*3.5. Statistical Analysis*

To confirm the hypothesis resulting from the analysis of the presented results: samples 3, 6 and 10 are not typical goldenrod honeys, and a multivariate statistical analysis was performed. In the first step, principal components analysis (PCA) was used, by which the samples were grouped using the projection of cases onto the factor plane (PC 1 × PC 2) (Figure 5). Samples 3 and 10 clearly differ from the others; also sample 6 can be considered as different, located on the negative side of the PC 1 principal component. Among the honey samples considered typical for goldenrod, two groups can be distinguished: samples 4, 6, 7 and 9 as well as 1, 2 and 5.

A cluster analysis was also performed, taking into account the tested honey parameters (Figure 6). The analysis showed the greatest differences for sample 10, which is confirmed by the highest bond distance. For samples 3 and 6, some similarities were demonstrated with other samples considered typical, 5 and 9, respectively, which may indicate the presence of admixtures to the base honey, which significantly changes its properties. The highest degree of similarity was shown by samples 1, 2, 4, 7 and 8, which was presented earlier with the use of SDS-PAGE protein and polyphenolic HPTLC profiles.

**Figure 6.** Tree diagram based on the average values of tested parameters for goldenrod honey samples (1–10).
