad- ministered to participants at ages 21, 24, 27, 30, and 33. From age 11 to 33, annual retention rates averaged 90%, with 92% of the still-living sample having been interviewed at age 33 (deceased n � 23 by age 33). At age 33, 90% of participants participated in face-to-face interviews, 7% completed web surveys, 2% submitted paper surveys, and 1% completed interviews by telephone.

Because a main focus of these analyses was the influence of the family environment (family of cohabitation) in adulthood, we chose to examine family environment at ages 27 and 30, a time when the majority of participants had formed families with live-in spouses or romantic partners. Two time points, ages 27 and 30, were selected to maximize the number of cohabitating participants. Accordingly, participants who did not report a spouse or live-in romantic partner at either age 27 or 30 (n � 311) were excluded from the analyses. Due to their low representation, Native Amer- icans (n � 28) were excluded from these analyses, bringing the final analysis sample to 469 participants. Of these, 237 (51%) were female, 110 (23%) self-identified as African American, 106 (23%) as Asian American, and the majority reported being married during at least one time point at ages 27–30 (n � 322, 69%).

Throughout the analyses, items in the adolescent subscales were combined and averaged across ages 10 –18. Measures of family of cohabitation were averaged over ages 27 and 30. Composites were created for cases in which at least half of the data points across the waves were present. Items with different response scales were standardized prior to combining. See Appendix 3 for detailed information about the measures.

Measures

Family of origin general family environment (ages 10 –18). Family of origin general environment measures included youth report of family management, family conflict, family involvement, and bonding to family members. For all scales, items were recoded as necessary so that higher scores indicate more of the construct (e.g., more bonding, more conflict). Measures were all highly reliable across adolescence: family management average reliability from age 10 to age 18 � � .83, conflict � � .82, positive involvement � � .78, and bonding � � .81. Composite measures were used as indicators of a latent General Family Environment construct (see Supplemental Materials, Appendix 2 for loading coefficients for all latent factors).

Family of origin family alcohol environment (ages 10 –16). Family alcohol environment measures included parent drinking, parent drinking attitudes, and involvement of participants in family drinking (e.g., getting or opening a drink for a family member), all completed by parents. Adolescent parent drinking (reliability across adolescence � � .89), parent prodrinking attitudes (reliabil- ity across adolescence � � .82), and involvement in family drink- ing (reliability across adolescence � � .81) measures were used as indicators of a latent alcohol family environment construct.

Family of origin family tobacco environment (ages 10 –16). Family of origin smoking environment measures included parent’s report of parent smoking, parent smoking attitudes, and youth involvement in parent smoking (e.g., getting or lighting cigarettes for family members). Preliminary testing indicated a high degree of overlap in parental smoking and drinking attitudes. Accord- ingly, in the models described below, the residual covariances of these two variables were estimated. Adolescent parent smoking (reliability across adolescence � � .94), parent prosmoking atti- tudes (reliability across adolescence � � .80), and involvement (reliability across adolescence � � .61) in family member smoking measures were used as indicators of a family of origin tobacco family environment latent construct.

Delinquent behavior (ages 10 –11). Baseline behavior prob- lems were assessed during the fall and spring of fifth grade when most participants were 10 and 11, respectively. Participants re- ported whether they had ever engaged in any of eight delinquent behaviors, including hitting a teacher, damaging property, picking fights, and being arrested. Items were assessed either as 1 (Yes) or 2 (No) or on a 4-point scale ranging from 1 (Never) to 4 (More than 4 times). Items were recoded such that engaging in any of the behaviors at least once at either time point was recoded as 1 and not engaging coded as 0. Items were summed up for a total Delinquent Behavior score (� � .75).

BD (ages 14 –18). Behavioral disinhibition was measured at ages 14, 15, 16, and 18 by five items that assessed the frequency of risky or impulsive behavior, such as engaging in risk taking on a dare and disregarding consequences. Items were assessed on a 5-point scale anchored at 1 (never) and 5 (2–3 times a month). Items were summed and then averaged across waves creating a single summative score of BD (reliability across adolescence � � .82).

Alcohol and tobacco use (age 18). Past-month alcohol use (beer, wine, wine coolers, whiskey, gin, or other liquor) was assessed with a single item. Responses were capped at 30. Past- month cigarette use was assessed on a 5-point scale anchored at 1

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

1154 EPSTEIN, HILL, BAILEY, AND HAWKINS

(not at all) and 5 (about a pack a day or more). Responses were recoded to reflect the number of cigarettes per pack (e.g., about half a pack a day was recoded to 10, and about a pack a day or more was recoded to 30).

Family of cohabitation general family environment (ages 27–30). Assessments of family of cohabitation general family environment were based on interactions with a spouse or live-in romantic partner. Family of cohabitation general family environ- ment measures included participant report of conflict with partner, involvement with partner, and partner bonding. Items within sub- scales were combined to parallel those in the family of origin general environment. Measures of family of cohabitation conflict (� � .83), involvement (� � .77), and bonding (� � .78) were each used as an indicator of a latent general family environment construct (see Appendix 2).

Family of cohabitation partner drinking (ages 27–30). At each point, participants indicated whether a live-in romantic part- ner or spouse drank alcohol heavily (yes/no). Participants were coded as having a heavily drinking partner if they answered “yes” for at least one of the two time points.

Family of cohabitation partner smoking (ages 27–30). Par- ticipants indicated whether a live-in romantic partner or spouse smoked (yes/no). Participants were coded as having a smoking partner if they answered “yes” for at least one of the two time points.

Adult comorbid problem behavior (age 33). Five adult problem behaviors were measured at age 33: tobacco dependence, alcohol abuse or dependence, other drug abuse or dependence, past-year involvement in crime, and sexual risk behavior. Alcohol abuse or dependence, tobacco dependence, illicit drug abuse, high- risk sexual behavior, and crime were each used as indicators of a latent factor of comorbid problem behavior (see Appendix 2).

Control variables. Key demographic control variables re- lated to BD, family environment, and adult risk behavior are included here. Gender and ethnicity were self-reported. Childhood socioeconomic status was assessed by eligibility for the National School Lunch/School Breakfast program at any time in Grades 5, 6, or 7, and was taken from school records. Dichotomous variables of gender, African American ethnicity, Asian American ethnicity, and socioeconomic status were used as controls.

Results

Analyses

All models were estimated using Mplus version 6.1 (Muthén & Muthén, 1998 –2007; Schafer & Graham, 2002). Measures of partner alcohol and tobacco use and the five indicators of the problem behavior latent factor were declared as ordered categor- ical, and the weighted least squares mean and variance-adjusted (WLSMV) estimator was used. The WLSMV estimator applies somewhat more stringent assumptions than full information max- imum likelihood, but still uses the full data set to estimate missing data (see Asparouhov & Muthen, 2010). In the present study, missing data on the outcome variables was 3% for cumulative criminal behavior; 6% for cumulative sexual risk behavior; and 6.2% for the alcohol, tobacco, and drug-related outcomes. Estima- tion of missing data using WLSMV is appropriate when the

amount of missing dependent variable data is not substantial, such as in the present study.

Family of origin general and alcohol- and tobacco-specific environments, family of cohabitation general environment, and comorbid problem behavior were modeled as latent variables (see Appendix 2 for indicator loadings). We used, the chi-square sta- tistic and three indices of model fit (comparative fit index [CFI], Tucker-Lewis Index [TLI], and root-mean-square error of approx- imation [RMSEA]) to evaluate the model fit throughout. Tables 1 and 2 contain intercorrelations of all modeled variables and de- scriptive statistics of the dependent variables.

