Sensory discrimination testing methodology selection based on beverage complexity

Discrimination Testing;Statistical Power;Beverages;Complexity;Dimensionality

Bloom, David J

关键词: Discrimination Testing;    Statistical Power;    Beverages;    Complexity;    Dimensionality;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/89180/BLOOM-DISSERTATION-2015.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Sensory discrimination testing is a vital tool used by sensory professionals within the food industry.While many testing methods are available for selection, various methods have proven to differ in power. By utilizing a more powerful method, sensory professionals can limit the resources needed for testing and increase their ability to find significant differences between products.The inherent variability of food and beverage systems in addition to multidimensional changes often made during reformulation of food products can make the task of method selection challenging. The goal of this study was to determine the optimal sensory discrimination methodologies for use with multidimensional beverage systems with confusable difference comparison. Complex beverage systems with product variations relevant to actual testing within the food industry were utilized to compare experimental results to those expected from theoretical modeling predictions.The findings from this study led to the exploration of Researcher-Designated and Panelist-Articulated specific discrimination methods to determine which procedure resulted in greater power for use in comparison of complex beverage systems.Once a more powerful specific testing procedure was established, it was then used in product comparisons with non-specific methods to determine how the degree of difference between samples influenced complex beverage systems in one-dimensional or multidimensional formulation changes. A model beverage system was created which could be altered in formulation to create multidimensional changes between samples and alter d’. The model beverage was utilized to study the impact of sample dimensionality and d’ for both specified and non-specified methods. Differences in the proportion of correct response, method power, and overdispersion were observed between methods when tested with the model beverage system. The findings from these studies emphasize the need for sensory professional to look beyond the categorization of methods by specified versus non-specified.Findings suggest methods should instead be fit to the range of d’ and multidimensional makeup relevant to the samples used in testing.