3 - Statistical methods and tools for analysing sensory food texture

Abstract

A proper understanding of statistical methods involved in the analysis of sensory data is important to ensure that the correct conclusions are made regarding the data. Sensory analysis methods use a range of measurement scales, and the number of assessors used for an experiment may also vary from one to another. These and other factors make it important to use proper procedures for data analysis. This chapter outlines a range of univariate and multivariate statistical methods used in the analysis of sensory analysis data. It will briefly outline each statistical method but will not describe the theory in any great detail. Only the essential points are outlined, and wherever possible, comments are provided to explain certain aspects further. The emphasis will be on the application of appropriate statistical methods and the interpretation of the results. Four examples of typical data from preference tests, difference tests and sensory descriptive analysis are provided to illustrate proper procedures that should be considered when analysing sensory data. Every example is analysed using R, which is a freely available software. Data sets and complete R scripts to do all the analyses described in this chapter can also be downloaded from http://dx.doi.org/10.1016/B978-1-78242-334-8.00003-1.