Factorial Design for Efficient Experimentation

Abstract

While James Dyson's eponymous company, Dyson Ltd., based in Malmesbury, United Kingdom, is known to make high-quality vacuum cleaners, Dyson might have saved himself and his company a lot of time and money if he had been aware of the factorial design approach to experimentation. In fact, as discussed below, the statement of Dyson on cause and effect is incorrect. It is possible to learn what improves a system by changing more than one input variable at a time. Furthermore, this learning can be done more efficiently and with the additional benefit of acquiring information about input-variable interactions that cannot be revealed in the one-at-a-time changes advocated by Dyson. Unfortunately, he is not alone in his beliefs. Many in industry, government, and elsewhere share such misconceptions. The aim of this article is to present some of the key ideas in factorial designs and demonstrate how this approach offers both greater efficiency and insight than one-at-a-time changes.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 35 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.