ENBIS Spring Meeting 2018

4 – 6 June 2018; Florence, Italy Abstract submission: 17 November 2017 – 20 April 2018

Planning Experiments to Improve Food Product Quality

6 June 2018, 11:00 – 12:10


Submitted by
Geoff Vining
Geoff Vining (Virginia Tech)
The statistical design and analysis of experiments has been crucial for food quality and production for more than 100 years, dating back at least to Fisher’s work at the Rothamsted Experimental Station. Experimental design and analysis were extremely important to the statistics group at General Foods founded Mavis Carroll in 1955. Many other companies have discovered the need for proper design and analysis of experiments since then.

Until very recently the focus of experimental design and analysis for food quality and production has focused on either single characteristics or a small number of characteristics. The typical analysis has built separate models for each response. Typically, only when trying to define “optimal” operating conditions have analysts considered the joint behavior of the characteristics of interest.

In many cases, the behavior that really controls the quality is a profile over time, for example, the temperature of cookies as they bake over time. The cookie being overheated or underheated at critical portions of the baking process can have a significant effect on subjective evaluations of the final product such as taste.

This talk begins with a brief review of the history of statistical design and analysis of experiments for food quality and production. It then shifts its attention to the analysis of profiles. It outlines some of the critical issues underlying the proper analysis and the challenges that the research community faces. This talk illustrates its basic points in the context of a on-going case study involving shrimp in Peru.

Return to programme