Skip to main content
. 2017 May 3;20(3):383–387. doi: 10.1093/ntr/ntx095

Table 2.

Multilevel Linear Regression Models Examining Product Age of Initiation Among Sexual and Gender Minority (SGM) and Non-SGM, Current Tobacco Product-Using Young Adultsa

Mean (SE) B p
Cigarettes (n = 684)
Heterosexual 16.05 (.13) ref
Gay or lesbian 15.60 (.34) −0.46 .19
Bisexual 15.97 (.32) −0.08 .80
queer, transgender, or other 16.44 (.49) 0.39 .44
E-Cigarettes (n = 516)
Heterosexual 19.19 (.07) ref
Gay or lesbian 19.35 (.25) 0.15 .55
Bisexual 19.12 (.19) −0.08 .71
Queer, transgender, or other 17.85 (.41) −1.34 .001
Cigars (n = 263)
Heterosexual 17.39 (.17) ref
Gay or lesbian 16.69 (.74) −0.70 .36
Bisexual 17.24 (.60) −0.15 .81
Queer, transgender, or other 16.78 (.69) −0.61 .38
Hookah (n = 536)
Heterosexual 17.70 (.12) ref
Gay or Lesbian 18.48 (.44) 0.78 .08
Bisexual 17.83 (.35) 0.13 .72
Queer, transgender, or other 17.41 (.59) −0.29 .63
Smokeless tobacco (n = 100)
Heterosexual 16.81 (.39) ref
Gay or lesbian 16.76 (1.41) −0.05 .97
Bisexual 13.15 (1.24) −3.66 .004
Queer, transgender, or other 16.70 (1.19) −0.11 .93

SE = standard error; B = unstandardized regression coefficient

aOne regression model is fit for each tobacco product, controlling for age, sex, race/ethnicity, and school type (2-year vs. 4-year). Heterosexual serves as the reference group. Each model includes college attended as a random effect.