A Weighted Statistical Network Modeling Approach to Product Competition Analysis

Table 2

Estimated coefficients of the 2013 co-consideration network for a weighted network and binary networks.

Input variables	Weighted	Binary 1	Binary 2	Binary 3
Network configuration effect
Sum/intercept	−9.54	−7.24	−8.80	−11.72
Main effect (nodal attributes)
Import	−1.46	−1.00	−1.31	−1.59
Price (log2)	0.27	0.18	0.21	0.25
Power (log2)	0.05	0.07	0.01	0.04
Fuel consumption	−0.03	−0.06	−0.02	−0.01
Brand origin (USA)	1.42	1.17	1.43	1.72
Brand origin (Europe)	1.11	0.82	1.09	1.31
Brand origin (Japan)	0.45	0.46	0.58	0.67
Brand origin (Korean)	0.75	0.69	0.90	1.07
Homophily effect (dyadic attributes)
Market segment matching	1.16	0.76	0.93	1.10
Brand origin matching	0.87	0.82	0.97	1.13
Price difference (log2)	−1.90	−1.35	−1.83	−2.25
Power difference (log2)	−0.06.	0.12	0.04	−0.12
Fuel consumption difference	−0.30	−0.26	−0.38	−0.41

Note: .