Principal component analysis on the sensory and instrumental data (Fig. 4) showed that the first and second principal components explained, respectively, 48.5% and 25.9% of the observed variation (74.4% in total). The significance level of Bartlett’s test of sphericity was 0.000, which indicates that the correlation matrix is not an identity matrix and that principal component analysis can be ap- plied to the data (Hair et al., 1998).
All sensory attributes (hardness, crispness, adhesiveness, fracturability and chewiness) and the instrumental forces with a guillotine and with a ‘‘V’’ shape contributed to explain the variance of principal component 1, while compression, puncture and shear tests explained the variance of principal component 2. These re- sults show that instrumental forces derived from the cut tests present stronger correlation with sensory attributes, as seen in Table 4.
Also notice that puncture test, as seen in Table 4, is negatively correlated with sensory attributes, because it is positioned in the quadrant opposite to them.