Table 2

Regression estimates of the annual effect of smoke-free restaurant policies on restaurant business, 1992 to 1998

Best estimate*Range of uncertainty†
*Regression coefficient expressed as an annual change.
†95% confidence interval for regression coefficient expressed as an annual change.
The unit of observation for the regressions is town-month.
The study period covered January 1992 through December 1998.
The natural logarithm of the dependent variable was used in all models.
Each model included independent variables controlling for quarterly seasonal variation, income, and the adoption status of bordering towns.
Source: Center for Health Economics Research analysis of the Massachusetts Department of Revenue meals tax data.
Model 1: All establishments
Time trend (TIME)3.67%+3.14% to +4.19%
High restrictive restaurant policy
post-adoption time trend (TIMEH)0.25%−1.32% to +1.81%
Other restaurant policy
post-adoption time trend (TIMEO)0.53%−0.20% to +1.26%
Model 2: Alcohol serving establishments
Time trend (TIME)4.95%+3.63% to +6.27%
High restrictive restaurant policy
post-adoption time trend (TIMEH)2.50%−1.48% to +6.48%
Other restaurant policy
post-adoption time trend (TIMEO)−0.65%−2.43% to +1.13%