Combining Process Variables and Ingredient Components in Mixing Experiments

This paper reviews the use of statistical designs formed by combining simplex-centroid designs in the ingredient or mixture components with factorial arrangements in the process variables. Three data sets are presented from a fish patty experiment in which three fish species (mullet, sheepshead, and croaker) were blended and the patties were prepared according to the settings of three process variables (oven cooking temperature, oven cooking time, and deep fat frying time). The response values are the average overall acceptance rating of the patties and the average texture rating of the patties. Detailed steps in the data analysis are provided to illustrate the techniques that are used and the type of information that is acquired from experiments containing mixture components and process variables.