1. Abstract A longitudinal study designed to follow children of alcohol and drug dependent fathers from adolescence into adulthood RISK began in 1993 and is currently ongoing (the present study is based on Time 1 data collected between 1993-1998 and Time 2 data collected between 1998-2003) Adolescents are recruited directly through the community (e.g., YMCA/YWCA, high schools) and indirectly through their parents (e.g., newspaper advertisements, presentations at alcohol/drug treatment programs) 200 adolescent girls and boys (62% girls; 68 Caucasian) and their fathers Age range = 15-19 years old; Mean age = 16.76 (SD=1.35) Mean household income = 4.69 (SD=2.42) on a scale ranging from 1=0-$9,999/year to 9=$150,000/year or more (equivalent of about $40,000/year) Working class families from an inner-city Sample Negative Affect The fear, hostility, sadness, and guilt scales from The Positive and Negative Affect Schedule (PANAS; Watson, Clark, & Tellegen, 1988) were used to assess negative affect. Each of the PANAS scales includes a list of adjectives that are responded to. For example, the adjectives for the fear scale are: Afraid, scared, frightened, nervous, jittery, and shaky. When completing the PANAS, individuals are asked to respond to each adjective by rating how much they usually feel that way in general, on a scale ranging from 1=“very slightly or not at all” to 5=“extremely.” Previous research has shown that the PANAS scales are reliable indicators of affect (Watson, 1988; Watson & Clark, 1991). In the RISK sample, the Cronbach alpha coefficients for the sadness, fear, guilt, and hostility scales were.83,.84,.86, and.78, respectively. Onset of Substance Use The youth were asked how old they were when they first began drinking regularly and when they first used marijuana. Based on their responses to these questions, they were classified as follows: Age Began Regular DrinkingAge 1 st Used Marijuana 0=have not begun to drink regularly0=never used 1=19-22 years of age 1=19-22 years of age 2=17-18 years of age 2=17-18 years of age 3=15-16 years of age 3=15-16 years of age 4=14 years of age or younger4=13-14 years of age 5=11-12 years of age 6=10 years of age or younger *Paternal alcoholism, paternal depression, and the negative affect components were assessed at Time 1. The substance use measures were assessed at Time 2. Parental Alcoholism and the Onset of Substance Use: An Examination of the Role that Negative Affect Plays Christine McCauley Ohannessian, University of Delaware and Victor Hesselbrock. University of Connecticut School of Medicine Structural equation modeling was used to examine the relations between paternal alcoholism, the negative affect components, and the onset of adolescent substance use. The specified model is presented in Figure 1. Four models were run. Model 1 tested a saturated model (which included non-specified direct paths). Model 2 tested the specified model (this model also included direct paths that were significant in Model 1). Model 3 only included paths that were significant in Model 2. Model 4 was identical to Model 3, except that it allowed for disturbance terms with modification indices greater than 4.0 to be correlated. Most studies that have examined the association between negative affect and substance use have used a general indicator of negative affect. However, certain components of negative affect may be more closely linked to substance use than others. Therefore, the primary goal of this study was to assess whether specific negative affect components play differential roles in the relationship between paternal alcoholism and substance use. Consistent with the literature, hostility was found to play a significant mediating role in the relationship between parental alcoholism and substance use. More specifically, COAs were found to have higher levels of hostility in comparison to non- COAs. In turn, higher levels of hostility predicted earlier marijuana use. None of the other negative affect components were found to significantly mediate the relationship between paternal alcoholism and adolescent substance use. Results from this study suggest that substance abuse prevention programs (particularly those designed for COAs) would be well advised to specifically target hostility, rather than general negative affect. Moreover, they highlight the importance of separately assessing specific components of negative affect when examining the role that negative affect plays in substance use. Conclusions * Supported by NIAAA P50AA03510 and NIAAA K01AA015059 Notes. Standardized regression coefficients are presented. Only significant paths (p<.05) are shown. For presentation purposes, correlations between disturbance terms are not shown. Results Model 1 (the saturated model) did not fit the data well {X 2 (7)=430.08, p=.00; NFI=.85; CFI=.85; RMSEA=.55). The only direct path that was significant was between paternal alcoholism and age at first marijuana use (β=.15, p<.05). This path was retained in subsequent models. Model 2 and Model 3 also did not fit the data well {X 2 (10)=432.76, p=.00; NFI=.85; CFI=.85; RMSEA=.46; X 2 (20)=439.09, p=.00; NFI=.84; CFI=.85; RMSEA=.32; respectively). In contrast, the overall fit of Model 4 fit the data quite well {X 2 (13)=13.15, p=.44; NFI=.99; CFI=1.00; RMSEA=.01). As shown in Figure 2, paternal alcoholism directly predicted age at first marijuana use. In addition, paternal alcoholism predicted the onset of marijuana use indirectly via hostility. Significant indirect paths also were observed between paternal depression and regular drinking. More specifically, paternal depression significantly predicted youth fear; in turn, youth fear significantly predicted age when regular drinking began. Sobel tests were conducted to formally test for mediation. Hostility was found to significantly mediate the relationship between paternal alcoholism and the onset of marijuana use (c.r.=1.98, p<.05). None of the other negative affect components significantly mediated the relationship between paternal alcoholism and adolescent substance use. The RISK Project Analyses This study was designed to separately examine the relations between four fundamental components of negative affect (fear, hostility, sadness, and guilt) and the onset of substance use. The primary aim of the study was to examine the potential mediating roles that these components of negative affect play in the relationship between paternal alcoholism and the initiation of substance use. The sample included 200 15-19 year old adolescents (68% Caucasian; 62% girls; 56% COAs) and their fathers. All participants were followed up five years after the initial testing (Time 2). Results from structural equation modeling (SEM) indicated that higher levels of fear and hostility and lower levels of guilt were associated with earlier substance use initiation. In addition, hostility was found to significantly mediate the relationship between paternal alcoholism and the onset of marijuana use. These results highlight the importance of separately examining different components of negative affect when investigating the role that negative affect plays in adolescent substance use. Paternal Alcoholism The Semi-Structured Assessment for the Genetics of Alcoholism (SSAGA) was administered to obtain lifetime diagnoses of alcohol dependence and depression in the fathers (depression was included as a covariate in the present investigation because it commonly is comorbid with alcohol dependence). The SSAGA is a clinical psychiatric interview with good psychometric properties (kappas range from.74- 1.00). Measures Figure 1 The Specified Model Figure 2 The Final Model Paternal Alcoholism Paternal Depression Fear Hostility Guilt Age First Used Marijuana Age Regular Drinking Began.15** -.14**.21** -.19*.15*.17* Paternal Alcoholism Paternal Depression Fear Hostility Guilt Age First Used Marijuana Age Regular Drinking Began Sadness