Statistical analyses were performed using the freeware programs
PanelCheck v.1.4.0 (www.PanelCheck.com) and R (ver. 2.13.1,
R Development Core Team) together with SIMCA-P þ Ver. 12 from
Umetrics (Umeå, Sweden). PanelCheck was used to monitor
assessor performance during training and for calculation of sensory
product differences. A three-way analysis of variance (ANOVA)
(effect: products, sensory assessors, replicates) on the sensory data
and a two-way Univariate ANOVA (effects: product, replicates) on
data from the instrumental measurements was applied for calculation
of significant product effects in R. Bonferroni least significant
differences and Pearson's correlation coefficients (r) was additionally
calculated using R. Statistical significances were defined at
p 0.05. Multivariate data analysis in form of principal component
analysis (PCA) on mean data from DA and instrumental measurements
was performed in Simca-Pþ in order to visualize the relationship
between products when analyzed by DA and TPA,
respectively. Data was auto-scaled and the multivariate models
were performed with full cross validation.