Table 3.

Power of the logistic regression analysis associated with various plain packaging effects on smoking prevalence (estimated value from fitted model highlighted in bold). Pseudo data were generated using immediate and gradual effect models. Power estimates were obtained with 100,000 Monte Carlo repetitions.

Reduction in smoking prevalence	Power
	Immediate effect	Gradual effect
5%	0.09	0.06
10%	0.22	0.09
12.1%	0.30	0.11
15%	0.44	0.15
20%	0.69	0.24
25%	0.88	0.36