Modeling age 33 problem behavior as a latent variable allowed us to partition variance of the five indicators into shared variance represented by the problem behavior latent factor and nonshared variance unique to each of the individual behaviors (e.g., variance uniquely associated with tobacco dependence). To test the hypoth- eses regarding the comorbid problem behavior versus the drug- specific effects, we examined associations between predictors and problem behavior as well as between predictors and individual indicators. A path between a predictor and the latent construct thus represents the effect on shared variance in comorbid problem behavior, whereas a path between the same predictor and the residual variance of an indicator represents the effect of the pre- dictor on the nonshared, unique, or specific variance in that indi- cator. This approach has been used in the past to model deviance (McGee & Newcomb, 1992; Newcomb et al., 2002).

Two structural equation models were estimated. The first model expanded on the work of Bailey et al. (2011) that linked general and drug-specific family environments to comorbid problem be- havior at age 24. We used the same measures of general family adolescent environment as Bailey et al., and the same measure of family smoking and drinking environments with the exception of having excluded sibling smoking and drinking due to low factor loadings. In addition, we used a comparable set of outcome mea- sures as Bailey et al., but now operationalized at age 33. We also expanded the model in two ways. First, we included BD as a measure of individual vulnerability and tested whether it moder- ated the relations between family environments and comorbid problem behavior. Second, we controlled for initial behavior prob- lems by adding early delinquency at ages 10 –11 into the model.

We first estimated a model that included all of the hypothesized effects (see nonmediated paths in Appendix 1) and competing hypotheses simultaneously. That is, we tested both general and drug-specific effects of family environments on comorbid problem behavior and unique variances of alcohol and tobacco misuse at age 33 in the same model. We also tested general and drug-specific effects of BD and early delinquency, and the association between delinquency and unique variance in criminal acts. In order to minimize suppression, at this stage we dropped nonsignificant nonhypothesized paths. The complete set of tested paths, esti- mates, and confidence interval for all models can be found in Appendix 2. In the second model, we investigated whether family of cohabitation environments during young adulthood mediated the relations between family of origin influences and age 33 outcomes. We also tested age 18 alcohol and tobacco use as potential mediators between adolescent and adult environments.

We used Mplus to explicitly model nonnormally distributed outcomes. Measures of problem behavior are nonnormative in nonclinical populations and were here modeled as ordered cate-

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

1155GENERAL AND DRUG-SPECIFIC ENVIRONMENTS

T ab

le 1

C o rr

el a ti

o n s

B et

w ee

n M

o d el

V a ri

a b le

s (P

a rt

ic ip

a n ts

W h o

R ep

o rt

ed H

a vi

n g

a S p o u se

o r

L iv

e- In

P a rt

n er

a t

A g e

2 7 – 3 0 )

V ar

ia bl

e 1

2 3

4 5

6 7

8 9

10 11

12 13

14 15

16 17

18 19

1. F

am .

en v.

(1 0–

18 )

— 2.

A lc

. en

v. (1

0– 18

) �

.0 3

— 3.

T ob

. en

v. (1

0– 18

) �

.0 2

.4 9�

� �

— 4.

B eh

. di

si nh

ib it

io n

� .2

7� � �

.1 9�

� �

.1 9�

� —

5. D

el in

qu en

t be

h. �

.2 8�

� �

.0 8

.2 0�

� �

.2 9�

� �

— 6.

F am

. en

v. (2

7– 30

) .3

3� � �

� .0

8 �

.0 9

� .1

5� �

� .1

8� � �

— 7.

P ar

tn er

dr in

ks �

.1 5†

.1 6†

.2 0�

.1 3†

� .0

2 �

.3 5�

� �

— 8.

A lc

. us

e (1

8) �

.1 9�

� �

.1 3†

.0 8

.2 5�

� �

.1 9�

� �

� .1

5� � �

� .0

1 —

9. C

ig .

us e

(1 8)

� .1

8� � �

.0 4

.3 1�

� �

.2 4�

� �

.1 5�

� �

� .0

4 .0

0 .1

8� � �

— 10

. P

ar tn

er sm

ok es

� .1

6� .1

3† .2

8� � �

.2 1�

� �

.1 3�

� .2

0� �

.4 8�

� �

.1 0�

.2 1�

� �

— 11

. A

lc .

di ag

no si

s (3

3) �

.2 6�

� .1

8� .0

7 .2

9� � �

.1 0

� .2

2� �

.2 3�

.2 6�

� �

.1 2

.3 7�

� �

— 12

. T

ob .

di ag

no si

s (3

3) �

.1 9�

� .0

4 .3

3� � �

.3 6�

� �

.2 0�

� �

.1 8�

� .1

7 .1

1 .3

8� � �

.4 0�

� �

.3 0�

� —

13 .

D ru

g di

ag no

si s

(3 3)

� .3

0� �

.0 4

.0 7

.3 1�

� �

.2 1�

� �

.1 9�

.2 8�

.2 5�

� �

.0 8

.5 4�

� �

.6 6�

� .5

3� � �

— 14

. C

ri m

e (3

3) �

.1 9�

� .1

3 .1

0 .3

2� � �

.2 3�

� �

� .1

9� �

.3 2�

� �

.0 9

� .0

4 .2

7� �

.4 8�

� �

.3 8�

� �

.6 5�

� �

15 .

S ex

ua l

ri sk

(3 3)

� .2

5� � �

.1 4�

.2 5�

� �

.3 2�

� �

.2 1�

� �

� .4

0� � �

.3 6�

� �

.1 3�

.0 9

.3 2�

� �

.5 2�

� �

.3 8�

� �

.6 4�

� �

.4 0�

� —

16 .

G en

de r

(m al

e) �

.1 0�

.0 7

.0 1

.2 5�

� �

.2 3�

� �

� .0

6 �

.1 4�

.1 2�

� .0

1 �

.0 2

.3 2�

� �

.1 1

.1 3�

� .2

2� �

.0 5

— 17

. A

si an

A m

er ic

an �

.0 3

� .4

3� � �

� .2

8� � �

� .2

5� � �

� .0

9 .1

1� �

.1 1

� .1

3� �

� .1

7† �

.1 9�

� �

.1 9�

� .3

2� � �

� .3

0� �

� .0

8 �

.2 4�

� �

.0 0

— 18

. A

fr ic

an A

m er

ic an

� .0

3 �

.1 3�

.0 3

.0 9�

.2 0�

� �

� .2

1� � �

� .0

5 .1

0� �

.0 3

.0 7

.1 2†

.2 2�

� .1

4� .2

7� � �

.0 3

� .3

0� � �

— 19

. S

E S

� .1

5� �

� .2

5� � �

.1 3�

� .0

1 .2

0� � �

� .0

7 �

.0 4

.0 5

.0 3

.0 4

� .0

6 .0

7 .0

2 .0

6 .1

4� �

� .0

5 .2

0� � �

.3 2�

� �

—

N o te

. G

en er

al fa

m il

y en

vi ro

nm en

t is

co de

d to

re fl

ec t

ge ne

ra l

po si

ti ve

fa m

il y

fu nc

ti on

in g.

E th

ni ci

ty re

fe re

nc e

gr ou

p is

W hi

te .

F am

. en

v. �

F am

il y

en vi

ro nm

en t;

A lc

. en

v. �

A lc

oh ol

en vi

ro nm

en t;

T ob

. en

v. �

T ob

ac co

en vi

ro nm

en t;

B eh

. �

B eh

av io

r; C

ig .

� C

ig ar

et te

; S

E S

� so

ci oe

co no

m ic

st at

us .

† p

� .1

0. �

p �

.0 5.

� �

p �

.0 1.

� � �

p �

.0 01

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

1156 EPSTEIN, HILL, BAILEY, AND HAWKINS

gorical. However, because Mplus does not estimate residual vari- ance of categorical variables, the more traditional approach of regressing residual variance of the indicator on the predictors is not available. Bailey et al. (2011) used phantom latent variables to partition residual variance of the indicators. In this article, we chose a different approach where the indicators are regressed directly onto the predictors without first formally partitioning residual variance. The two approaches yield identical unstandard- ized estimates, and we tested both approaches to ensure model integrity. We chose to present standardized estimates from the second approach because of its relative visual simplicity and greater ease of replication for future research.

