Table 2.
Fitted logistic regression models. Final model highlighted in bold. AIC = Akaike Information Criterion.
| Model | Parameter estimates (standard error) | Reduction in smoking prevalence (95% CI) | Significance | AIC |
|---|---|---|---|---|
| Time (year) | -0.0567 (0.0048) | 5.5% (4.6%, 6.4%) | P < 0.001 | 930.58 |
|
Time (year) Plain packaging |
-0.0530 (0.0054) -0.1287 (0.0895) |
5.5% (4.2%, 6.2%) 12.1% (-4.8%, 26.2%) |
P < 0.001 P = 0.15 |
930.47 |
| Time (year) 25% tax increase |
-0.0605 (0.0074) 0.0478 (0.0697) |
5.9% (4.5%, 7.2%) -4.9% (-20.3%, 8.5%) |
P < 0.001 P = 0.49 |
932.11 |
| Time (year) Smoke-free policies |
-0.0510 (0.0107) -0.0501 (0.0846) |
5.0% (3.0%, 6.9%) 4.9% (-12.3%, 19.4%) |
P < 0.001 P = 0.55 |
932.23 |
| Time (year) Graphic health warnings |
-0.0588 (0.0091) 0.181 (0.0660) |
5.7% (4.0%, 7.4%) -1.8% (-15.9%, 10.5%) |
P < 0.001 P = 0.78 |
932.50 |