Responsibles: W. Dewé, Collaborateur Scientifique, ULg - B. Boulanger, Chargé de cours adjoint, ULg

Objectives:

The objective is to review different statistical methodologies that are used during the life of an analytical method. Indeed, statistical experimental designs may be used for the development and the optimization of the analytical method. As analytical methods need a calibration for quantification, linear and non-linear regressions are reviewed in details, including diagnostics like residual analysis, lack-of-fit test, … Analysis of variance models are also presented for estimating validation criteria like bias, repeatability, intermediate precision, … Finally, the different statistical approaches that are frequently used to take a decision about the validity of the analytical method, before or after its use in routine are presented. Examples are detailed for illustration purpose.

 

Target audience:

This training is dedicated to the scientists who develop, optimize and validate quantitative analytical methods.

 

Program:

1. Statistical design of experiment

• Box-Behnken Design
• Central Composite Design

2. Calibration

• Linear regression
• Non-linear regression
• Diagnostics

3. Analysis of variance

• Analysis of variance model
• Fixed effect versus random effect
• Variance component estimation

4. Decision making

• Equivalence testing
• Tolerance intervals
• Control charts

5. Examples

 

 

These trainings can be dispatched internally or in your company.

For more information, please contact us info@arlenda.com or use this form