An effective strategy for the analysis of response profiles

Abstract

An effective strategy for the analysis and comparison of profiles is proposed and illustrated with case studies. This useful strategy makes effective use of profile data by using all the data in each individual profile to create two statistics that describe the profile behavior: Profile “Level” and Profile “Shape.” The profile Level relates to the area under the profile. The Shape statistic is related to the rate of increase of the profile over time. These characteristics enable practical interpretations regarding the factors studied in the experiment. The method is easy to use requiring only straight line regression and design of experiments analysis procedures. A Modified Principal Component Analysis is recommended as an alternative approach to create profile level and shape statistics when the linearized profile model does not give an adequate description of the data.

Keywords:

• profiles
• response curves
• profile modeling
• profile comparison
• analysis strategy

Acknowledgments

The author is pleased to thank the editor and referees for their helpful comments and suggestions which improved the presentation of this article.

Additional information

Notes on contributors

Ronald D. Snee

Ron D. Snee is Founder of Snee Associates, LLC, a firm dedicated to the successful implementation of process and organizational improvement initiatives, using Quality by Design, Continued Process Verification and other data-based improvement approaches that produce bottom line results. Prior to entering the consulting field he worked at DuPont for 24 years in a variety of assignments including eight years in pharmaceutical development. He also serves as an Adjunct Professor in the Pharmaceutical programs at Temple and Rutgers Universities. Ron is an Honorary Member of American Society for Quality, Fellow of the American Statistical Association and American Association for the Advancement of Science and Academician in the International Academy for Quality. He has coauthored seven books and published more than 330 articles in the quality, improvement and statistics literature. Ron received his BA from Washington and Jefferson College and MS and PhD degrees from Rutgers University.