Table 4.

Fixed Effects Analysis of the Association between County-Level Characteristics and Individual Health Behaviors, U.S. National Longitudinal Study of Youth, 1992–2012

	β Coefficient [95% CI]
	Smoker Unadjusted	Smoker Adjusted	Daily Cigarettes Unadjusted	Daily Cigarettes Adjusted	Binge Drinking Unadjusted	Binge Drinking Adjusted	Daily Drinks Unadjusted	Daily Drinks Adjusted
County characteristics
Unemployment rate	−0.0036*	− 0.0034*	− 0.056*	− 0.053*	− 0.000045	0.00020	0.032*	0.026*
	[− 0.0060, − 0.0012]	[−0.0058, − 0.00092]	[−0.098, − 0.013]	[−0.098, − 0.0084]	[−0.0030, 0.0029]	[− 0.0029, 0.0033]	[0.0082, 0.055]	[0.0025, 0.050]
Income (per $1000)	−0.00052	−0.00056	− 0.0049	−0.0069	0.00095*	0.00077	0.0062*	0.0042
	[−0.0012, 0.00018]	[− 0.0013, 0.00017]	[− 0.017, 0.0074]	[−0.020, 0.0063]	[0.00008, 0.0018]	[−0.00013, 0.0017]	[0.00079, 0.012]	[−0.0013, 0.0097]
% < high school	− 0.00020	−0.00015	0.0068	0.0086*	− 0.000033	−0.000060	− 0.0020	−0.0015
	[−0.00072, 0.00032]	[−0.00070, 0.00040]	[− 0.0010, 0.015]	[0.00013, 0.017]	[− 0.00059, 0.00052]	[− 0.00065, 0.00053]	[−0.0053, 0.0012]	[− 0.0049, 0.0019]
Ln(Household income)		−0.00023		0.048*		0.0059*		0.050*
		[−0.0024, 0.0019]		[0.0057, 0.091]		[0.0039, 0.0079]		[0.037, 0.064]
Education (ref < HS)
High school		−0.0015		−0.10		−0.0063		0.045
		[−0.025, 0.022]		[−0.57, 0.36]		[−0.029, 0.017]		[−0.16, 0.25]
Some college		0.000067		−0.14		−0.015		0.051
		[−0.027, 0.027]		[−0.68, 0.39]		[−0.042, 0.012]		[−0.17, 0.28]
College or more		0.0013		0.34		−0.0042		0.14
		[−0.035, 0.038]		[−0.34, 1.02]		[−0.043, 0.034]		[−0.14, 0.43]
Married		−0.031*		−0.49		−0.044*		−0.21*
		[−0.047, − 0.015]		[− 0.80, − 0.18]		[−0.065, − 0.022]		[− 0.34, − 0.085]
Ln(Weeks unemployed)		0.0060*		0.12*		0.0017		0.032*
		[0.0021, 0.0099]		[0.046, 0.20]		[−0.0030, 0.0064]		[0.00084, 0.064]
Number of children		−0.0016		0.066		0.0044		0.0094
		[−0.0062, 0.0030]		[−0.018, 0.15]		[−0.0013, 0.010]		[−0.026, 0.045]
No. Observations	42,996	42,996	40,735	40,735	39,402	39,402	44,272	44,272
No. Individuals	9246	9246	9206	9206	8943	8943	9064	9064

*p < 0.05. HS = high school. County-level characteristics included annual unemployment rate, inflation-adjusted per capita personal income, and percent with less than a high school education. Analyses were conducted using multivariable linear regressions with imputed data, including fixed effects at the individual level to adjust for time-invariant individual characteristics. Linear probability models were used for binary outcomes. Additional controls included fixed effects for year