Are we there yet? Australian road safety targets and road traffic crash fatalities

BMC Public Health

Table 6 Regression results by jurisdiction for trends in annual fatality rates: 1971-2009 and 1992-2009

1971-2009	Preferred Model	Variable^a	Coefficient	Std. Error^b	t-stat	R ²	SE^c	BG^d	J-B^e
Australia⁽⁺⁾	exponential	α	84.05	3.03	27.75***	0.97	0.078	18.86***	1.71
		x	-0.04	0.002	-27.92***
NSW^{(■ +)}	exponential	α	90.53	5.45	16.62***	0.94	0.125	17.85***	1.28
		x	-0.04	0.00	-16.20***
Vic.⁽⁺⁾	exponential	α	87.58	2.88	30.37***	0.95	0.11	11.32***	0.95
		x	-0.04	0.00	-29.51***
Qld⁽⁺⁾	power	α	662.77	30.73	21.56***	0.97	0.095	9.3***	1.29
		lnx	-86.90	4.05	-21.47***
SA	exponential	α	75.16	2.93	25.68***	0.94	0.103	0.13	1.33
		x	-0.04	0.001	-24.76***
WA	power	α	525.37	45.83	11.46***	0.91	0.128	12.75***	0.92
		lnx	-68.81	6.03	-11.40***
Tas.	power	α	519.00	37.03	14.02***	0.84	0.172	0.13	20.43***
		lnx	-67.96	4.88	-13.94***
ACT	exponential	α	92.75	7.10	13.06***	0.81	0.25	1.85	0.26
		x	-0.05	0.00	-12.76***
NT	exponential	α	59.01	5.52	10.70***	0.73	0.195	0.58	0.30
		x	-0.03	0.002	-10.05***
1992-2009
Australia^(■)	exponential	α	62.6	3.09	20.29***	0.96	0.034	0.05	0.73
		x	-0.03	0.001	-19.58***
NSW	exponential	α	67.76	6.25	10.85***	0.87	0.069	0.05	9.74***
		x	-0.03	0.003	-10.51***
Vic.^(■)	exponential	α	60.92	7.36	8.27***	0.80	0.08	2.19	1.33
		x	-0.03	0.00	-8.00***
Qld^{(■ +)}	power	α	542.62	83.62	6.49***	0.78	0.102	5.86**	1.38
		lnx	-71.09	11.00	-6.46***
SA^(■)	exponential	α	72.4	9.77	7.41***	0.76	0.107	1.2	1.28
		x	-0.04	0.004	-7.18***
WA^(■)	power	α	339.80	74.91	4.54***	0.56	0.108	1.5	0.19
		lnx	-44.39	9.86	-4.50***
Tas.	power	α	223.70	146.77	1.52	0.12	0.212	0.74	1.72
		lnx	-29.12	19.31	-1.51
ACT^(■)	exponential	α	60.22	25.88	2.33**	0.24	0.28	0.03	0.11
		x	-0.03	0.01	-2.27**
NT^(■)	exponential	α	53.15	21.24	2.50**	0.26	0.234	0.5	0.21
		x	-0.02	0.01	-2.35**

Notes: ^a α indicates the natural logarithm of the constant; ^b Std. Error indicates the standard error of the estimator; ^c SE is the standard error of the regression; ^d BG is the Obs* R-squared statistic from the Breusch-Godfrey Lagrange multiplier test for serial correlation (1 lag); ^e J-B is the statistic from the Jarque-Bera test for the normality of the residuals. The preferred model for each jurisdiction is indicated by ^(■) and ⁽⁺⁾ indicates the equation has been estimated using the Newey-West heteroscedasticity and autocorrelation consistent (HAC) covariance estimator. ***, ** and * indicate statistical significance at 1, 5, and 10 percent respectively. Other abbreviations are as defined in the text.

ISSN: 1471-2458