Statistics in marketing and consumer research
Author: Baron, Steve
Source: Journal of Customer Behaviour, Volume 9, Number 3, Autumn 2010 , pp. 229-242(14)
Publisher: Westburn Publishers Ltd
Abstract:There is a long history of using statistical methods in academic marketing and consumer research, with structural equation modelling (SEM) being the currently most popular method. There is a much longer history of statistics as an academic discipline. I am one of several academics who have wondered whether the dominance of SEM and related multivariate analyses in marketing research is a healthy trend. Although undoubtedly some SEM-based studies can lead to illuminating new marketing constructs, the majority seem to offer little more than obvious conclusions, while at the same time presenting deceptively precise and seemingly objective research findings. Yet, it is clear, from a study of the early pioneers of statistical methods and tests, that statistics can be a fascinating and illuminating discipline if treated as an enabler, rather than an unthinking driver, of investigations. This commentary compares the approaches of statisticians who founded regression, correlation, t, F and χ2 tests, with those of many SEM advocates, and questions the value and practice of SEM studies in marketing and consumer research.
Document Type: Commentary
Publication date: September 1, 2010