ISSN:
1750-3841
Source:
Blackwell Publishing Journal Backfiles 1879-2005
Topics:
Agriculture, Forestry, Horticulture, Fishery, Domestic Science, Nutrition
,
Process Engineering, Biotechnology, Nutrition Technology
Notes:
The use of correlation coefficients, F-statistics and LSDs was described for measuring judge performance in terms of agreement, reliability, discrimination, stability, and variability. The technique was applied to a number of wine-quality evaluation experiments. It was shown that a single analysis of variance of all judges' scores in an experiment will often be inappropriate. Further, it was demonstrated that judges' performances varied over time. It was, therefore, recommended that each judge's performance be monitored continually and that when judges were unreliable and nondiscriminating, they would be ignored in drawing conclusions from the experiment. The analysis for wine differences should be based on a separate analysis of each set of homogeneous judges as determined from the measures for agreement, reliability, and variability.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1111/j.1365-2621.1987.tb14061.x
Permalink