An inter-collaborative tasting exercise was undertaken in which panels
at seven different brewing companies compared the flavour of four bee
rs in duplicate. Statistical analysis demonstrated that the results fr
om individual panels were repeatable but there was little consistency
between panels. Generalised Procrustes Analysis demonstrated that inco
nsistencies between panels were due to differences in the use of flavo
ur terminology and differences in the use of scoring scales. It is sug
gested that consistency could be significantly improved by introducing
standardised flavour methods and training across the brewing industry
.