Multivariate analyses predicting perceived harm measures with self-reported frequency of exposure measures (N=1449)
| Independent variables | Breathing vapour is harmful to health | Concern about health impact of vapour | Breathing vapour is more harmful compared with breathing smoke |
|---|---|---|---|
| B (95% CI) | B (95% CI) | B (95% CI) | |
| Ad exposure | 0.124 (−0.102 to 0.350) | 0.049 (−0.184 to 0.283) | −0.028 (−0.135 to 0.079) |
| Other media exposure | 0.019 (−0.263 to 0.302) | 0.036 (−0.265 to 0.337) | 0.068 (−0.067 to 0.203) |
| Interpersonal discussion | −0.245* (−0.476 to −0.015) | −0.173 (−0.423 to 0.076) | −0.134* (−0.246 to −0.022) |
| Age (years) | 0.007 (−0.001 to 0.015) | 0.012** (0.004 to 0.020) | 0.001 (−0.002 to 0.004) |
| Sex | |||
| Female | 0.145 (−0.088 to 0.379) | 0.211 (−0.035 to 0.456) | −0.001 (−0.107 to 0.106) |
| Race/ethnicity (non-Hispanic Caucasian is referent) | |||
| African-American | 0.267 (−0.200 to 0.735) | 0.283 (−0.212 to 0.778) | 0.212* (0.014 to 0.411) |
| Hispanic | 0.301 (−0.093 to 0.696) | 0.357 (−0.047 to 0.760) | 0.175 (−0.016 to 0.366) |
| Other | 0.301 (−0.176 to 0.778) | 0.246 (−0.242 to 0.735) | 0.344** (0.097 to 0.590) |
| Education (high school or below is referent) | |||
| Some college | −0.048 (−0.341 to 0.244) | −0.145 (−0.451 to 0.160) | −0.075 (−0.204 to 0.055) |
| College graduate or higher | 0.17 (−0.140 to 0.480) | 0.147 (−0.173 to 0.468) | −0.007 (−0.135 to 0.122) |
| Annual household income (<$25 000 is referent) | |||
| $25 000–$49 999 | 0.097 (−0.308 to 0.502) | 0.029 (−0.399 to 0.457) | −0.043 (−0.223 to 0.136) |
| ≥$50 000 | 0.214 (−0.163 to 0.591) | 0.177 (−0.216 to 0.571) | −0.049 (−0.220 to 0.122) |
| Health status | 0.024 (−0.111 to 0.158) | 0.055 (−0.086 to 0.196) | −0.012 (−0.075 to 0.050) |
| Smoking status (non-smoker is referent) | |||
| Former | −0.487*** (−0.769 to −0.205) | −0.372* (−0.669 to −0.075) | −0.104 (−0.219 to 0.010) |
| Current | −1.119*** (−1.516 to −0.722) | −0.992*** (−1.424 to −0.559) | −0.133 (−0.321 to 0.055) |
| Tried e-cigarettes at least once (never is referent) | |||
| Yes but not in the past 30 days | −0.623** (−1.074 to −0.171) | −0.981*** (−1.450 to −0.511) | −0.297** (−0.521 to −0.073) |
| Yes in the past 30 days | −0.850** (−1.404 to −0.297) | −1.088*** (−1.712 to −0.465) | −0.462** (−0.765 to −0.158) |
| Observed others vaping | −0.113 (−0.396 to 0.171) | −0.047 (−0.362 to 0.269) | −0.032 (−0.196 to 0.132) |
| Constant | 3.467 | 3.343 | 2.298 |
| R2 | 0.125 | 0.129 | 0.080 |
Cell entries are unstandardised coefficients from multivariate regressions adjusting for all variables in the table. Self-reported exposure measures are frequency of exposure on scale with a maximum value of 4.
As suggested by one reviewer, we performed a sensitivity analysis to group news and late-night TV together and analysed information from social media as a separate predictor. The substantive results were identical to what we reported here. Frequency of interpersonal discussion was associated with reduced perceptions that breathing vapour is harmful to health and reduced perceptions of breathing vapour being more harmful than smoke. Frequency of exposure to e-cigarette information from social media was not a significant predictor for the three perceived harm outcomes. As suggested by another reviewer, we also performed a sensitivity analysis to obtain the bootstrapped SEs because of non-normality of the outcome variables and found that the substantive conclusions were very similar to the above analysis with the exception of one additional significant finding—frequency of other media exposure was associated with higher perceived comparative harm of SHV. These sensitivity analyses are available from the authors on request.
*p<0.05, **p<0.01, ***p<0.0005.