Grading Sewing Operator Skill Using Principal Component Analysis and Ordinal Logistic Regression

In the apparel manufacturing process, productivity and quality are somewhat determined by operator skill level. Predicting worker skill level is very important for effective production operation management. However, the current methods for ranking skill level in the manufacturing industry have been based on the subjective evaluation of managers and have failed both in predicting the operator skill level needed for planning and in encouraging operators to develop new skills for quality and productivity. This article develops a new method for grading sewing worker skill levels that employs updated knowledge from experts involved in training, coaching and managing operations in factories. This approach uses the Delphi method combined with principal component analysis to define and classify six qualitative variables that effect on three aspects of operator skill, including coordination skill, sustaining skill, and tool operating skill. Based on these three variables, ordinal logistic regression is applied to grade skill levels, with a statistically significance result.