Table A3

The effect of college education on smoking behavior at different ages. Panel data with fixed effects, marginal effects of conditional logit. Males

Birth Cohort	Dependent variable: smoked at age 17 and smoked at ages 25 to 60
Birth Cohort	1910–19	1920–29	1930–39	1940–49	1950–59	1960–69	1970–79
	At age 25
College	0.058 [0.0521]	0.029 [0.0258]	0.032^** [0.0160]	0.007 [0.0158]	−0.009 [0.0167]	−0.033 [0.0309]	0.044 [0.0466]
Observations	4374	7590	7736	10744	10930	5690	1506
	At age 30
College	0.059^** [0.0293]	0.062^** [0.0262]	0.038^** [0.0188]	−0.004 [0.0183]	−0.034^* [0.0200]	−0.003 [0.0299]
Observations	4,790	7,924	8,196	11,230	9,774	4,662
	At age 35
College	0.029 [0.0251]	0.051^** [0.0228]	0.039^** [0.0183]	−0.058 [0.0138]	−0.037^* [0.0194]	−0.064 [0.0418]
Observations	4,838	8,076	8,386	10,078	8,260	2,186
	At age 40
College	0.072^** [0.0306]	0.061^*** [0.0220]	0.039^* [0.0204]	−0.001 [0.0176]	−0.072^*** [0.0230]
Observations	4,786	8,186	8,546	8,530	5,854
	At age 45
College	0.057^* [0.0296]	0.048^** [0.0229]	0.025 [0.0199]	0.014 [0.0200]	−0.067^** [0.0324]
Observations	4,784	8,174	7,830	6,938	2,452
	At age 50
College	0.088^** [0.0336]	0.039^** [0.0188]	0.001 [0.0217]	0.012 [0.0241]
Observations	4746	8188	6638	4942
	At age 55
College	0.072^** [0.0305]	0.033 [0.0229]	−0.008 [0.0243]	−0.018 [0.0389]
Observations	4,758	7,278	5,420	2,012
	At age 60
College	0.068^** [0.0330]	0.025 [0.0262]	−0.016 [0.0257]
Observations	4,756	6,024	4,032

↵Analysis from smoking histories from the 1978, 1979, 1980, 1987, 1988, 1992, 1995, 1997, 1998, 1999, and 2000 National Health Interview Surveys for individuals age 25 and older at the time of the survey. Smoking histories are reconstructed using the age at smoking initiation and cessation. See also Note 5. Panel data regressions with individual fixed effects: The baseline smoking behavior is always taken at age College always takes the value 0 at age 17. At ages 25 and older, it takes the value 1 for college graduates and the value 0 otherwise. Additional controls include age, gender, year of birth, race, and survey year dummies. Robust standard errors in square brackets. ***, **, * denote statistical significance at the 1 percent, 5 percent, and 10 percent confidence levels, respectively.