The Effect of Different Inputs to Factor Analysis: an Example using Service Quality in UK Branch Banking

Factor analysis has long been used in service quality research to understand the dimensions of the construct. This research reinvestigates this construct using two different methodologies (classical test theory and item response theory) in order to assess the homogeneity of the dimensions across the methodologies in a retail branch banking sample taken from a larger network in the UK. The findings show that the two methodologies give different results. Furthermore the choice of correlation matrix to input into Confirmatory Factor Analysis may be more important than is currently thought in the literature as they give different results in this sample.