A Method for the Statistical Selection of a Panel of Rum Tasters

Abstract

Employees of an agricultural research unit were evaluated for selection as rum tasters. Prospects were classified and ranked considering their relative consistency in four organoleptic tests of four rum samples and their evaluation relative to those of an experienced rum taster. Statistical techniques used were variance analysis of Latin squares for the scores of the evaluation of 10 rum attributes and the calculation of a rum evaluation index for each taster using a linear discriminant function.