Family of Origin Environments, BD, and Age 33 Outcomes

Our first hypothesis concerned the effect of general and drug-specific family environments in the family of origin on comorbid problem behaviors and drug-specific outcomes at age 33, and our second hypothesis concerned the effect of adoles- cent BD on these outcomes. Accordingly, in the first model we tested these hypotheses by examining the relations between general family environment, alcohol environment, and tobacco environment in the family of origin, and problem behaviors at age 33 (see Figure 1). We tested the hypothesized associations between (a) general family environment and comorbid problem behavior at age 33, (b) alcohol environment and age 33 alcohol abuse or dependence, and (c) tobacco environment and age 33 tobacco dependence. Additionally, we examined the associa- tions between childhood BD and delinquency and age 33 out- comes. Control variables were allowed to correlate with the predictors, and were set to predict problem behavior. The fit indices showed good model fit, (�2(148) � 225.06, CFI � .95, TLI � .93, RMSEA � .03.

Consistent with predictions, positive family environment dur- ing adolescence had a protective effect and was negatively associated with comorbid problem behaviors at age 33, but was not uniquely associated with any specific behaviors. Also con- sistent with our hypotheses, smoking environment in the family of origin was uniquely linked with tobacco dependence in adulthood, suggesting that early exposure to tobacco may pre- dispose children to initiate and maintain smoking into adult- hood. However, family of origin alcohol environment was not

associated with unique variance of alcohol abuse or dependence at age 33. In support of the second hypothesis, BD was asso- ciated with an increased rate of engaging in comorbid problem behaviors, but not specific problem behaviors at age 33. Con- sistent with our prediction, there was a moderate association between BD and delinquent behavior. However, the hypothe- sized links between early delinquency and comorbid problem behavior or crime were not supported by the results.

The possible interaction between the three adolescent environ- ments and BD was explored using multigroup comparisons. We created two groups by cutting participants’ BD scores at the 33rd percentile. Sensitivity analysis changing the cutoff for the high-BD group to the 40th percentile yielded similar results. We performed the multigroup comparisons using the DIFFTEST function of Mplus (Muthén & Muthén, 1998 –2007). All factor loadings and structural parameters were constrained to be equal across the two groups in the constrained model. We then compared the uncon- strained model with three separate models in which the appropriate path between each of the three predictors and the outcome was estimated freely. The DIFFTEST procedure showed no significant interaction between BD and family of origin general environ- ment’s effect on problem behavior, �2 difference (�1) � 2.99, p � .05; family of origin alcohol environment’s effect on alcohol abuse or dependence, �2 difference (�1) � 1.15, p � .05; or family of origin tobacco environment’s effect on tobacco dependence, �2

difference (�1) � 1.97, p � .05. Thus, BD appeared to contribute additively to comorbid problem behavior in adulthood but did not moderate family of origin general environmental influence. That is, a positive family of origin environment had the same inhibiting effect on problem behavior regardless of the degree of partici- pants’ BD.

Associations between predictor variables (see Table 1) indicate that childhood BD was associated with less positive home envi- ronment and more pronounced alcohol and tobacco environments. BD was linked to delinquent behavior, which was in turn associ- ated with more prominent tobacco environment and less positive general family environment. Alcohol and tobacco environments were significantly intercorrelated, but neither was associated with general home environment. African American children tended to come from families where the alcohol environment was less pro- nounced. Male gender and identifying as African American were associated with more BD and delinquent behavior. Being male was

Table 2 Descriptive Statistics

Variable name M (SD) Range n(%) reporting � 0 behaviors/symptoms

Days in past month drank alcohol (age 18) 1.95 (4.29) 0–30 190 (40.5) Cigarettes per day, past month (age 18) 2.49 (6.19) 0–30 117 (24.9) Partner drinks heavily 0.15 (0.36) 0–1 70 (14.9) Partner smokes 0.39 (0.49) 0–1 184 (39.2) Comorbid problem behavior

Alcohol abuse or dependence diagnosis 0.13 (0.35) 0–1 63 (13.4) Tobacco dependence 0.18 (0.38) 0–1 78 (16.6) Illicit drug abuse or dependence diagnosis 0.08 (0.28) 0–1 37 (7.9) Crime 0.22 (0.59) 0–4 60 (12.8) Risky sexual behavior 0.40 (0.66) 0–4 171 (36.5)

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

1157GENERAL AND DRUG-SPECIFIC ENVIRONMENTS

also associated with less positive family environments. Asian American children, however, were less likely to exhibit symptoms of BD and were also less likely to come from smoking or drinking families. Lower socioeconomic status was associated with less positive family environment and lower family emphasis on alco- hol, but a greater presence of nicotine, and greater engagement in delinquent behavior.

Examining Adult Family of Cohabitation Environments: A Mediational Analysis

Our third and fourth hypotheses considered the effects of environmental influences in the family of cohabitation. We predicted an additive-mediational model in which both family of origin and family of cohabitation environments influence risk behavior, alcohol abuse or dependence, and tobacco depen- dence. We also tested whether the effects of drug-specific adolescent and adult environments were mediated by partici- pants’ substance use in late adolescence (age 18). Retaining all of the hypothesized paths from Model 1, we added the first block of age 18 alcohol and tobacco use as potential mediators between family of origin and family of cohabitation environ- ments (see Figure 2). Next, the second block of age 27–30 romantic partner variables were added as mediators between

age 18 substance use and age 33 outcomes. In general, for each dependent variable, we tested substance-concordant and general influences of environments (e.g., adolescent alcohol environ- ment to alcohol use at age 18; general family environment to age 18 alcohol use). Age 18 alcohol and tobacco use were set to mediate all adolescent variables and partner substance use. Each block of the potential mediators was regressed onto the demographic variables, and variables within a block were al- lowed to intercorrelate. The final model shown in Figure 2 fit the data well, �2(263) � 374.98, CFI � .95, TLI � .93, RMSEA � .03. As a sensitivity check, associations between age 18 substance use and general outcomes (age 27–30 general family environment and age 33 comorbid problem behavior) were tested separately and found to be nonsignificant. Addi- tionally, we tested whether the marital status of participants at ages 27 and 30 or the presence of children living in the home affected the results. The DIFFTEST procedure in WLSMV estimator showed that neither marital status nor the presence of children moderated findings. These changes were not included in the final model.

In accordance with our third hypothesis, there was a strong positive association between general environments in family of origin and family of cohabitation. Family of origin general envi-

Figure 1. Estimated model of adolescent environments and age 33 outcomes for participants who reported having a spouse or live-in dating partner at age 27–30 (n � 469), �2(148) � 225.06, comparative fit index � .95, Tucker-Lewis Index � .93, root-mean-square error of approximation � .03. Ethnicity referent is White. General Family Environment is coded to reflect general positive family functioning. All dependent variables are controlled for demographics, which are also correlated with the predictors. SES � socioeconomic status. �� p � .01. �� p � .001.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

1158 EPSTEIN, HILL, BAILEY, AND HAWKINS

ronment had a protective effect on the likelihood of having a substance-using partner in young adulthood. We also found con- tinuity of adolescent tobacco environment and choice of smoking partner, which was partially mediated by age 18 tobacco use. There were no direct effects of adolescent alcohol environment on partner drinking, although continuity from drinking family to alcohol use at 18 was suggested. Unlike tobacco, there was no indication that participants selected partners on the basis of their own alcohol use at age 18.

