Exclusive Content & Downloads from ASQ
The Case Against Normal Plots of Effects
Summary: When analyzing effects in an unreplicated experiment, normal or half-normal plots of the effects (also called Daniel plots) are a popular way to visualize them and to judge which are active. This article discusses the ways in which these plots can be confusing and misleading. There are other methods available that are less subjective, easier to explain, more powerful, and less likely to be misinterpreted. The author recommends against using Daniel plots, even as a supplement to one of these better analyses.
- Topics: Engineering
- Keywords: Daniel Plots, Half-normal plots, Fractional factorial design, Factorial designs, Effects, Unreplicated fractional experiment, Teaching
- Author: Lenth, Russel V.;
- Journal: Journal of Quality Technology