Table 3. Association Between Public Support for Tobacco Control Strategies, Current Smoking Behavior, and Demographic Characteristics, Tobacco Behavior and Public Opinion Survey, New York City, Years 2010, 2011, 2012, Combined.
Demographics | Favor Smoke-Free Parks |
Favor Smoke-Free Beaches |
||||
---|---|---|---|---|---|---|
Weighted Prevalence (95% CI) | Multivariable regression |
Weighted Prevalence (95% CI) | Multivariable regression |
|||
Adjusted OR (95% CI) | Adjusted P Value, χ2 | Adjusted OR (95% CI) | Adjusted P Value, χ2 | |||
Smoking | ||||||
Current smoker | 32 (30–34) | 0.39 (0.34-.44) | <.001 | 32 (30–34) | 0.36 (.31-.41) | <.001 |
Nonsmoker | 55 (53–57) | 1 [Reference] | 56 (54–58) | 1 [Reference] | ||
Sex | ||||||
Female | 56 (53–58) | 1.45 (1.23–1.70) | <.001 | 56 (54–58) | 1.43 (1.22–1.69) | <.001 |
Male | 46 (43–48) | 1 [Reference] | 47 (44–50) | 1 [Reference] | ||
Age group, y | ||||||
18–24 | 61 (54–68) | 1.40 (0.98–2.02) | .03 | 57 (50–64) | 1.17 (0.82–1.67) | .048 |
25–44 | 57 (54–61) | 1.30 (1.04–1.62) | 58 (55–62) | 1.33 (1.07–1.66) | ||
45–64 | 49 (46–52) | 1.02 (0.83–1.25) | 50 (47–53) | 1.05 (0.86–1.29) | ||
≥65 | 49 (45–52) | 1 [Reference] | 50 (47–54) | 1 [Reference] | ||
Race/ethnicity | ||||||
Non-Hispanic black | 54 (50–58) | 1.27 (1.04–1.56) | <.001 | 50 (46–54) | 0.94 (0.77–1.16) | <.002 |
Hispanic/Latino | 62 (58–67) | 1.59 (1.24–2.04) | 61 (57–66) | 1.53 (1.22–1.92) | ||
Asian/other | 61 (55–67) | 1.59 (1.17–2.15) | 60 (54–66) | 1.43 (1.07–1.92) | ||
Non-Hispanic white | 47 (44–49) | 1 [Reference] | 50 (48–53) | 1 [Reference] | ||
Language spoken most often at home | ||||||
Other than English | 65 (60–69) | 1.47 (1.13–1.91) | .004 | 62 (57–67) | –a | –a |
English | 50 (48–52) | 1 [Reference] | 51 (49–53) | |||
Education | ||||||
Less than high school | 56 (49–62) | –a | –a | 52 (45–59) | –a | –a |
High school graduate/some college | 52 (49–55) | 50 (47–53) | ||||
College graduate or more | 51 (48–53) | 54 (52–57) |
Abbreviations: CI, confidence interval; OR, odds ratio.
We initially included smoking status, sex, age group, race/ethnicity, language spoken at home, and education in the logistic regression models. We removed terms with P values greater than .10 individually. We found no significant interactions between smoking status and demographic characteristics in the final models.