Abstract:

A consumer product, such as a cookie, produces multidimensional sensory effects when it is eaten. In order to produce sensorially consistent cookies, it is necessary to develop and use methods that account for multidimensionality. For instance, a consumer may notice how dry or how hard a cookie is or how much moisture it absorbs when chewed. Some attributes may be independent of others; some may be interrelated. For example, we expect dryness and hardness to be related. In this report we show how to set multivariate specifications and how to establish control charts for sensory measures of food and beverage products. The ideas discussed have numerous applications in the analysis of multivariate attribute data.