The fourth hypothesis specified a mediated model in which the effects of family of origin environments on age 33 outcomes were partially mediated by family of cohabitation variables. As ex- pected, results indicated strong continuity from adolescent smok- ing and drinking to age 33 substance use problems (see Figure 2). Consistent with hypotheses, however, after accounting for age 18 substance use and adding the young adulthood variables, a number of direct associations between adolescent predictors and adult outcomes remained significant. Family of origin general environ- ment continued to play a protective role against engaging in comorbid problem behavior at age 33, indicating a lasting protec- tive effect of positive family functioning during adolescence well into adulthood. A trend toward intergeneration continuity in smok- ing behavior was indicated by the association between family of origin tobacco environment and greater likelihood of developing tobacco dependence at age 33, over and above initiating smoking by age 18 and having a smoking partner during ages 27–30. The

effects of BD on comorbid problem behavior also persisted after the age 18 substance use and partner environments were added to the model.

Similar to family of origin general environment, general environment in the family of cohabitation showed a trending protective effect on comorbid problem behavior at age 33. Having a drinking partner or smoking partner was strongly associated with engaging in comorbid problem behavior, par- tially mediating the relationship between family environments in adolescence and problem behavior at age 33. However, substance-specific effects of partner drug use were not supported. Next, indirect effects were computed using the bias-corrected boot- strap confidence intervals (BCBOOTSTRAP; Shrout & Bolger, 2002). There were three significant ( p � .05) indirect effects on age 33 tobacco dependence via age 18 smoking: adolescent gen- eral family environment (probit � � �.04), BD (� � .07), and adolescent smoking environment (� � .10). Adolescent general family environment had indirect effects on problem behavior at age 33 (� � �.06) through partner smoking and on alcohol abuse and dependence at 33 (� � �.03) through partner drinking. Finally, tobacco environment had an indirect effect on comorbid problem behavior through partner smoking (� � .07).

In regards to demographic controls, results indicated a strong negative relationship between identifying as African American and positive family environment in the family of cohabitation. Com- pared with women, men were less likely to report a heavily

Figure 2. Estimated model of adolescent and adult environments and age 33 outcomes for participants who reported having a spouse or live-in dating partner at age 27–30 (n � 469), �2(263) � 374.98, comparative fit index � .95, Tucker-Lewis Index � .93, root-mean-square error of approximation � .03. Ethnicity referent category is White. General Family Environment is coded to reflect general positive family functioning. All dependent variables are controlled for demographics, which are also correlated with the predictors. Estimated, but now shown in the figure, are the correlations between general family environment in the family of cohabitation (A), partner drinking (B) and partner smoking (C), which were AB � �.35��, AC � �.12 , BC � .46��, and correlation between alcohol and tobacco use at age 18, .12��. p � .10. � p � .05. �� p � .01. �� p � .001.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

1159GENERAL AND DRUG-SPECIFIC ENVIRONMENTS

drinking partner. Being male and African American predicted greater comorbid problem behavior at age 33. Regression coeffi- cients and confidence intervals are available in the Supplemental Materials, Appendix 2.

Discussion

The conceptual and methodological approaches of this work illustrate three organizing principles for representing the social environment in complex models of addiction. The first principle concerns a clear delineation of a functional domain of influence, such as family, peer, school/work, and neighborhood. The present study focused on the family domain. Second, within each domain, general functioning can be distinguished from the drug-specific aspects of that domain. In the present work, we examined the differential impact of positive general family environment from those influences that are specifically related to tobacco or alcohol. The third principle calls for locating a social environment within its developmental context. In the present study, different patterns of prediction emerged for adolescent and adult family environ- ments. The present study is also based on the organizing heuristic of examining general deviance as measured by comorbid involve- ment in multiple problem behaviors as compared with involvement only in specific component problem behaviors. Directly modeling the comorbidity between substance use and other externalizing behaviors has allowed us to investigate both general predictors of comorbid problem behaviors and specific predictors of alcohol and tobacco problems that are not comorbid with other problem be- haviors.

General Versus Specific Predictors of General Versus Specific Outcomes

Our first major finding concerned identifying environmental factors that uniquely predict alcohol and tobacco problems, over and above their effect on comorbid problem behavior. Results indicate that general family functioning in adolescence predicted comorbid problem behavior at age 33 and that exposure to tobacco in the family of origin was uniquely linked to tobacco dependence in adulthood. These findings are consistent with previous findings by Bailey et al. (2011) on age 24 outcomes. However, the analo- gous association between adolescent family alcohol environment and later alcohol abuse or dependence at age 33 was not replicated. It is possible that the effect of early alcohol environment is stronger during emerging adulthood but not sustained later in life.

We also examined the effects of BD as a person-level risk factor previously linked to both adult substance use and problem behav- ior (Button et al., 2007; Fu et al., 2002) and controlled for initial behavioral problems at ages 10 –11. We found that BD had a strong direct effect on comorbid problem behavior above and beyond the impact of environmental factors. Early delinquency and BD were moderately related, although unlike BD, delinquency did not have an independent effect on comorbid behavior prob- lems. We also investigated whether the adverse effects of BD were either moderated by consistently positive adolescent family func- tioning or exacerbated by exposure to alcohol and tobacco influ- ences. None of the interactions between BD and the three family environments were significant, suggesting that these influences were additive and not multiplicative. It is possible that BD inter-

acts with only certain aspects of the family environment (e.g., consistently poor family management as in Hill et al., 2010) or only during specific sensitive periods in development. Future studies need to continue exploring the potential interactions be- tween environmental influences and person-level factors.

General and Drug-Specific Environmental Continuity

The second major finding concerned the environmental conti- nuity of general and drug-specific environments in the family of origin to environments related to cohabiting partnerships in young adulthood. Early general family environmental factors, such as the amount of family conflict and the strength of bonding, appeared to be highly predictive of the quality of romantic relationships in adulthood. This is consistent with the Social Development Model (Hawkins & Weis, 1985) as well as with findings from literatures on parenting and attachment (e.g., Leveridge et al., 2005; Mick- elson, Kessler, & Shaver, 1997; Shaver & Brennan, 1992). More- over, a positive general family environment in adolescence was associated with a lesser likelihood of having a smoking and a drinking partner during young adulthood. These effects suggest that practices in the family of origin, such as conflict resolution and child monitoring, have important and long-lasting implications for both general and drug-specific outcomes.

Consistent with prediction, we found continuity from family of origin smoking environment to choosing a smoking partner. This relationship was partially mediated by smoking behavior at age 18, suggesting that children of smokers are more likely to smoke themselves and to choose to partner with a smoker (e.g, Falba & Sindelar, 2008; Kuo et al., 2007). The direct effect of tobacco environment on choice of partner, however, indicates an additional influence that family of origin has on later life choices. For example, children raised in smoking families may become accus- tomed to the smell of tobacco and its near-constant presence, possibly making the odors familiar and even pleasing in another person (e.g., Etcheverry & Agnew, 2009; Forestell & Mennella, 2005). Continuity in alcohol family environment did not emerge in our analyses, although there was a trend suggesting that presence of alcohol in the family during adolescence increases the likeli- hood of drinking at age 18. Paired with a nonsignificant connection to age 33 alcohol abuse or dependence, this finding may indicate that adolescent family alcohol environment is a weak predictor of long-term offspring outcomes and choices. It is possible that alcohol use assessed in the present study reflected normative moderate alcohol use and thus was not predictive of offspring problem behavior.

The differential pattern of results for alcohol and tobacco sug- gests the possibility that parental tobacco use differs from parental drinking in its visibility and accessibility to the child. It may be possible to shield children from parental alcohol use by engaging in drinking late in the evening or only occasionally. Furthermore, although parents’ moderate drinking is not discouraged in society, many parents disapprove of their children’s drinking during child- hood and adolescence. Thus, there is an inherent contradiction between some parents’ drinking behavior and their attitudes to- ward alcohol that may weaken the relation between overall family alcohol environment and children’s alcohol problems. However, children raised in smoking families who are exposed to tobacco through observation of parental behavior, the smell of cigarettes,

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

1160 EPSTEIN, HILL, BAILEY, AND HAWKINS

and inhalation of secondhand smoke are also less likely to expe- rience parental discouragement from smoking. Because of the highly addictive nature of nicotine and the high stability of smok- ing behavior in adults (Chassin, Presson, Pitts, & Sherman, 2000), children of smokers are exposed to tobacco throughout the day for many years, have early opportunities to initiate tobacco use them- selves, and have an available supply of the parents’ tobacco products. These patterns of exposure may explain the strong con- tinuity in tobacco-related behavior in our analyses. Finally, it is possible that there are genetic mechanisms unique to nicotine that are transmitted from parents to children or that secondhand smoke exposure during sensitive periods in early development alters children’s neurochemistry in a way that makes children of smokers more susceptible to later tobacco dependence (Volkow & Li, 2005).

