Skip to main content
. 2018 Dec 4;8(12):e025370. doi: 10.1136/bmjopen-2018-025370

Table 4.

Model testing mediation of the associations between use of RYO cigarettes and quit attempts by weekly spend on smoking (see figure 1)

Coeff. SE P values* Bootstrap 95% CI Effect ratio
Total effect (path c) −0.02 0.006 <0.001
Direct effect (path c′) −0.0003 0.006 0.964
Indirect effect (via mediator) −0.02 0.003 <0.001 (−0.03 to −0.02) 1.0

Analyses are adjusted for age, sex, social grade, region, survey year and daily cigarette consumption.

*P values shown for indirect effects are derived from the Sobel test for consistency with total and direct effects; however, bootstrap 95% CIs provide a more robust indication of significant mediation (see Method for more details).

Coeff, coefficient; RYO, roll-your-own.