Estimation Methods for Detecting Non-Additive Effects in Structural Equation Models: An Exhibition of the Two-Stage Least Squares Method in Consumer Service Evaluation Networks
The Marketing Review, Vol. 10, No. 2, pp. 135-147
12 Pages Posted: 3 Aug 2009 Last revised: 20 Jul 2011
Date Written: August 2, 2009
Marketers call for the examination of more complex functional structures of the links between service evaluation constructs. However, it seems that including and estimating non-additive effects in the structural equation modeling framework is a daunting task. Therefore, marketers need to be acquainted with the use of these methodologies in a concise and simple manner. Building from Joreskog (1998) who suggest using Bollen’s (1995) estimation method, the present study reviews the relevant literature and exhibits the use of Bollen’s (1995) two-stage least squares estimation method in the context of a consumer service evaluation network.
Keywords: Curvilinear Effects, Two-Stage Least Squares, Latent Variables, Service Quality
JEL Classification: C13, M31
Suggested Citation: Suggested Citation