The 2007 sample includes data from 2,931 students who self-identified as Black only. A more detailed description of the sampling methods and study procedures is published elsewhere (Brenner et al., 2004). Sample weights were included in analyses to adjust for survey nonresponse and sample selection probabilities. Primary sampling units (PSU) and stratum variables were included overnight delivery to account for the complex sampling design. All missing data were handled using the full information maximum likelihood (FIML) capabilities of Mplus. Measures The 2007 YRBS included 95 items covering a range of domains including substance use, mental health and other health-related behaviors. Tobacco use was measured by five questions.

Respondents indicated whether they had ever smoked a cigarette (even a puff or two), age first smoked a whole cigarette, number of days smoked in the past 30 days, number of cigarettes smoked per day in the past 30 days, and whether or not they had ever smoked everyday for 30 days. To measure suicidality, the following items were used: whether or not the respondent had ever felt sad/hopeless for 2 weeks, whether they had seriously thought about ending their own life, whether they had made plans to end their own life, and number of times they attempted to end their own life. Finally, gender and educational level were included as demographic covariates. Analysis Plan LCA was used to identify homogenous subgroups of youth based first on response patterns to smoking behaviors and second on indicators of suicidality, controlling for age and gender in each of these models.

After determining the appropriate number of classes for each set of behaviors, a final model was run in which the association between smoking class membership and suicidal behavior class membership was assessed while controlling for age and gender effects on each through multinomial logistic regression (Lewinsohn, Rohde, Seeley, & Baldwin, 2001). All analyses were conducted using Mplus 5.21 (Lubke & Muth��n, 2005; McCutcheon, 1987) and incorporated sample weights, stratum, and cluster variables to account for the complex sampling design of the YRBS. To determine the number of classes for both smoking and suicidal behaviors, an initial series of models were conducted. For each, an initial one-class (no covariates) model was assessed followed by a series of models with covariates specifying increased number of classes (e.g., two-class, three-class, etc.) representing different patterns of tobacco use behavior. Similar procedures were also Carfilzomib implemented for suicidality indicators.