Table 3.
Multivariable Ordered Probit Regression Assessing E-Cigarette Harm Perceptiona
| Average marginal effects predicting e-cigarettes are more harmfulc | ||||
|---|---|---|---|---|
| OR (95% CI)b | Prob. (%) | 95% CI | p | |
| Ever use of cigarettes | ||||
| No | 1.00 | Reference | ||
| Yes | 0.92 (0.85, 1.01) | −0.9 | (−1.8, 0.1) | .07 |
| Ever use of other tobacco productsd | ||||
| No | 1.00 | Reference | ||
| Yes | 0.93 (0.86, 1.00) | −0.8 | (−1.5, −0.1) | .03 |
| All tobacco products are dangerous | ||||
| Strongly agree | 1.00 | Reference | ||
| Agree | 0.76 (0.70, 0.82) | −2.8 | (−3.5, −2.1) | <.001 |
| Disagree | 0.75 (0.65, 0.88) | −2.8 | (−4.1, −1.5) | <.001 |
| Strongly disagree | 1.06 (0.91, 1.23) | 0.7 | (−1.3, 2.7) | .48 |
| Sex | ||||
| Male | 1.00 | Reference | ||
| Female | 1.27 (1.20, 1.34) | 2.6 | (2.0, 3.2) | <.001 |
| Age | ||||
| ≤ 12 | 1.00 | Reference | ||
| 13 | 1.01 (0.89, 1.14) | 0.1 | (−1.2, 1.4) | .89 |
| 14 | 0.97 (0.87, 1.10) | −0.3 | (−1.4, 0.9) | .67 |
| 15 | 1.08 (0.96, 1.20) | 0.8 | (−0.4, 2.0) | .20 |
| 16 | 1.05 (0.95, 1.17) | 0.5 | (−0.6, 1.7) | .33 |
| ≥ 17 | 1.14 (1.03, 1.26) | 1.5 | (0.4, 2.6) | .01 |
| Race/ethnicity | ||||
| Non-Hispanic White | 1.00 | Reference | ||
| Non-Hispanic Black | 1.24 (1.13, 1.36) | 2.4 | (1.2, 3.6) | <.001 |
| Hispanic | 1.28 (1.17, 1.40) | 2.8 | (1.7, 4.0) | <.001 |
| Non-Hispanic Asian | 1.24 (1.11, 1.38) | 2.4 | (1.0, 3.8) | .001 |
| Other/multiple | 1.11 (0.97, 1.26) | 1.0 | (−0.4, 2.4) | .15 |
| Family member uses tobacco | ||||
| No | 1.00 | Reference | ||
| Yes | 0.88 (0.83, 0.93) | −1.4 | (−1.9, −0.8) | <.001 |
| Parent/guardian discussed not using a tobacco product, past 12 months | ||||
| No | 1.00 | Reference | ||
| Yes | 0.99 (0.93, 1.06) | −0.1 | (−0.8, 0.6) | .80 |
| Cut 1 | 0.36 (0.26, 0.45) | |||
| Cut 2 | 1.67 (1.59, 1.76) | |||
aModel assesses increasing levels of harm perception of e-cigarettes relative to conventional cigarettes. This includes individuals who responded to the question that “electronic cigarettes or e-cigarettes, such as Ruyan or NJOY are (less harmful, equally harmful, or more harmful) than regular cigarettes?” and selected one of those three responses; it excludes individuals who reported that they, “didn’t know enough about these products” to answer and those that had, “never heard of electronic cigarettes or e-cigarettes.”
bOdds ratios (ORs) displayed reflect estimates from the weighted multivariate ordered probit model in which all listed covariates are included, with 95% confidence intervals (CI) calculated by Taylor series linearization to account for the complex survey design.
cAverage marginal effects estimate the average change of altering each covariate (e.g., having used other tobacco products compared to not having used other tobacco products) among the study population, estimating the likelihood that an individual perceived e-cigarettes as more harmful than conventional cigarettes.
dOther tobacco use includes use of cigars, dip, pipes, hookah, snus, or dissolvable tobacco products.