An evaluation of the limitations of, and alternatives to, the Co-Plot methodology

The Co-Plot technique has become stereotyped in the form of a series of steps that are automatically followed with little regard for the characteristics of the data and the type of analysis they support. While there are obvious extensions to this type of analysis, the methodology has remained unchanged over many years. Here we argue that the application of the Co-Plot to binary, nominal data, as is done in published papers, is inappropriate. We frame the Co-Plot in the general literature of multivariate analysis in order to show that it is a simplification of standard multivariate methods. We use a recently published example on mapping MBA programmes to show how a similar, although more statistically appropriate, analysis based on multidimensional scaling methods produces very different results.