Problem/Objective: This study is to assess the influences of home smoking restrictions on adolescent smoking behaviors, by taking into account of demographic characteristics, parental socioeconomic status, parental smoking, and unobserved family environment.

Data: Young respondents (aged 15-18) in Tobacco Use Supplement of Current Population Surveys (TUS-CPS) of 1995-6, 1998-9, 2001-2, and 2003. The CPS is a short-term monthly longitudinal study of a nationally representative sample that contains information on about 240,000 individuals from about 50,000 households within the given survey period. All permanent household members who were 15 or older were interviewed. The TUS-CPS, administered as a supplement of CPS, contained information on smoking behaviors, cessation, attitudes toward smoking in public places.

Methods: Structural Equation model or multilevel analysis: Sibling resemblance model in specific. We used the unique features of CPS household data to construct a sample of sibling pairs by matching young respondents with their siblings. Using sibling model, we can control for unobserved heterogeneity (i.e., shared family environment) and obtain better and more accurate estimates of effects of smoking restrictions in home.

Results: After adjusting demographic and socioeconomic characteristics and unobserved shared family environment, we expect influences of smoking restrictions in home and parental smoking sustains.

Conclusions: Pediatricians and other physicians should encourage parents with young children not to smoke at home.