line with a main objective. This question may focus on understanding the impact of one or more factors on specific product characteristics (instrumental techniques in formulation development). These factors could be related to formulation components, processing parameters, or process time. Sometimes, the objective is to optimize formulations or processes to reduce costs or improve product quality (principles of food formulation optimization).
Experimental Design
Experimental design refers to the method of conducting experiments, minimizing research costs, and achieving targeted outcomes based on statistical theories. For instance, when studying the impact of multiple factors on product characteristics, there are various possible combinations of factor levels on the properties. The selection of each factor level should be based on a statistical plan. Proper experimental design ensures the reproducibility of experiments and results while facilitating the application of analysis of variance (ANOVA) techniques in data analysis.
Effect of a Single Factor
When the goal is to examine the impact of a single factor on product characteristics or a target response, a single-factor method is used. In this approach, the researcher determines the levels of the factor. The analysis of this type of experiment can be conducted using one-way ANOVA in statistical software such as Minitab (software skills for food industry professionals).
Effect of Multiple Factors
When studying the impact of multiple factors, their interactions must also be considered. One factor may suppress or amplify the effect of another. In such cases, different experimental designs can be implemented based on research objectives.
Factorial Design
A completely randomized factorial design is suitable when all possible conditions must be examined. This design includes all possible combinations of different factor levels. For example, if Factor A has three levels, Factor B has two levels, and Factor C has three levels, then the total number of experimental conditions would be
3×2×3=18
However, a minimum number of repetitions must be considered, which depends on the error degrees of freedom, which must be lower than 10. Although factorial designs provide a comprehensive analysis of variable effects, they increase research costs due to the large number of treatments. ANOVA allows for evaluating individual factor effects and the interactions. The regression model derived from this design can be used for formulation or process optimization.
Response Surface Designs
Unlike factorial designs, response surface designs use a geometric model to select key points from all possible conditions while incorporating central point errors, thus reducing the number of required treatments. Common response surface designs include the Central Composite and Box–Behnken designs. The geometric settings of these designs influence accuracy and experimental range. The analysis involves model fitting, evaluating significance, and assessing factor effects and interactions. Design Expert software is a suitable tool for experiment planning and result analysis. The model derived from this design can also be used for formulation or process optimization.
Mixture Designs
In some formulation optimization studies, the total sum of the examined factor levels must remain constant. For instance, when optimizing an emulsifier formulation, a combination of three emulsifiers may be used, and their total percentage must always equal 100%. In such cases, increasing one factor necessarily decreases others. Mixture designs are appropriate for such conditions. The analysis involves model fitting, evaluating factor effects and interactions, and optimizing formulations based on the developed model.