Family of Origin Influences and Family of Cohabitation Mediators

The third set of findings concerned the mediational role that adult environment plays in predicting age 33 outcomes. Our results indicated that both sets of general family environments had an effect on comorbid problem behavior. The long-reaching influence of adolescent family functioning is consistent with Moffit’s (1993a) notion that life course antisocial tendencies are rooted in genetic and early environmental factors and that risk-taking tra- jectories are set early on. The protective effect of positive envi- ronment in the family of cohabitation, however, suggests that targets for preventive interventions extend into adulthood.

With regard to tobacco dependence, we generally found that early family contexts continued to predict tobacco dependence at age 33, even after accounting for smoking behavior at age 18 and having a smoking partner. We did not find a parallel effect for either family of origin or having a drinking partner for alcohol abuse or dependence, over and above age 18 alcohol use. Although substance-specific effects of partner behavior were not evident, both partner smoking and drinking were associated with more comorbid problem behavior, reiterating the notion that both ado- lescent and young adult environmental influences play an impor- tant role in predicting problem behavior.

Finally, consistent with predictions, results indicated that greater childhood BD increased the risk of engaging in problem behavior at age 33, even after including baseline problem behavior and family of cohabitation environments in the model. Although the links between BD, tobacco, and alcohol problems have been reported in prior studies (Brook et al., 2006; Hill et al., 2000), our results suggest that BD plays a greater role in predicting comorbid problem behavior than the unique variance in alcohol or tobacco problems. It is possible that the associations with BD in other studies of tobacco and alcohol addiction emerged as a result of substantial variance that these problems share with problem be- havior in general. Partitioning shared variance of comorbid prob- lems from unique variance of tobacco and alcohol problems should help identify drug-specific predictors that can be addressed with drug-specific interventions.

Some limitations should be considered when interpreting the findings. First, the SSDP sample is a school-based urban sample from the Pacific Northwest. Second, due to the relatively small number of Native Americans in the sample, they were not included

in analyses. This is an important demographic group with unique risk factors, and future studies need to closely examine person– environment predictors of tobacco and alcohol addiction in this and other minority populations. Third, drug-specific environments in the family of cohabitation were measured with a single item that may not have captured sufficient variability in partner relation- ships. Furthermore, although possible effects of marriage status and the presence of children in the home were tested in this study, other family structure variables, such as relationship duration, need to be considered. Studies in this area should also examine partner attitudes and partner-provided opportunities for alcohol use both as they relate to childhood alcohol environment and as predictors of future alcohol dependence. Finally, using a longitudinal design is not sufficient to conclusively determine causation. However, we have included a number of controls in our model that, although not exhaustive, provide a reasonable platform for causal inference (Bullock, Harlow, & Mulaik, 1994).

Conclusions and Implications for Subsequent Research

This study presents an innovative approach to examining person– environment predictors of alcohol and tobacco problems. A major strength of this study lies in the separation of shared variance (comorbid problem behavior) from variance in tobacco and alcohol problems, which helps distinguish causes of general risk-taking behavior from those causes specific to alcohol and tobacco dependence. This approach has important implications for future research, particularly for emerging work in gene– environment interplay in the development of addiction. The pres- ent study offers a model for conceptualizing environmental influ- ences suitable for later use in studies of gene– environment interplay that is broad enough to be flexible in multiple research studies, yet specific enough to identify targets for preventive intervention.

References

Andrews, J. A., Hops, H., & Duncan, S. C. (1997). Adolescent modeling of parent substance use: The moderating effect of the relationship with the parent. Journal of Family Psychology, 11, 259 –270. doi:10.1037/ 0893-3200.11.3.259

Asparouhov, T., & Muthen, B. (2010). Weighted least squares estimation with missing data [Technical report]. Retrieved from http://www.statmodel.com/ download/GstrucMissingRevision.pdf

Bachman, J. G., O’Malley, P. M., Schulenberg, J. E., Johnston, L. D., Bryant, A. L., & Merline, A. C. (2002). The decline of substance use in young adulthood: Changes in social activities, roles, and beliefs. Mah- wah, NJ: Lawrence Erlbaum Associates.

Bailey, J. A., Hill, K. G., Meacham, M. C., Young, S. E., & Hawkins, J. D. (2011). Strategies for characterizing complex phenotypes and environ- ments: General and specific family environmental predictors of young adult tobacco dependence, alcohol use disorder, and co-occurring prob- lems. Drug and Alcohol Dependence, 118, 444 – 451. doi:10.1016/ j.drugalcdep.2011.05.002

Bailey, J. A., Hill, K. G., Oesterle, S., & Hawkins, J. D. (2006). Linking substance use and problem behavior across three generations. Journal of Abnormal Child Psychology, 34, 263–282. doi:10.1007/s10802-006- 9033-z

Bricker, J. B., Peterson, A. V., Jr., Andersen, M., Leroux, B. G., Rajan, K., & Sarason, I. G. (2006). Close friends’, parents’, and older siblings’

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

1161GENERAL AND DRUG-SPECIFIC ENVIRONMENTS

http://dx.doi.org/10.1037/0893-3200.11.3.259

http://www.statmodel.com/download/GstrucMissingRevision.pdf

http://dx.doi.org/10.1016/j.drugalcdep.2011.05.002

http://dx.doi.org/10.1007/s10802-006-9033-z

smoking: Reevaluating their influence on children’s smoking. Nicotine & Tobacco Research, 8, 217–226. doi:10.1080/14622200600576339

Brook, J. S., Ning, Y., & Brook, D. W. (2006). Personality risk factors associated with trajectories of tobacco use. The American Journal on Addictions, 15, 426 – 433. doi:10.1080/10550490600996363

Bullock, H. E., Harlow, L. L., & Mulaik, S. A. (1994). Causation issues in structural equation modeling research. Structural Equation Mod- eling: A Multidisciplinary Journal, 1, 253–267. doi:10.1080/ 10705519409539977

Button, T. M. M., Rhee, S. H., Hewitt, J. K., Young, S. E., Corley, R. P., & Stallings, M. C. (2007). The role of conduct disorder in explaining the comorbidity between alcohol and illicit drug dependence in adolescence. Drug and Alcohol Dependence, 87, 46 –53. doi:10.1016/j.drugalcdep .2006.07.012

Catalano, R. F., & Hawkins, J. D. (1996). The social development model: A theory of antisocial behavior. In J. D. Hawkins (Ed.), Delinquency and crime: Current theories (pp. 149 –197). New York, NY: Cambridge University Press.

Chassin, L., Pitts, S. C., & Prost, J. (2002). Binge drinking trajectories from adolescence to emerging adulthood in a high-risk sample: Predictors and substance abuse outcomes. Journal of Consulting and Clinical Psychol- ogy, 70, 67–78. doi:10.1037/0022-006X.70.1.67

Chassin, L., Presson, C. C., Pitts, S. C., & Sherman, S. J. (2000). The natural history of cigarette smoking from adolescence to adulthood in a midwestern community sample: Multiple trajectories and their psycho- social correlates. Health Psychology, 19, 223–231. doi:10.1037/0278- 6133.19.3.223

Chassin, L., Presson, C., Rose, J., Sherman, S. J., & Prost, J. (2002). Parental smoking cessation and adolescent smoking. Journal of Pediat- ric Psychology, 27, 485– 496. doi:10.1093/jpepsy/27.6.485

Cloninger, R. C., Sigvardsson, S., & Bohman, M. (1996). Type I and Type II alcoholism: An update. Alcohol Health and Research World, 20, 18 –23.

