3.2.1. Emotional Responses

According to Cochran's Q test results (Table S1), 23 emotional terms were significant, including "adventurous," "satisfied," "active," "calm," "affectionate," "energetic," "enthusiastic," "friendly," "glad," "good," "happy," "interested," "joyful," "loving," "peaceful," "pleased," "pleasant," "bored," "disgusted," "worried," "aggressive," "polite", and "wild." Figure 6 shows the correspondence analysis (CA) and principal coordinate analysis (PCoA) results. The CA displays the relationships between emotional terms obtained based on the CATA method and three chocolate products considering the contextual effect. As shown in Figure 6a, the principal component one (PC1) and principal component two (PC2) were 63.43% and 22.86%, respectively, which explained 86.29% of data variability in total. According to CA results, milk chocolate and white chocolate under PVR and B were found to share similar profiles, which were associated with both neutral and positive emotional descriptors such as "peaceful," "pleasant," "good," "satisfied," "glad," "pleased," and "polite." Milk and white chocolate under NVR were also associated with positive terms, such as "affectionate," "interested," "happy," "loving," "joyful," and "friendly." Dark chocolate had highly distinctive groups of emotional terms under NVR and PVR/B. With regard to dark chocolate under NVR, it was related to ardent descriptors, including "adventurous," "energetic," "wild," "active," and "enthusiastic." In contrast, dark chocolate under PVR and B were related to negative terms, such as "bored," "worried," "disgusted," and "aggressive."

The PCoA results show the relationship between emotional terms and the overall liking scores of the three chocolate products under three different contextual settings (Figure 6b). Only the terms "aggressive," "disgusted," "worried," and "bored" were selected in relation to the lowest mean values (<5.0) for the overall liking of chocolate products under different environments. However, terms such as "pleased," "glad," "good," "loving," "friendly," "peaceful," "pleasant," "affectionate," "satisfied," and "joyful" contributed to higher overall liking scores of chocolate products considering the contextual effect (>5.0). *Foods* **2020**, *9*, x FOR PEER REVIEW 9 of 19

**Figure 3.** Just-About-Right (JAR) frequencies and penalty analysis results <sup>1</sup> regarding milk chocolate attributes under different environments <sup>2</sup> . <sup>1</sup> Penalty analysis was associated with the overall liking scores (9-point hedonic scale). <sup>2</sup> M-PVR: milk chocolate positive VR; M-NVR: milk chocolate negative VR; M-B: milk chocolate sensory booth. **Figure 3.** Just-About-Right (JAR) frequencies and penalty analysis results <sup>1</sup> regarding milk chocolate attributes under different environments <sup>2</sup> . <sup>1</sup> Penalty analysis was associated with the overall liking scores (9-point hedonic scale). <sup>2</sup> M-PVR: milk chocolate positive VR; M-NVR: milk chocolate negative VR; M-B: milk chocolate sensory booth.

*Foods* **2020**, *9*, x FOR PEER REVIEW 10 of 19

**Figure 4.** Just-About-Right (JAR) frequencies and penalty analysis results <sup>1</sup> regarding white chocolate attributes under different environments <sup>2</sup> . <sup>1</sup> Penalty analysis was associated with the overall liking scores (9-point hedonic scale). <sup>2</sup> W-PVR: white chocolate positive VR; W-NVR: white chocolate negative VR; W-B: white chocolate sensory booth. **Figure 4.** Just-About-Right (JAR) frequencies and penalty analysis results <sup>1</sup> regarding white chocolate attributes under different environments <sup>2</sup> . <sup>1</sup> Penalty analysis was associated with the overall liking scores (9-point hedonic scale). <sup>2</sup> W-PVR: white chocolate positive VR; W-NVR: white chocolate negative VR; W-B: white chocolate sensory booth.

*Foods* **2020**, *9*, x FOR PEER REVIEW 11 of 19

**Figure 5.** Just-About-Right (JAR) frequencies and penalty analysis results <sup>1</sup> regarding dark chocolate attributes under different environments <sup>2</sup> . <sup>1</sup> Penalty analysis was associated with the overall liking scores (9-point hedonic scale). <sup>2</sup> D-PVR: dark chocolate positive VR; D-NVR: dark chocolate negative VR; D-B: dark chocolate sensory booth. **Figure 5.** Just-About-Right (JAR) frequencies and penalty analysis results <sup>1</sup> regarding dark chocolate attributes under different environments <sup>2</sup> . <sup>1</sup> Penalty analysis was associated with the overall liking scores (9-point hedonic scale). <sup>2</sup> D-PVR: dark chocolate positive VR; D-NVR: dark chocolate negative VR; D-B: dark chocolate sensory booth.

*Foods* **2020**, *9*, x FOR PEER REVIEW 12 of 19

**Figure 6.** (**a**) Correspondence analysis (CA) of emotional terms for chocolate products tasted under different contextual settings <sup>1</sup> ; (**b**) Principal coordinate analysis (PCoA) of emotional terms regarding the overall liking scores. <sup>1</sup> M-PVR: milk chocolate positive VR; M-NVR: milk chocolate negative VR; M-B: milk chocolate sensory booth; W-PVR: white chocolate positive VR; W-NVR: white chocolate negative VR; W-B: white chocolate sensory booth; D-PVR: dark chocolate positive VR; D-NVR: dark chocolate negative VR; D-B: dark chocolate sensory booth. **Figure 6.** (**a**) Correspondence analysis (CA) of emotional terms for chocolate products tasted under different contextual settings <sup>1</sup> ; (**b**) Principal coordinate analysis (PCoA) of emotional terms regarding the overall liking scores. <sup>1</sup> M-PVR: milk chocolate positive VR; M-NVR: milk chocolate negative VR; M-B: milk chocolate sensory booth; W-PVR: white chocolate positive VR; W-NVR: white chocolate negative VR; W-B: white chocolate sensory booth; D-PVR: dark chocolate positive VR; D-NVR: dark chocolate negative VR; D-B: dark chocolate sensory booth.

