2.5. Statistical analysis Sensory data was analyzed using a three-way ANOVA including the type of melon, the assessor and the tasting session as fixed factors. The double interaction ‘type of melon x assessor’ was also included in the preliminary analysis but dropped later since it was not significant (p > 0.05) for any of the six texture descript-tors assessed. Physical-chemical data was submitted to a one-way ANOVA (melon type as a fixed effect) over the mean values obtained for each experimental unit (each individual melon). In all cases a Tukey’s HSD post hoc test (p ≤ 0.05) was performed to test the existence of statistical differences between the different melon types. A Pearson’s correlation analysis and a Principal Component Analysis were carried out to monitor the relationship between all the variables measured over the same samples. The univariate and multivariate discriminate ability of the measured variables was evaluated through discriminant analysis. Different confusion matrices were also computed to verify the percentage of correctly