Conway, K. P., Compton, W. M., & Miller, P. M. (2006). Novel ap- proaches to phenotyping drug abuse. Addictive Behaviors, 31, 923–928. doi:10.1016/S0306-4603(06)00146-8

Donnellan, M. B., Larsen-Rife, D., & Conger, R. D. (2005). Personality, family history, and competence in early adult romantic relationships. Journal of Personality and Social Psychology, 88, 562–576. doi: 10.1037/0022-3514.88.3.562

Engels, R. C. M. E., Knibbe, R. A., de Vries, H., Drop, M. J., & van Breukelen, G. J. P. (1999). Influences of parental and best friends’ smoking and drinking on adolescent use: A longitudinal study. Journal of Applied Social Psychology, 29, 337–361. doi:10.1111/j.1559- 1816.1999.tb01390.x

Engels, R. C. M. E., Vermulst, A. A., Dubas, J. S., Bot, S. M., & Gerris, J. (2005). Long-term effects of family functioning and child character- istics on problem drinking in young adulthood. European Addiction Research, 11, 32–37. doi:10.1159/000081414

Engels, R. C. M. E., Vitaro, F., Den Exter Blokland, E., de Kemp, R., & Scholte, R. H. J. (2004). Influence and selection processes in friendships and adolescent smoking behaviour: The role of parental smoking. Jour- nal of Adolescence, 27, 531–544. doi:10.1016/j.adolescence.2004 .06.006

Etcheverry, P. E., & Agnew, C. R. (2009). Similarity in cigarette smoking attracts: A prospective study of romantic partner selection by own smoking and smoker prototypes. Psychology of Addictive Behaviors, 23, 632– 643. doi:10.1037/a0017370

Falba, T. A., & Sindelar, J. L. (2008). Spousal concordance in health behavior change. Health Services Research, 43, 96 –116. doi:10.1111/ j.1475-6773.2007.00754.x

Fischer, J. L., Fitzpatrick, J., & Cleveland, H. H. (2007). Linking family functioning to dating relationship quality via novelty-seeking and harm-

avoidance personality pathways. Journal of Social and Personal Rela- tionships, 24, 575–590. doi:10.1177/0265407507079257

Fleming, C. B., Kim, H., Harachi, T. W., & Catalano, R. F. (2002). Family processes for children in early elementary school as predictors of smok- ing initiation. Journal of Adolescent Health, 30, 184 –189. doi:10.1016/ S1054-139X(01)00327-5

Forestell, C. A., & Mennella, J. A. (2005). Children’s hedonic judgments of cigarette smoke odor: Effects of parental smoking and maternal mood. Psychology of Addictive Behaviors, 19, 423– 432. doi:10.1037/0893- 164X.19.4.423

Fu, Q., Heath, A. C., Bucholz, K. K., Nelson, E., Goldberg, J., Lyons, M. J., . . . Eisen, S. A. (2002). Shared genetic risk of major depression, alcohol dependence, and marijuana dependence: Contribution of antiso- cial personality disorder in men. Archives of General Psychiatry, 59, 1125–1132. doi:10.1001/archpsyc.59.12.1125

Galaif, E. R., Stein, J. A., Newcomb, M. D., & Bernstein, D. P. (2001). Gender differences in the prediction of problem alcohol use in adult- hood: Exploring the influence of family factors and childhood maltreat- ment. Journal of Studies on Alcohol, 62, 486 – 493.

Guo, J., Hawkins, J. D., Hill, K. G., & Abbott, R. D. (2001). Childhood and adolescent predictors of alcohol abuse and dependence in young adult- hood. Journal of Studies on Alcohol, 62, 754 –762.

Harter, S. L. (2000). Psychosocial adjustment of adult children of alcohol- ics: A review of the recent empirical literature. Clinical Psychology Review, 20, 311–337. doi:10.1016/S0272-7358(98)00084-1

Hawkins, J. D., Catalano, R. F., & Miller, J. Y. (1992). Risk and protective factors for alcohol and other drug problems in adolescence and early adulthood: Implications for substance-abuse prevention. Psychological Bulletin, 112, 64 –105. doi:10.1037/0033-2909.112.1.64

Hawkins, J. D., Kosterman, R., Catalano, R. F., Hill, K. G., & Abbott, R. D. (2005). Promoting positive adult functioning through social de- velopment intervention in childhood: Long-term effects from the Seattle Social Development Project. Archives of Pediatrics and Adolescent Medicine, 159, 25–31. doi:10.1001/archpedi.159.1.25

Hawkins, J. D., & Weis, J. G. (1985). The social development model: An integrated approach to delinquency prevention. Journal of Primary Prevention, 6, 73–97. doi:10.1007/BF01325432

Hill, K. G., Hawkins, J. D., Bailey, J. A., Catalano, R. F., Abbott, R. D., & Shapiro, V. (2010). Person-environment interaction in the prediction of alcohol abuse and alcohol dependence in adulthood. Drug & Alcohol Dependence, 110, 62– 69. doi:10.1016/j.drugalcdep.2010.02.005

Hill, K. G., Hawkins, J. D., Catalano, R. F., Abbott, R. D., & Guo, J. (2005). Family influences on the risk of daily smoking initiation. Jour- nal of Adolescent Health, 37, 202–210. doi:10.1016/j.jadohealth .2004.08.014

Hill, K. G., White, H. R., Chung, I-J., Hawkins, J. D., & Catalano, R. F. (2000). Early adult outcomes of adolescent binge drinking: Person- and variable-centered analyses of binge drinking trajectories. Alcoholism: Clinical and Experimental Research, 24, 892–901. doi:10.1111/j.1530- 0277.2000.tb02071.x

Hops, H., Tildesley, E., Lichtenstein, E., Ary, D., & Sherman, L. (1990). Parent-adolescent problem-solving interactions and drug use. American Journal of Drug and Alcohol Abuse, 16, 239 –258. doi:10.3109/ 00952999009001586

Hu, M-C., Davies, M., & Kandel, D. B. (2006). Epidemiology and corre- lates of daily smoking and nicotine dependence among young adults in the United States. American Journal of Public Health, 96, 299 –308. doi:10.2105/AJPH.2004.057232

Iacono, W. G., Carlson, S. R., Taylor, J., Elkins, I. J., & McGue, M. (1999). Behavioral disinhibition and the development of substance-use disor- ders: Findings from the Minnesota Twin Family Study. Development & Psychopathology, 11, 869 –900. doi:10.1017/S0954579499002369

Iacono, W. G., Malone, S. M., & McGue, M. (2008). Behavioral disinhi- bition and the development of early-onset addiction: Common and

