Binomial generalised estimating equation models, choice between legal, illegal and RYO cigarettes
| Binomial model 1 | Binomial model 2 | Binomial model 3 | Binomial model 4 | |
|---|---|---|---|---|
| Variable | RYO=1, others=0 | RYO=1, legal=0, illegal=. | RYO=1, illegal=0, legal=. | Illegal=1, legal=0, RYO==. |
| Price ratio variables | ||||
| Price manufactured-legal/price RYO | 1.069** (1.01 to 1.132) | – | – | – |
| Price manufactured-legal/price RYO | – | 1.095** (1.016 to 1.18) | – | – |
| Price manufactured-illegal/price RYO | – | – | 1.052 (0.905 to 1.223) | – |
| Price manufactured-legal/illegal | – | – | – | 1.024 (0.96 to 1.092) |
| Price ratio elasticity | 0.344*** (0.096 to 0.591) | 0.459*** (0.193 to 0.725) | 0.065 (−0.061 to 0.191) | 0.050 (−0.097 to 0.196) |
| Policy exposure index | ||||
| Antismoking | 0.899** (0.821 to 0.984) | 0.913** (0.836 to 0.997) | 0.898 (0.765 to 1.055) | 1.089 (0.974 to 1.218) |
| Warning labels | 0.979 (0.76 to 1.261) | 0.984 (0.764 to 1.267) | 0.993 (0.694 to 1.421) | 0.943 (0.718 to 1.238) |
| Marketing | 1.064 (0.959 to 1.18) | 1.051 (0.942 to 1.173) | 1.027 (0.876 to 1.204) | 0.981 (0.86 to 1.119) |
| Number of observations | 4197 | 3726 | 994 | 3674 |
| Demographic variables | ||||
| Gender: female | 1 | 1 | 1 | 1 |
| Gender: male | 4.145*** (2.953 to 5.818) | 3.826*** (2.683 to 5.455) | 5.564*** (3.545 to 8.733) | 0.793 (0.566 to 1.111) |
| Education: <middle school | 1 | 1 | 1 | 1 |
| Education: middle school | 0.509*** (0.311 to 0.832) | 0.459*** (0.282 to 0.746) | 0.832 (0.416 to 1.665) | 0.465*** (0.295 to 0.733) |
| Education: high school | 0.477*** (0.292 to 0.779) | 0.414*** (0.257 to 0.668) | 0.841 (0.456 to 1.55) | 0.441*** (0.284 to 0.684) |
| Education: >high school | 0.219*** (0.113 to 0.424) | 0.200*** (0.103 to 0.387) | 0.526 (0.164 to 1.692) | 0.290*** (0.148 to 0.568) |
| Age 18–24 | 1 | 1 | 1 | 1 |
| Age 25–39 | 1.409 (0.787 to 2.522) | 1.325 (0.719 to 2.442) | 0.887 (0.41 to 1.92) | 1.693** (1.035 to 2.769) |
| Age 40–54 | 1.775* (0.952 to 3.31) | 1.767* (0.94 to 3.321) | 0.831 (0.393 to 1.757) | 3.016*** (1.718 to 5.293) |
| Age 55+ | 2.254** (1.164 to 4.363) | 2.208** (1.11 to 4.393) | 0.844 (0.373 to 1.907) | 2.776*** (1.417 to 5.437) |
| Income: $U0–$U4500 | 1 | 1 | 1 | 1 |
| Income: $U4501–$U8000 | 0.719 (0.435 to 1.188) | 0.690 (0.424 to 1.122) | 0.857 (0.447 to 1.643) | 0.504*** (0.318 to 0.799) |
| Income: $U8001–$U15 000 | 0.526** (0.302 to 0.916) | 0.474*** (0.273 to 0.822) | 0.689 (0.315 to 1.509) | 0.497*** (0.3 to 0.824) |
| Income: >$U15 000 | 0.321*** (0.179 to 0.575) | 0.288*** (0.165 to 0.503) | 0.705 (0.315 to 1.578) | 0.263*** (0.159 to 0.435) |
| Income: missing | 0.656 (0.321 to 1.339) | 0.635 (0.3 to 1.343) | 0.750 (0.358 to 1.57) | 0.510** (0.262 to 0.991) |
| Marital: married | 1 | 1 | 1 | 1 |
| Marital: Separated/divorced/widowed | 0.652* (0.397 to 1.071) | 0.674 (0.392 to 1.148) | 0.661 (0.361 to 1.209) | 1.167 (0.767 to 1.775) |
| Marital: domestic partnership | 1.425 (0.817 to 2.486) | 1.491 (0.858 to 2.591) | 1.247 (0.687 to 2.263) | 1.467* (0.987 to 2.18) |
| Marital: single | 0.992 (0.579 to 1.701) | 1.030 (0.598 to 1.773) | 1.081 (0.554 to 2.109) | 1.029 (0.619 to 1.71) |
| # of observations | 4197 | 3726 | 994 | 3674 |
Note: all regressions are weighted using sample weights and also contain control Montevideo dummy and dummies for years 2008–2009 (wave2), 2010–2011 (wave3) and 2012 (wave 4), and indicator for residents from Montevideo. Robust SEs clustered at the segment level are obtained.
*p≤0.1, **p≤0.05, ***p≤0.01. ORs and elasticities are reported and 95% CIs are in parentheses. OR=exp(b), b is the coefficient. Elasticity=the average of d(log y)/d(log x).
RYO, roll-your-own.