main effects of judge (j), EC treatment (t) and replication (r), as well as the interaction between judge × treatment (j × t), judge × replication (j × r) and treatment × replication (t × r). For terms with significant j × t effects (juiciness-by-mouth, p ≤ 0.01), the F-value was examined in more detail using techniques outlined by Goniak and Noble (1987) (F = MSt/MSj×t = 0.48). The recalculated F-value for this term reflected that it was not being used consistently enough and was dropped from further analysis.