Table 4.

Binary and Ordered Logistic Regression Odds Ratios, Likelihood That Covariates’ Mean Over-Time Variables Predict Homicide Trajectory Group Membership.

	Binary logistic regression			Ordered logistic
	Model 1	Model 2	Model 3
	Low vs. medium	Medium vs. high	Low vs. high	Model 4
Population size (In)	—	—	—	1.79*
Development indexa	1.62*	—	5.34**	1.63**
Female labor force	0.95	1.13	1.06	1.07*
Percent divorced males	1.38	1.49	2.61**	1.92**
Sex ratio (males/females)	1.19	0.75**	—	0.76*
Percent urban population	—	0.95*	1.02	0.91**
Net migration (immigrant-emigrant)	0.99*	—	1.00**	—
Latin American region	—	57.21**	—	139.45**
n	47	49	48	65
Log pseudolikelihood	−16.24	−22.58	−7.84	−33.72
Pseudo R2	.47	.33	.76	.51

^aIndex includes Gini, GDP per cap, infant mortality, % male ages 15–29, and shadow economy.

^*p < .05.

^**p < .01. (one-tailed tests). Robust standard errors used.