This workshop will discuss the basics of experimental design and multivariate statistics for analysis of food and beverage samples. The focus will be on fundamentals of sound experimental design for collecting data from multiple analytical platforms (GC, LC, MS, ICP, etc.) with large numbers of samples and analysis by multivariate statistical approaches. This 1.5-day workshop will also provide opportunities to discuss real world data sets featuring tips on pitfalls and limitations of common statistical approaches and how to choose appropriate sampling techniques and statistical tools to meet project objectives. Emphasis will be on use of Agilent’s Mass Profiler Professional Statistical analysis package, although other statistical packages may also be discussed. Example data sets will be chosen for discussion and evaluation throughout the workshop.

Specific Topics:

1. Experimental Design Overview—Discussion of experimental design of complex data sets and emphasizing issues associated with numbers of samples required, number of analytical variables measured, and differences between biological and analytical replications.
2. Introduction to One-way Analysis of Variance (ANOVA)—Use of one-way ANOVA to identify statistically significant attributes/variables that can be further analyzed using multivariate statistics, like PCA
3. Introduction to Principal Component Analysis (PCA)—Applications of PCA and exploratory analysis tools to understand relationships among samples and a discussion of limitations of PCA analysis
4. Introduction to Partial Least Squares (PLS) Analysis—Statistical tools for predictive modeling of multifactorial data sets
5. Statistical Sampling Approaches—Sound sampling protocols to ensure representative samples as well as an overview of basic sample preparation techniques.
6. Real World Examples—Throughout the workshop real world data sets will be used to highlight specific points, showing strengths and limitations of the various statistical approaches and highlighting common pitfalls in the application and interpretation of statistical outputs.