Table 5

Sensitivity analysis results for regression of smoking prevalence among total population

Control variablesSmoking prevalenceSensitivity test 1
(excluding Canada and Australia)		Sensitivity test 2
(excluding the UK)		Sensitivity test 3
(excluding Australia, Canada and the UK)
Panel A: using R score as price measure
 POS display bans−0.086*
(−2.38)
[−0.015]
{−0.065}		−0.095*
(−2.44)
[−0.017]
{−0.072}		−0.086*
(−2.36)
[−0.015}
{−0.065}		−0.095*
(−2.42)
[−0.017]
{−0.072}
Panel B: using EIU mean price as price measure
 POS display bans−0.093*
(−2.22)
[−0.016]
{−0.072}		−0.107**
(−2.35)
[−0.019]
{−0.082}		−0.093*
(−2.21)
[−0.016]
{−0.071}		−0.107*
(−2.34)
[−0.019]
{−0.082}
 EIU mean price−0.015**
(−2.60)
[−0.003]
{−0.063}		−0.018**
(−3.13)
[−0.003]
{−0.073}		−0.015**
(−2.64)
[−0.003]
{−0.064}		−0.018***
(−3.19)
[−0.003]
{−0.074}
Panel C: using MPOWER price for the most popular brand as price measure
 POS display bans−0.077*
(−2.13)
[−0.014]
{−0.058}		−0.085*
(−2.27)
[−0.015]
{−0.064}		−0.077*
(−2.11)
[−0.014]
{−0.058}		−0.084*
(−2.26)
[−0.015]
{−0.064}
 MPOWER price for the most popular brand−0.024†
(−1.71)
[−0.004]
{−0.073}		−0.034**
(−2.57)
[−0.006]
{−0.103}		−0.024†
(−1.72)
[−0.004]
{−0.072}		−0.035**
(−2.59)
[−0.006]
{−0.103}

  • t-Statistics are in parentheses. SEs were clustered at the country level. Marginal effects are in square brackets. Per cent reduction in smoking prevalence and elasticities are in curly brackets.

  • *P<0.05, **P<0.01, ***P<0.001, †P<0.1. 

  • EIU, Economist Intelligence Unit; POS, point of sale.