Mirror, Mirror on the Car, Who's the Fairest of Them by Far? A Case Study in Statistical Engineering

• Activity Number: 62
• Type: Topic Contributed
• Date/Time: Sunday, July 31, 2011 : 4:00 PM to 5:50 PM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Jennifer H. Van Mullekom*+
• Companies: DuPont
• Keywords: Statistical Engineering ; Case Study ; DOE ; Multivariate Statistics

Abstract

The concept of "statistical engineering" has received much press in the many of the industrial statistics and quality journals. A panel discussion on the subject during JSM 2010 led to many attendees acknowledging the prevalence of the application of this concept in the manufacturing, financial, and healthcare industries. This discussion motivated the need for case studies to be communicated to build further awareness of these applications. One such application comes from the automotive paint business at DuPont. A series of statistical techniques was use to relate the appearance of a coated article to human perception. These include DOE, consumer and expert preference studies, principal components analysis, multi-dimensional scaling, Procrustes analysis, and a few more. This series of techniques has then been applied to multiple paint colors and resulted in the submission of a patent application. With respect to business implications, it allows the business to easily classify applications for a particular finish thus reducing testing cost. Ultimately, it may be used to set a new standard for the automotive paint appearance.