. 2012 Oct 4;23(3):458–464. doi: 10.1093/eurpub/cks127

Table 2.

Estimation results of bivariate probit model with selectivity correction and ordinary probit: female sample

Explanatory variables	Bivariate probit with selectivity correction				Ordinary probit
	Participate or not		Meeting the recommended level		Meeting the recommended level
	Coef^a.	ME	Coef^a.	ME	Coef^a.	ME
Degree qualification	0.196***	0.077			0.140**	0.013
Employed			0.198***	0.054	0.217***	0.019
Age	−0.018***	−0.007			−0.010**	−0.001
Ethnicity^b
Mixed	0.105	0.046	−0.179	−0.045
Asian	−0.422***	−0.155	0.422***	0.133
Black	−0.046	−0.016	0.002	0.001
Chinese	0.036	0.015	−0.246	−0.060
Income			0.000**	0.000	0.000**	0.000
No. of children	−0.002	−0.001
Drinkers	0.314***	0.123
Smokers	−0.195***	−0.075
Health status^c
Good health	0.487***	0.181			0.570**	0.039
Fair health	0.280***	0.111			0.231	0.023
Obese	−0.107**	−0.045	−0.163**	−0.042	−0.160	−0.013
Voluntary activity	0.114**	0.046
Urban residence			0.170**	0.044
Seasonal effect^d
Summer	0.203***	0.081	0.154**	0.043	0.254***	0.025
Spring	0.061	0.024	0.017	0.005	0.065	0.006
Autumn	0.059	0.024	−0.002	−0.001	0.032	0.003
Region of residence^e
North east	−0.123	−0.050
North west	−0.214***	−0.089
Yorkshire	−0.133**	−0.054
East Midlands	−0.037	−0.014
West Midlands	−0.095	−0.040
East	0.010	0.000
London	−0.189***	−0.081
South west	−0.018	−0.011
Constant	−0.201*		−1.328***		−2.272***
Observations	7818		3349		7818
Rho	−0.364		−0.364
	P = 0.001		P = 0.001

Coef, Coefficients; ME, Marginal effects.

a: The estimated parameters and asterisks show significance level of 1% (***), 5% (**) and 10% (*).

b: Reference category: white.

c: Reference category: poor health.

d: Reference category: winter.

e: Reference category: south east.

Hausman specification test (null hypothesis: difference in coefficients in two models is not systematic): Chi² (7) = 107.32; P = 0.001.