‘Average of Normals’ Used as Control of Accuracy, and a Comparison with other Controls

Abstract

Hoffmann's ‘average of normals’ control system has been evaluated. Evaluation is based on theoretical viewpoints and on a comparison with other control principles (reference sample, and the use of two different analytical methods for the same test). Our investigations demonstrate that ‘the average of normals’ is suitable as daily control of systematic errors for a number of tests if each series includes 15–20 sample results or more, and also suitable for periodical check of stability in analytical level. ‘Average of normals’ is not applicable to the evaluation of analytical precision. The method forms an excellent supplement to the reference sample control. A simple method for the determination of the sensitivity of ‘average of normals’ is described.

Key Words:

• Accuracy
• average of normals
• precision
• quality control
• reference sample