11-8

A. A. OLABI, Food Science, Cornell University, Stocking Hall, Ithaca, NY 14853, C. E. McCulloch, Biometrics, Cornell University, Warren Hall, Ithaca, NY 14853, D. A. Levitsky, Nutritional Science, Cornell University, Savage Hall, Ithaca, NY 14853, and J. B. Hunter, Agricultural and Biological Engineering, Cornell University, Riley-Robb Hall, Ithaca, NY 14853.

Control foods have been used in taste panels to provide reference samples in the case of quality control studies or descriptive panels. However, the use of control foods in acceptability panels is rare. Control foods can be useful in incomplete designs and can also increase the precision of statistical tests. In addition, control foods can be used to reduce between-session variation in food acceptability scores.

This work attempts to pinpoint possible benefits of control foods in acceptability panels.

Assessments of two sets of foods, which have mostly common items, were conducted with two groups of 20-24 panelists each. At each session, panelists rated the acceptability, on a nine-point hedonic scale, of six individual foods including one control food. The two groups of panelists assessed 161 and 143 foods respectively, including two control foods, a whole-wheat bread, and a cracker. PROC Mixed procedure in SAS was used for the statistical analyses.

There was no significant panel-food interaction in the case of control foods, indicating that the control foods were rated similarly by the two groups of panelists. Session was included as a continuous variable in the model to detect any trend in acceptability scores with successive sessions. Session effect was not significant (p < 0.05). ANOVA was conducted first by giving the same number to the control foods (in the food variable) and second by giving them different numbers in different sessions, thus considering them different foods. The F test values for the two methods were 10.73 and 9.02, indicating a higher precision in the first case.

Control food(s) protect against session-to-session variation especially in cases where uncontrollable trends or variations are expected.