Evaluating kaizen event performance using data envelopment analysis for interval and ordinal data

This paper proposes the analysis of Kaizen events using Data Envelopment Analysis with special consideration to perceptual performance variables. Kaizen events are used widely to achieve enterprise process improvement. This study continues the empirical investigation of Kaizen event performance findings from a multi-site field study through a specialized application of Data Envelopment Analysis (DEA). More commonly, DEA is used to analyze quantitative, linear objective data. However many process improvement studies, including Kaizen events, require the analysis of perceptual (subjective) data. Perceptual survey data is often viewed as continuous or interval data. However, there has been debate as to whether or not perceptual data should be modeled as interval or ordinal data, particularly when using DEA. This research proposes a modified approach to DEA that considers perceptual data as ordinal data.