Random-Parameters Behavioral Models to Investigate Determinants of Perceived Safety in Railway Stations

Table 3

Binomial Logit estimated models.

	MODEL 1		MODEL 2		MODEL 3
Specification	Binomial Logit		Binomial Logit		Binomial Logit
Description	All attributes		Consistent attributes in sign only and interactions with socioeconomic (SE) characteristics		Consistent attributes in sign only and interactions with SE characteristics and travel habits
#observations	1208		1208		1208
Log-likelihood function (constants)	-835.5		−835.5		−835.5
Log-likelihood function (fitted)	-648.1		−636.6		−627.9
Pseudo-R2	0.224		0.238		0.249
Akaike information criterion	1316.3		1297.2		1285.9
Variable	Coeff.	t-ratio	Coeff.	t-ratio	Coeff.	t-ratio
Surveillance cameras	0.46	6.25	0.47	6.35	0.45	6.01
Security personnel	0.89	11.81	1.07	11.09	1.10	11.22
Security personnel, employed			−0.45	−3.22	−0.47	−3.34
Commercial activities	0.08	1.21	0.08	1.10	0.08	1.18
Road crossings	1.12	6.49	1.05	8.56	0.78	4.37
Road crossings, freq. car high					0.49	2.09
Tactile path and signage	−0.09	−0.72
Greenery	0.25	2.18	0.54	3.15	0.56	3.21
Greenery, female			−0.53	−2.29	−0.54	−2.32
Intermodal infrastructure	0.45	3.69	0.46	3.79	0.27	1.84
Intermodal infrastructure, freq. PT high					0.67	2.70
Artificial lighting	0.09	0.73	0.09	0.74	0.09	0.74
Crowding	0.34	3.81	0.07	0.60	0.10	0.84
Crowding, female			0.51	2.94	0.46	2.62
Decorum and maintenance	1.11	8.88	1.12	9.00	1.00	7.36
Decorum and maintenance, freq. rail low					0.72	2.16

The symbols , , and indicate significance at 1%, 5%, and 10% level.