Overall interpretation of taste and perception

Holistic by DMRI™ method is applicable to describe products with sensory differences and by comparing the results with a sensory descriptive analysis, it is possible to see which sensory properties that may explain variations in the holistic assessment among the products. The average of common sensory descriptors (X) was related to the average consumer perception of holistic words (Y) by PLSR and describes how HBG, HPP11, HPG1.1 and HPL influence the taste in the four products. The analysis is performed on two consumer groups who tasted either liver pate and wiener sausages (Figure 17) or salami and meatball (Figure 18).

Figure 17: APLS correlation loading for average of sensory descriptors and product matrix in blue (X) and Holistic in red (Y). The two first dimensions showing the difference among HPL, HPP11, HPG1.1 and a reference (REF) in Liver pate (L) together with HPG1.1 and a reference (REF) in Wiener Sausages (W). Full cross validation with all six products. The inner and outer circles indicate 50 and 100 % explained variance, respectively.

The two first principal components describe 66 % of the variance in taste of wiener sausages and liver pates (Figure 17). The horizontal dimension describes novelty going from familiar and traditional to different and strange. Familiar and traditional is associated with rose appearance and spicy odor in the reference of wiener sausages while different and strange is associated with bitter flavor, firmness in

Results and discussion

81

texture and brown appearance in HPL and HPP11 in Liver pate. The vertical dimension describes the quality parameters in which discount and chemical are associated to sour and off flavor, which are negatively correlated to the liver pate reference, which is perceived as exclusive and natural. Liver pate containing HPG1.1 and the perception of appetizing is described with 50 % variance but is mostly correlated to exclusive and natural as well. Salty and spicy flavors show a slight correlation to traditional and familiar and are anti-correlated with boring. The consumers did not perceive boring with significant differences and it could be interpreted that the consumers did not have a general conventional perception of boring and the word can be associated in different way. As the tested products are common consumed Danish meat products, it means that they are well known by the Danish consumers and therefore can be perceived as boring by default. However, if the tasted product is perceived as boring, it could be interpreted, as the product does not differ from the reference taste expected by the consumer. HPG1.1 in wiener sausage can slightly be associated to sour and off flavor, but must be interpreted with caution, as only 50 % of the variance is explained. The result agree with the descriptive results as HPL is significantly different from HPG1.1, HPP11 and the reference in the brown and rosa appearance along with bitter flavor and HPP11 is significantly different from HPG1.1, HPL and the reference in off flavor in liver pate. HPG1.1 in liver pate can be associated to the liver pate reference which is substantiated in the descriptive attributes whereas HPG1.1 did not differ from the reference attributes, except from burnt odor and firmness in texture (Table 18 page 74). The wiener sausages reference was significantly different in the rosa appearance (Table 20 on page 76, wiener).

Results and discussion

82

Figure 18: APLS correlation loading for average of sensory attributes and product matrix in blue (X) and holistic perception in red (Y). The two first dimensions showing the difference among HBG, HPG1.1 and a reference (REF) in Salami (S) together with HPP11, HPG1.1 and a reference (REF) in Meatball (F). Full cross validation with all six products. The inner and outer circles indicate 50 and 100 % explained variance, respectively.

The two first principal components describe 93 % of the variance in the taste of salami and meatball (Figure 18). The holistic words are clearly clustered in two anti-correlated groups. Left group encompass strange, different, chemical, boring and discount which can be associated with off odor, off flavor, off after flavor and bitter flavor in meatball with HPP11 and HPG1.1. These are reversely correlated with familiar, traditional, natural, appetizing and exclusive. The meatball and salami reference seems to be associated to salt flavor, but only explained with 50 % variance. In spite of this the salt flavor are not described with significant difference in the meatball attributes where as it is in the salami reference. The salami reference is anti-correlated with the off and bitter modalities. HBG in salami seems to be particular firm and low in spicy flavor.

The result from PLS shows the relationship between the variations between common sensory descriptors and consumer perception of the products and how different hydrolysates is connected.

When comparing the different product categories, there is a clear difference between hydrolysate and product and how consumers perceive the taste of the products. The clusters of anti-correlated holistic

Results and discussion

83

words is clear in meatball and salami which is perceived with higher significant differences among the consumers (Table 22, page 78) compared with liver pate and wiener sausages (Table 23, page 79). The result could express a tendency in the perception of liver pate and sausages as different among the consumers whereas meatball and salami is perceived more similar. Moreover it could be interpreted as if the hydrolysates are less prominent in the liver pate and wiener sausages as compared to the salami and meatball products. It seems that the liver pate and wiener sausages can be classified by novelty and quality parameters (Figure 17 on page 80). The off and bitter modalities is prominent in products with the hydrolysates compared to the references in the descriptive analysis (Table 18, Table 19, Table 20 & Table 21), whereas they by the consumers in general are associated with strange, different, chemical, boring and discount (Figure 17 and Figure 18).

Results and discussion

84