K. Steinsholt, ARE ASSESSORS LEVELS OF A SPLIT-PLOT FACTOR IN THE ANALYSIS OF VARIANCE OF SENSORY PROFILE EXPERIMENTS, Food quality and preference, 9(3), 1998, pp. 153-156
With simple theoretical and practical examples from research in the mi
lk processing industry, it is argued that assessors should be treated
statistically as levels of a fixed 'split-plot' factor in analysis of
variance. The calculations will then be equal to using the assessors a
s 'repeated measures'. If the assessors are considered to be random va
riables, the calculations will be similar to using the assessor as a b
lock factor. Published by Elsevier Science Ltd. All rights reserved.