Statistical Designs and Panel Training/Experience for Sensory Analysis

Samples from six experimental treatments (in a factorial arrangement) were presented to a three-member, trained-experienced panel and an eight-member semi-trained panel in either a randomized complete block (RCB) design or a balanced incomplete block (BIB) design. The experiment was repeated three times using chicken, turkey, and poultry frankfurters representing three levels of variation in experimental units. Flavor and textural characteristics were evaluated. Neither design was consistently more advantageous than the other; however, residual error mean squares for textural characteristics scored by the trained-experienced panel were lower for the BIB design. The three-member, trained-experienced panel had lower residual error mean squares than did the semi-trained panel.