Respondents without children were coded as ��no.�� Social Variables Dichotomous indicators were created for having anyone else in the household who smokes cigarettes and having visitors smoke inside at least 1 day per week. Based on previous work (Borland et al., 2006), respondents were asked how many of the five adults they spend the most time selleck Belinostat with on a regular basis: (a) smoke cigarettes or (b) do not allow smoking in their own homes. Responses were dichotomized into a majority (��3) versus fewer. All respondents were also asked, ��If your lease said no smoking was allowed in your unit, how hard would it be to keep other household members from smoking in your unit? Visitors?�� (Hennrikus et al., 2003). Those who reported ��somewhat�� or ��very�� hard to either question (versus ��not at all�� hard) were classified as ��hard to ask.

�� Environmental/Community Variables Data collectors recorded (yes/no) whether each unit had a covered front porch and if it was on the ground floor. A dichotomous indicator (yes/no) for smelling cigarette smoke that came from other apartments or the hallway at least once a year (SHS incursions) was created to be comparable to other studies (Hennrikus et al., 2003; Hewett et al., 2007; King et al., 2010). Perceptions of safety were assessed by asking respondents how safe they would feel (4-point Likert-type scale) being out alone in the daytime/nighttime near their building (Robinson, Lawton, Taylor, & Perkins, 2003). Neighborhood cohesion was measured using an existing 5-item Likert scale (e.g.

, people around here are willing to help their neighbors) (Sampson, Raudenbush, & Earls, 1997). The composite scale of averaged items ranged from 1 to 5 with higher scores indicating better cohesion (Cronbach��s alpha = .81). Respondents who identified one or fewer neighborhood issues (4-item scale; e.g., people who do not keep up their property or yards) (Robinson, Lawton, Taylor, & Perkins, 2003) as ��somewhat of a problem�� and none as a ��big problem�� were assigned ��better�� neighborhood condition; all others were assigned ��worse�� condition. Data Analysis The prevalence of each individual, social, and environmental/community factor was compared between smokers and nonsmokers. Smoking status and demographic characteristics were also compared between respondents who supported smoke-free policies and those who did not.

Chi-square tests of association with Rao�CScott adjustment were used for categorical variables and Kruskal�CWallis tests or analysis of variance for continuous variables. Length of stay was natural log-transformed to improve normality; geometric means and 95% confidence intervals are reported for this variable. Associations between independent variables (i.e., demo Dacomitinib graphic, individual, social, and environmental/community factors) and each dependent variable were tested using multiple logistic regression analyses.