Logistic regression model comparing never-smokers to ex-smokers
| Variable | Odds ratio | 99% Confidence interval |
|---|---|---|
| Age | 1.08 | 1.04–1.11 |
| Gender | 1.23 | 1.06–1.43 |
| Ethnicity (vs Euro-Americans) | ||
| African-American | 0.35 | 0.27–0.45 |
| Hispanic-Americans | 0.76 | 0.61–0.95 |
| Other | 0.80 | 0.63–1.01 |
| Income level | 1.02 | 0.95–1.10 |
| Education level | 0.84 | 0.76–0.94 |
| Marital status | 1.49 | 1.24–1.80 |
| Percentage of friends who smoke | 1.89 | 1.78–2.01 |
| Smoking is socially attractive | 2.19 | 2.06–2.32 |
| Alcohol use | 1.29 | 1.21–1.38 |
| Smokeless tobacco use | 1.94 | 1.61–2.34 |
Dependent variable coded 1 = ex-smoker; 0 = never-smoker. Gender coded 0 = man; 1 = woman. Marital status coded 0 = single; 1 = married. Percentage of friends who smoke coded as 1 = <20%; 2 = 20–49%; 3 = 50–79%; 4 = ⩾80%. Social attractiveness of smoking rated on a 5-point scale from 1 = strongly disagree to 5 = strongly agree. Alcohol use coded as 1 = no use; 2 = drink a few times per year, 3 = drink a few times per month, 4 = drink a few times per week, 5 = drink every day. Smokeless tobacco use coded 0 = non-user; 1 = user.