Statistical approaches in surface finishing. Part 3. Design-of-experiments

J. R. Smith, C. Larson
  • Transactions of the IMF, November 2019, Taylor & Francis
  • DOI: 10.1080/00202967.2019.1673530

Design-of-experiments: A how-to guide

Photo by Carlos Muza on Unsplash

Photo by Carlos Muza on Unsplash

What is it about?

This paper is the third and final part of a short series of articles aimed towards describing some of the various statistical methods and approaches that have been used in surface finishing. Here, various types of Design-of-Experiments (DOE) techniques are described, such as full and fractional factorial design, central composite design, response surface methodology and Taguchi methods. Their application in the design of surface finishing related experiments are then reviewed, ending with a worked example of a two-level 2^4 full factorial DOE.

Why is it important?

Design-of-experiments is an excellent approach when you have many different variables that control the outcome of a process/experiment. It provides a method such that only a limited number of experiments need to be performed rather than needing to change each variable at once, thereby saving time and money.


Dr James R Smith
University of Portsmouth

Free downloads of the post-print versions of the Part 1 (parametric stats), Part 2 (non-parametric stats) and Part 3 (design-of-experiments) are available from the Resources section (right-hand-side panel). Links to screenshot images showing how to perform the tests in SPSS are also shown in the Resources section (right-hand-side panel).

Read Publication

The following have contributed to this page: Dr James R Smith