Table 1.

Variables predictive of county-level foodservice business review volume.

Variable	(a) Variables selected in shrinkage procedure Coefficient (SE)	Variables significant in (a) Coefficient (SE)
Log of population	1.257 (0.079)^***	1.295 (0.065)^***
Restaurants per 1000 population	0.812 (0.111)^***	0.881 (0.095)^***
Schools per 1000 population	1.373 (1.702)	–
Percent with bachelor degree	0.012 (0.013)	–
Percent with high school degree	0.042 (0.019)^*	0.053 (0.015)^***
Intercept	−13.75 (1.628)^***	−14.889
	R² = 0.85	R² = 0.849
	F_5,125 = 142.8^***	F_3,127 = 238.7^***

Initial model shown in column two included all variables selected via the regularization procedure. A second regression model comprising only the significant variables is given in column three.

p < 0.05;

^**

p < 0.01;

^***

p < 0.001.