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

1162 EPSTEIN, HILL, BAILEY, AND HAWKINS

http://dx.doi.org/10.1080/14622200600576339

http://dx.doi.org/10.1080/10550490600996363

http://dx.doi.org/10.1080/10705519409539977

http://dx.doi.org/10.1016/j.drugalcdep.2006.07.012

http://dx.doi.org/10.1037/0022-006X.70.1.67

http://dx.doi.org/10.1037/0278-6133.19.3.223

http://dx.doi.org/10.1093/jpepsy/27.6.485

http://dx.doi.org/10.1016/S0306-4603%2806%2900146-8

http://dx.doi.org/10.1037/0022-3514.88.3.562

http://dx.doi.org/10.1111/j.1559-1816.1999.tb01390.x

http://dx.doi.org/10.1159/000081414

http://dx.doi.org/10.1016/j.adolescence.2004.06.006

http://dx.doi.org/10.1037/a0017370

http://dx.doi.org/10.1111/j.1475-6773.2007.00754.x

http://dx.doi.org/10.1177/0265407507079257

http://dx.doi.org/10.1016/S1054-139X%2801%2900327-5

http://dx.doi.org/10.1037/0893-164X.19.4.423

http://dx.doi.org/10.1001/archpsyc.59.12.1125

http://dx.doi.org/10.1016/S0272-7358%2898%2900084-1

http://dx.doi.org/10.1037/0033-2909.112.1.64

http://dx.doi.org/10.1001/archpedi.159.1.25

http://dx.doi.org/10.1007/BF01325432

http://dx.doi.org/10.1016/j.drugalcdep.2010.02.005

http://dx.doi.org/10.1016/j.jadohealth.2004.08.014

http://dx.doi.org/10.1111/j.1530-0277.2000.tb02071.x

http://dx.doi.org/10.3109/00952999009001586

http://dx.doi.org/10.2105/AJPH.2004.057232

http://dx.doi.org/10.1017/S0954579499002369

specific influences. Annual Review of Clinical Psychology, 4, 325–348. doi:10.1146/annurev.clinpsy.4.022007.141157

Jackson, K. M., & Sher, K. J. (2005). Similarities and differences of longitudinal phenotypes across alternate indices of alcohol involvement: A methodologic comparison of trajectory approaches. Psychology of Addictive Behaviors, 19, 339 –351. doi:10.1037/0893-164X.19.4.339

Jackson, K. M., Sher, K. J., & Schulenberg, J. E. (2005). Conjoint devel- opmental trajectories of young adult alcohol and tobacco use. Journal of Abnormal Psychology, 114, 612– 626. doi:10.1037/0021-843X.114 .4.612

Johnson, J. L., & Leff, M. (1999). Children of substance abusers: Overview of research findings. Pediatrics, 103, 1085–1099.

Johnston, L. D., O’Malley, P. M., Bachman, J. G., & Schulenberg, J. E. (2011). Monitoring the future national survey results on drug use, 1975–2010. Volume I: Secondary school students. Ann Arbor: Institute for Social Research, The University of Michigan.

Kreek, M. J., Nielsen, D. A., Butelman, E. R., & LaForge, K. S. (2005). Genetic influences on impulsivity, risk taking, stress responsivity and vulnerability to drug abuse and addiction. Nature Neuroscience, 8, 1450 –1457. doi:10.1038/nn1583

Kuo, P-H., Wood, P., Morley, K. I., Madden, P., Martin, N. G., & Heath, A. C. (2007). Cohort trends in prevalence and spousal concordance for smoking. Drug and Alcohol Dependence, 88, 122–129. doi:10.1016/ j.drugalcdep.2006.09.021

Latendresse, S. J., Rose, R. J., Viken, R. J., Pulkkinen, L., Kaprio, J., & Dick, D. M. (2008). Parenting mechanisms in links between parents’ and adolescents’ alcohol use behaviors. Alcoholism: Clinical and Experi- mental Research, 32, 322–330. doi:10.1111/j.1530-0277.2007.00583.x

Laub, J. H., Nagin, D. S., & Sampson, R. J. (1998). Trajectories of change in criminal offending: Good marriages and the desistance process. Amer- ican Sociological Review, 63, 225–238. doi:10.2307/2657324

Leveridge, M., Stoltenberg, C., & Beesley, D. (2005). Relationship of attachment style to personality factors and family interaction patterns. Contemporary Family Therapy: An International Journal, 27, 577–597. doi:10.1007/s10591-005-8243-9

Loeber, R. T., & Dishion, T. (1983). Early predictors of male delinquency: A review. Psychological Bulletin, 94, 68 –99. doi:10.1037/0033- 2909.94.1.68

Maggs, J. L., & Schulenberg, J. E. (2004). Trajectories of alcohol use during the transition to adulthood. Alcohol Research & Health, 28, 195–201.

McGee, L., & Newcomb, M. D. (1992). General deviance syndrome: Expanded hierarchical evaluations at four ages from early adolescence to adulthood. Journal of Consulting and Clinical Psychology, 60, 766 –776. doi:10.1037/0022-006X.60.5.766

McGue, M., Iacono, W. G., & Krueger, R. (2006). The association of early adolescent problem behavior and adult psychopathology: A multivariate behavioral genetic perspective. Behavior Genetics, 36, 591– 602. doi: 10.1007/s10519-006-9061-z

Merline, A., Jager, J., & Schulenberg, J. E. (2008). Adolescent risk factors for adult alcohol use and abuse: Stability and change of predictive value across early and middle adulthood. Addiction, 103, 84 –99. doi:10.1111/ j.1360-0443.2008.02178.x

Mickelson, K. D., Kessler, R. C., & Shaver, P. R. (1997). Adult attachment in a nationally representative sample. Journal of Personality and Social Psychology, 73, 1092–1106. doi:10.1037/0022-3514.73.5.1092

Moffitt, T. E. (1993a). Adolescence-limited and life-course-persistent an- tisocial behavior: A developmental taxonomy. Psychological Review, 100, 674 –701. doi:10.1037/0033-295X.100.4.674

Moffitt, T. E. (1993b). The neuropsychology of conduct disorder. Develop- ment & Psychopathology, 5, 135–151. doi:10.1017/S0954579400004302

Moffitt, T. E. (2003). Life-course-persistent and adolescence-limited anti- social behavior: A 10-year research review and a research agenda. In

B. B. Lahey, T. E. Moffitt, & A Caspi (Eds.), Causes of conduct disorder and juvenile delinquency (pp. 49 –75). New York, NY: Guilford Press.

Moffitt, T. E., & Caspi, A. (2001). Childhood predictors differentiate life-course persistent and adolescence-limited antisocial pathways among males and females. Development and Psychopathology, 13, 355– 375. doi:10.1017/S0954579401002097

Muthén, L. K., & Muthén, B. O. (1998 –2007). Mplus user’s guide (4th ed.). Los Angeles, CA: Author.

Neighbors, B., Kempton, T., & Forehand, R. (1992). Co-occurrence of substance abuse with conduct, anxiety, and depression disorders in juvenile delinquents. Addictive Behaviors, 17, 379 –386. doi:10.1016/ 0306-4603(92)90043-U

Newcomb, M. D., Abbott, R. D., Catalano, R. F., Hawkins, J. D., Battin- Pearson, S., & Hill, K. (2002). Mediational and deviance theories of late high school failure: Process roles of structural strains, academic com- petence, and general versus specific problem behavior. Journal of Coun- seling Psychology, 49, 172–186. doi:10.1037/0022-0167.49.2.172

Newcomb, M. D., & Loeb, T. B. (1999). Poor parenting as an adult problem behavior: General deviance, deviant attitudes, inadequate fam- ily support and bonding, or just bad parents? Journal of Family Psy- chology, 13, 175–193. doi:10.1037/0893-3200.13.2.175

Olmsted, M. E., Crowell, J. A., & Waters, E. (2003). Assortative mating among adult children of alcoholics and alcoholics. Family Relations, 52, 64 –71. doi:10.1111/j.1741-3729.2003.00064.x

Rhule-Louie, D. M., & McMahon, R. J. (2007). Problem behavior and romantic relationships: Assortative mating, behavior contagion, and desistance. Clinical Child and Family Psychology Review, 10, 53–100. doi:10.1007/s10567-006-0016-y

Rutter, M., Moffitt, T. E., & Caspi, A. (2006). Gene-environment interplay and psychopathology: Multiple varieties but real effects. Journal of Child Psychology and Psychiatry, 47, 226 –261. doi:10.1111/j.1469- 7610.2005.01557.x

Ryan, S. M., Jorm, A. F., & Lubman, D. I. (2010). Parenting factors associated with reduced adolescent alcohol use: A systematic review of longitudinal studies. Australian and New Zealand Journal of Psychiatry, 44, 774 –783. doi:10.1080/00048674.2010.501759

Schafer, J. L., & Graham, J. W. (2002). Missing data: Our view of the state of the art. Psychological Methods, 7, 147–177. doi:10.1037/1082- 989X.7.2.147

Schulenberg, J. E., Bryant, A. L., & O’Malley, P. M. (2004). Taking hold of some kind of life: How developmental tasks relate to trajectories of well-being during the transition to adulthood. Development and Psycho- pathology, 16, 1119 –1140. doi:10.1017/S0954579404040167

Schulenberg, J., O’Malley, P. M., Bachman, J. G., Wadsworth, K. N., & Johnston, L. D. (1996). Getting drunk and growing up: Trajectories of frequent binge drinking during the transition to young adulthood. Jour- nal of Studies on Alcohol, 57, 289 –304.

