Discriminating between measurement scales using non-nested tests and two-stage least squares estimators: The case of market orientation

The use of multi-item scales in marketing research is common practice, with many researcher's borrowing 'off the shelf' scales. Yet, there is little literature which provides researchers with techniques for discriminating between prima facie, similar measures of constructs. This study develops a simple procedure for discriminating between measurement scales, combining non-nested econometric tests with two-stage least squares (2SLS) estimators for latent variables. The authors examine two measures of the construct market orientation, MARKOR and MKTOR, which are central to the marketing discipline and are commonly employed. Results suggest that MKTOR, in general, is superior to MARKOR in explaining variations in measures of business performance. Evidence also suggests that results from non-nested tests based on standard OLS regressions and summated scales may be misleading.