Observations on the use of statistical methods in Food Science and Technology

Statistical methods are important aids to detect trends, explore relationships and draw conclusions from experimental data. However, it is not uncommon to find that many researchers apply statistical tests without first checking whether they are appropriate for the intended application. The aim of this paper is to present some of the more important univariate and bivariate parametric and non-parametric statistical techniques and to highlight their uses based on practical examples in Food Science and Technology. The underlying requirements for use of particular statistical tests, together with their advantages and disadvantages in practical applications are also discussed, such as the need to check for normality and homogeneity of variances prior to the comparison of two or more sample sets in inference tests, correlation and regression analysis.