A Network-Based Approach to Modeling and Predicting Product Coconsideration Relations

Table 4

Estimated coefficients and odds ratios of the dyadic model and ERGM.

Input variables	Dyadic Model		ERGM
	Est. coef.	Odds	Est. coef.	Odds
Network configurations of product interdependence
Edge/Intercept		0.00		0.00
Star effect (inverse measure)				7.20
Triangle effect				2.01
Baseline effects of vehicle attributes
Import		1.45		1.11
Price ()		0.98		1.01
Power ()		1.97		1.42
Fuel consumption (per 100 BHP)		1.21		1.13
Homophily effects of vehicle attribute matching and difference
Market segment matching		3.98		1.94
Make origin matching		3.60		1.69
Price difference ()		0.17		0.45
Power difference ()	0.08	1.09	0.13	1.14
Fuel consumption difference		0.92		0.93
Homophily effects of customer association
Distance of customers’ perceived characteristics.		0.66		0.74
Distance of customers’ demographics		0.56		0.69
Model performance
Null deviance	104,618
Bayesian Information Criterion (BIC)	16,005		14,021

Note. -value < 0.01, -value < 0.001.