Data are from the 2009 Youth Risk Behavior Survey (high school sample). An association between risk behavior, mood disorder, and asthma morbidity in adolescents and young adults has been identified but is not clearly understood. Risk behaviors, including smoking and substance use, are not only present in young people with asthma, they may increase in the presence of depression and are associated with poor treatment adherence.

The causal hypothesis is that asthma may cause depression which predicts risk behaviors. Other pathways may be involved (but cannot be tested with this data), including depression-induced changes in inflammatory processes resulting in worsening asthma, or the presence of a third, possibly genetic, factor explaining the co-morbidity of psychological and physical illness.

Variables in data file - fexamq5 (N=16,410) Design variables Variable name Primary sampling unit PSU Stratum STRATUM Weight WEIGHT Demographic Variables (Note: create indicator variables for the following) Sex SEX (1=female, 2=male) Grade GRADE (1=9th, 2=10th, 3=11th, 4=12th) Race/ethnicity RACEX (1=White, 2=Black, 3=Hispanic, 4=Other) Other Variables (all coded 0=No, 1=Yes) Ever had asthma (EVASTH) QN86 - Percentage of students who had ever been told by a doctor or nurse that they had asthma Depression (DEP) QN23 - Sad for 2 weeks in the past 12 months Past month any tobacco use (TOBUSE) QNANYTOB - Percentage of students who smoked cigarettes or cigars or used chewing tobacco, snuff, or dip on one or more of the past 30 days Binge drinking (BINGE) QN42 - Five + drinks at one time past 30 days Marijuana use (MJUSE) QN47 - Marijuana use past 30 days Past 30 days substance use (P30SU) - sum (of tobuse,binge,mjuse); recode as 0 none 1 any Test a causal (mediation) model reflecting the hypothesis that past year depression (dep) mediates the relationship between chronic asthma (ever) and recent substance use (p30su).

Examine demographic variables as potential confounders. Do not test for moderation. Support statements with results. Be sure to use the design effects when testing hypotheses. Evaluate and discuss the impact of selection bias, misclassification, and confounding on the results. Write a brief data analysis plan and results section (2-3 paragraphs).