5. Process Improvement
5.6. Case Studies
5.6.2. Sonoluminescent Light Intensity Case Study

5.6.2.7.

Important Factors: Half-Normal Probability Plot

Purpose

The half-normal probability plot is used to distinguish between important and unimportant effects.

Sample Half-Normal Probability Plot

Conclusions from the Half-Normal Probability Plot

We can make the following conclusions from the half-normal probability plot.

1. The points in the plot divide into two clear clusters:
   • An upper cluster (|effect| > 60).
   • A lower cluster (|effect| < 20).

2. The upper cluster contains the effects:
   X2, X7, X1*X3 (and confounding), X1, X3, X2*X3 (and confounding), X1*X2 (and confounding)

   These effects should definitely be considered important.

3. The remaining effects lie on a line and form a lower cluster. These effects are declared relatively unimportant.

4. The effect id's and the confounding structure are given on the far right (e.g., 13:13+27+46).