Skip to main content
. 2009 Mar 5;11(2):190–196. doi: 10.1093/ntr/ntn019

Table 3.

Multivariate logistic regression models comparing independent associates of moderate to heavy smoking with light daily smoking and with intermittent smoking status

Moderate to heavy daily versus light daily smokers (n = 1,661)
Moderate to heavy daily versus intermittent smokers (n = 1,458)
Respondent characteristic Odds ratio (95% CI) Odds ratio (95% CI)
Survey year
    2003 1.00 1.00
    2005 1.02 (0.76–1.37) .80 (0.55–1.17)
Gender
    Male 1.00 1.00
    Female 1.81 (1.30–2.52) 1.47 (1.08–2.01)
Age (years)
    18–34 1.00 1.00
    35–49 .66 (0.44–1.00) .49 (0.32–0.73)
    50–64 .53 (0.33–0.84) .46 (0.28–0.76)
    65+ .87 (0.49–1.53) 1.29 (0.68–2.41)
Race/ethnicity
    Non-Hispanic White 1.00 1.00
    Non-Hispanic Black 2.60 (1.59–4.25) 3.37 (1.79–6.37)
    Hispanic 6.63 (3.80–11.56) 11.59 (6.67–20.13)
    Non-Hispanic other 2.68 (1.35–5.33) 1.72 (0.85–3.48)
Education
    Less than high school 1.00 1.00
    High school graduate 1.53 (0.96–2.42) 1.59 (0.86–2.92)
    Some college 1.78 (1.05–3.02) 2.59 (1.48–4.54)
    College graduate 2.25 (1.27–3.98) 3.91 (2.13–7.18)
Income
    <US$25,000 1.00 1.00
    $25,000 to <$50,000 0.86 (0.57–1.30) 1.14(0.70–1.83)
    $50,000 to <$75,000 0.95 (0.56–1.59) 1.42 (0.87–2.33)
    $75,000 and greater 1.53 (0.92–2.54) 2.16 (1.30–3.59)
Insurance status
    No health insurance 1.00 1.00
    Has health insurance 1.10 (0.78–1.56) .94 (0.62–1.43)
Trust cancer information from doctor
    Not a lota 1.00 1.00
    A lot .85 (0.61–1.17) 1.48 (1.03–2.08)
Mean hours of television viewing on a typical weekdayb .92 (0.87–0.97)
Health/cancer information seeking on Internet
    No 1.00
    Yes 1.01 (0.67–1.52)
Intention to quitb
    No 1.00
    Yes 1.51 (1.05–2.18)
a

The following responses were included in the “not a lot” category to reflect level of trust rating: “some,” “a little,” and “not at all.”

b

These items were not included in this model because they were not significantly associated with smoking status at the bivariate level.