Examining Methods to Test Factor Patterns for Concordance

Publisher Website
Google Scholar
Abstract
To test whether the individual factor patterns of panelists and the panel factor pattern agreed, five grades of tea of the same type, ranging from low to high quality were studied. When a likelihood‐ratio (LR) test and cluster analysis (CA) were applied to the 55‐term correlation matrices of the seven individual panelists, according to the LR test only two of the matrices were homogeneous, but the CA test grouped all the matrices into the same cluster. The study demonstrated that the correlation matrices of the individual panelists or factor patterns themselves should be examined for agreement among panelists when factor analysis is used.