The author compared two methods for assessing the reliability of blind wine tastings of South African Sauvignon Blanc and Pinotage white and red wines. The first method required the taster to rank each of the 8 white and red wines from 1 for the most preferred to 8 for the least preferred, with tied ranks permitted when 2 or more wines were equally preferred. The second used scores representing a taster’s perceived quality of wine as: 50%-‐69% =Poor/Unacceptable; 70%-‐79%= Fair/Mediocre; 80%-‐89%= Good/Above Average; and 90%-‐100%= Excellent to Superior. Results indicated that the two methods provided quite different results with respect to the critical issue of identifying replicate wines; while the identification of replicate white and red wines proved successful when wine scores were used, not a single wine taster could successfully identify replicate white or red wines when using ranking methods. This finding, plus the fact that very different scorings can produce the very same rankings, such as 1, 2, and 3 being given scores of 60, 64, 66 (Poor/Unacceptable); 70,72, 76 (Fair/Mediocre); 80, 83, 87 (Good/Above Average; and 90, 95, 98. (Excellent to Superior), led to the conclusion that such flaws in the ranking method render it of very limited usefulness in the evaluation of the results of blind wine tasting.
This article was written by Karl Storchmann