3.2.2. Principal Component and Cluster Analyses of the Chocolate Products under Different Environments 3.2.2. Principal Component and Cluster Analyses of the Chocolate Products under Different Environments

The principal component analysis (PCA) and agglomerative hierarchical clustering (HCA) results are shown in Figure 7. PCA biplot visualized the associations between liking scores of ten attributes and the three chocolate products (milk, white, and dark) while considering the contextual effect (Figure 7a). The principal component one (PC1) and principal component two (PC2) were 91.72% and 6.64%, respectively, explaining totally 98.36% of data variability. Liking vectors of most attributes were well linked with the horizontal axis, which was PC1 (squared cosines varied from 0.91 to 0.99, data not shown). The liking vector of cocoa flavor was aligned with the vertical axis, The principal component analysis (PCA) and agglomerative hierarchical clustering (HCA) results are shown in Figure 7. PCA biplot visualized the associations between liking scores of ten attributes and the three chocolate products (milk, white, and dark) while considering the contextual effect (Figure 7a). The principal component one (PC1) and principal component two (PC2) were 91.72% and 6.64%, respectively, explaining totally 98.36% of data variability. Liking vectors of most attributes were well linked with the horizontal axis, which was PC1 (squared cosines varied from 0.91 to 0.99). The liking vector of cocoa flavor was aligned with the vertical axis, which was PC2 (squared cosine

was 0.53). Liking vectors of most attributes, except for cocoa flavor, were close to each other in Figure 7a, indicating their positive association. In addition, the liking vector of cocoa flavor was not associated with the liking vectors of hardness and texture as they were almost orthogonal. In terms of chocolate products, milk chocolate was highly associated with the liking of cocoa flavor under PVR and NVR, and milk chocolate under B was associated with the overall liking as well as the liking of bitterness, sweetness, smoothness, dairy flavor, aftertaste and taste/flavor. In addition, white chocolate was relatively associated with the liking of hardness and texture under PVR and NVR. However, dark chocolate was negatively correlated with the liking of all evaluated attributes regardless of the contextual effects. *Foods* **2020**, *9*, x FOR PEER REVIEW 13 of 19 addition, the liking vector of cocoa flavor was not associated with the liking vectors of hardness and texture as they were almost orthogonal. In terms of chocolate products, milk chocolate was highly associated with the liking of cocoa flavor under PVR and NVR, and milk chocolate under B was associated with the overall liking as well as the liking of bitterness, sweetness, smoothness, dairy flavor, aftertaste and taste/flavor. In addition, white chocolate was relatively associated with the liking of hardness and texture under PVR and NVR. However, dark chocolate was negatively

Figure 7b shows the dendrogram based on AHC for the nine chocolate–environment combinations (3 × 3 factorial design). Three main cluster groups were formed, which were (1) dark chocolate under all environments, (2) milk chocolate under all environments, and (3) white chocolate under all environments. correlated with the liking of all evaluated attributes regardless of the contextual effects. Figure 7b shows the dendrogram based on AHC for the nine chocolate–environment combinations (3 × 3 factorial design). Three main cluster groups were formed, which were (1) dark chocolate under all environments, (2) milk chocolate under all environments, and (3) white chocolate under all environments.

**Figure 7.** (**a**) Principal component analysis (PCA) biplot regarding liking scores <sup>1</sup> of chocolate attributes in different environments <sup>2</sup> ; (**b**) Dendrogram of agglomerative hierarchical clustering (AHC) grouping chocolate products under different environments <sup>2</sup> . <sup>1</sup> Liking scores were based on the 9-point hedonic scale (1 = dislike extremely, 5 = neither like nor dislike, 9 = like extremely) [19]. 2 M-PVR: milk chocolate positive VR; M-NVR: milk chocolate negative VR; M-B: milk chocolate sensory booth; W-PVR: white chocolate positive VR; W-NVR: white chocolate negative VR; W-B: white chocolate sensory booth; D-PVR: dark chocolate positive VR; D-NVR: dark chocolate negative VR; D-B: dark chocolate sensory booth. **Figure 7.** (**a**) Principal component analysis (PCA) biplot regarding liking scores <sup>1</sup> of chocolate attributes in different environments <sup>2</sup> ; (**b**) Dendrogram of agglomerative hierarchical clustering (AHC) grouping chocolate products under different environments <sup>2</sup> . <sup>1</sup> Liking scores were based on the 9-point hedonic scale (1 = dislike extremely, 5 = neither like nor dislike, 9 = like extremely) [19]. <sup>2</sup> M-PVR: milk chocolate positive VR; M-NVR: milk chocolate negative VR; M-B: milk chocolate sensory booth; W-PVR: white chocolate positive VR; W-NVR: white chocolate negative VR; W-B: white chocolate sensory booth; D-PVR: dark chocolate positive VR; D-NVR: dark chocolate negative VR; D-B: dark chocolate sensory booth.