Schulenberg, J. E., & Maggs, J. L. (2002). A developmental perspective on alcohol use and heavy drinking during adolescence and the transition to young adulthood. Journal of Studies on Alcohol, 14, 54 –70.

Shaver, P. R., & Brennan, K. A. (1992). Attachment styles and the “Big Five” personality traits: Their connections with each other and with romantic relationship outcomes. Personality and Social Psychology Bul- letin, 18, 536 –545. doi:10.1177/0146167292185003

Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and non- experimental studies: New procedures and recommendations. Psycho- logical Methods, 7, 422– 445. doi:10.1037/1082-989X.7.4.422

Simons, R. L., Stewart, E., Gordon, L. C., Conger, R. D., & Elder, G. H., Jr. (2002). A test of life-course explanations for stability and change in antisocial behavior from adolesence to young adulthood. Criminology, 40, 401– 434. doi:10.1111/j.1745-9125.2002.tb00961.x

Substance Abuse and Mental Health Services Administration. (2010). Center for Behavioral Health Statistics and Quality, national survey on

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

1163GENERAL AND DRUG-SPECIFIC ENVIRONMENTS

http://dx.doi.org/10.1146/annurev.clinpsy.4.022007.141157

http://dx.doi.org/10.1037/0893-164X.19.4.339

http://dx.doi.org/10.1037/0021-843X.114.4.612

http://dx.doi.org/10.1038/nn1583

http://dx.doi.org/10.1016/j.drugalcdep.2006.09.021

http://dx.doi.org/10.1111/j.1530-0277.2007.00583.x

http://dx.doi.org/10.2307/2657324

http://dx.doi.org/10.1007/s10591-005-8243-9

http://dx.doi.org/10.1037/0033-2909.94.1.68

http://dx.doi.org/10.1037/0022-006X.60.5.766

http://dx.doi.org/10.1007/s10519-006-9061-z

http://dx.doi.org/10.1111/j.1360-0443.2008.02178.x

http://dx.doi.org/10.1037/0022-3514.73.5.1092

http://dx.doi.org/10.1037/0033-295X.100.4.674

http://dx.doi.org/10.1017/S0954579400004302

http://dx.doi.org/10.1017/S0954579401002097

http://dx.doi.org/10.1016/0306-4603%2892%2990043-U

http://dx.doi.org/10.1037/0022-0167.49.2.172

http://dx.doi.org/10.1037/0893-3200.13.2.175

http://dx.doi.org/10.1111/j.1741-3729.2003.00064.x

http://dx.doi.org/10.1007/s10567-006-0016-y

http://dx.doi.org/10.1111/j.1469-7610.2005.01557.x

http://dx.doi.org/10.1080/00048674.2010.501759

http://dx.doi.org/10.1037/1082-989X.7.2.147

http://dx.doi.org/10.1017/S0954579404040167

http://dx.doi.org/10.1177/0146167292185003

http://dx.doi.org/10.1037/1082-989X.7.4.422

http://dx.doi.org/10.1111/j.1745-9125.2002.tb00961.x

drug use and health, 2009 and 2010. Retrieved from http:// www.oas.samhsa.gov/NSDUH/2K8NSDUH/tabs/INDEX.PDF

Taylor, J. E., Conard, M. W., O’Byrne, K. K., Haddock, C. K., & Poston, W. S. C. (2004). Saturation of tobacco smoking models and risk of alcohol and tobacco use among adolescents. Journal of Adolescent Health, 35, 190 –196. doi:10.1016/S1054-139X(04)00087-4

Tucker, J. S., Ellickson, P. L., & Klein, D. J. (2003). Predictors of the transition to regular smoking during adolescence and young adulthood. Journal of Adolescent Health, 32, 314 –324. doi:10.1016/S1054- 139X(02)00709-7

Volkow, N. D., & Li, T.-K. (2005). Drugs and alcohol: Treating and

preventing abuse, addiction and their medical consequences. Pharma- cology & Therapeutics, 108, 3–17. doi:10.1016/j.pharmthera.2005 .06.021

Young, S. E., Rhee, S. H., Stallings, M. C., Corley, R. P., & Hewitt, J. K. (2006). Genetic and environmental vulnerabilities underlying adolescent substance use and problem use: General or specific? Behavior Genetics, 36, 603– 615. doi:10.1007/s10519-006-9066-7

Received September 1, 2011 Revision received May 8, 2012

Accepted May 11, 2012 �

Mindfulness, Compassion and Human Development Call for Papers for a Special Section of Developmental Psychology

Editors: Robert W. Roeser and Jacquelynne S. Eccles

A growing body of evidence suggests that training in contemplative practices can facilitate the development of positive human qualities like mindfulness, empathy and compassion. New studies are documenting the neural and psychological mechanisms that underlie these positive human qualities, and some attention has been devoted to the social mechanisms by which they are developed and sustained. Only a handful of empirical studies have explicitly adopted a develop- mental perspective on the use of contemplative practices to develop these qualities and optimize human development across the lifespan, however. The goal of this special section is to showcase empirical research papers that redress this imbalance by focusing on key developmental questions such as:

● What is the normative developmental course of mindfulness and compassion; and how can we validly and reliably measure these constructs across time in children, adolescents and adults? For instance, with regard to mindfulness, when does the ability to become aware of one’s thoughts, feelings, and sensory experiences become possible? What are the developmental manifestations of compassion and how does this construct change over time? Are there periods of relatively greater plasticity in the development of these positive human qualities? Why?

● What are the interpersonal manifestations of mindfulness and compassion in the everyday contexts of human development? For instance, are there mindful and compassionate forms of parenting or teaching? What are the distinguishing features of these forms of socialization? How can we measure the social and behavioral features of mindfulness and compassion in naturalistic settings? Are there more and less age-appropriate ways of teaching mindfulness and compassion during childhood, adolescence and adulthood?

● Can mindfulness and compassion training facilitate the ability of key socialization agents (parents, teachers, mental health professionals) to foster optimal development in children, youth, and young adults, particularly those facing developmental challenges that present unique social-emotional challenges? Is there any evidence that training socialization agents directly provides indirect benefits for the children and adolescents in their care?

Potential contributors should submit a 2-page proposal for such an article by July 1, 2013. The special section editors will then select appropriate proposals and invite submission of full articles, which will then go through the normal review processes for Developmental Psychology. The full articles will be due no later than November 1, 2013. Submit manuscripts using the APA Manuscript Submission Portal: http://www.apa.org/pubs/journals/dev/. Inquiries, including ques- tions about appropriate topics, may be sent electronically to Robert W. Roeser at [email protected] or Jacquelynne S. Eccles at [email protected].

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

1164 EPSTEIN, HILL, BAILEY, AND HAWKINS

http://www.oas.samhsa.gov/NSDUH/2K8NSDUH/tabs/INDEX.PDF

http://dx.doi.org/10.1016/S1054-139X%2804%2900087-4

http://dx.doi.org/10.1016/S1054-139X%2802%2900709-7

http://dx.doi.org/10.1016/j.pharmthera.2005.06.021

http://dx.doi.org/10.1007/s10519-006-9066-7

http://www.apa.org/pubs/journals/dev/