WINETASTER ON 11/09/99 WITH 7 JUDGES AND 6 WINES BASED ON RANKS, IDENT=N Copyright (c) 1995-99 Richard E. Quandt
FLIGHT 1: Number of Judges = 7 Number of Wines = 6
Identification of the Wine: The judges' overall ranking:
Wine A is Thierry German, Marginale, '96 tied for 4th place Wine B is Stonefly Vineyards '97, Napa ........ 6th place Wine C is Chapellet '96, Napa ........ 3rd place Wine D is Pride Mountain '97, Sonoma ........ 1st place Wine E is Peju '94, Napa ........ 2nd place Wine F is L'Ecosse '95, Napa tied for 4th place
The Judges's Rankings
Judge Wine -> A B C D E F John 5. 6. 4. 2. 3. 1. Grant 6. 4. 5. 2. 3. 1. Bob E. 1. 4. 3. 2. 6. 5. Frank 5. 6. 2. 1. 3. 4. Bob A. 4. 6. 5. 1. 2. 3. Burt 3. 6. 2. 4. 1. 5. Bill 1. 5. 2. 3. 4. 6.
Table of Votes Against Wine -> A B C D E F
Group Ranking -> 4 6 3 1 2 4 Votes Against -> 25 37 23 15 22 25
( 7 is the best possible, 42 is the worst)
Here is a measure of the correlation in the preferences of the judges which ranges between 1.0 (perfect correlation) and 0.0 (no correlation):
W = 0.2980
The probability that random chance could be responsible for this correlation is quite small, 0.0640. Most analysts would say that unless this probability is less than 0.1, the judges' preferences are not strongly related. We now analyze how each taster's preferences are correlated with the group preference. A correlation of 1.0 means that the taster's preferences are a perfect predictor of the group's preferences. A 0.0 means no correlation, while a -1.0 means that the taster has the reverse ranking of the group. This is measured by the correlation R.
Correlation Between the Ranks of Each Person With the Average Ranking of Others
Name of Person Correlation R Frank 0.6667 Bob A. 0.6000 John 0.4058 Burt 0.1160 Grant -0.0580 Bill -0.0857 Bob E. -0.1449
The wines were preferred by the judges in the following order. When the preferences of the judges are strong enough to permit meaningful differentiation among the wines, they are separated by -------------------- and are judged to be significantly different.
1. ........ 1st place Wine D is Pride Mountain '97, Sonoma --------------------------------------------------- 2. ........ 2nd place Wine E is Peju '94, Napa 3. ........ 3rd place Wine C is Chapellet '96, Napa 4. tied for 4th place Wine A is Thierry German, Marginale, '96 5. tied for 4th place Wine F is L'Ecosse '95, Napa --------------------------------------------------- 6. ........ 6th place Wine B is Stonefly Vineyards '97, Napa We now test whether the ranksums AS A WHOLE provide a significant ordering. The Friedman Chi-square value is 10.4286. The probability that this could happen by chance is 0.0640 We now undertake a more detailed examination of the pair-wise rank correla- tions that exist between pairs of judges. First, we present a table in which you can find the correlation for any pair of judges, by finding one of the names in the left hand margin and the other name on top of a column. A second table arranges these correlations in descending order and marks which is significantly positive significantly negative, or not significant. This may allow you to find clusters of judges whose rankings were particularly similar or particularly dissimilar. Pairwise Rank Correlations Correlations must exceed in absolute value 0.89 for significance at the 0.05 level and must exceed 0.83 for significance at the 0.1 level John Grant Bob E. John 1.000 0.829 -0.314 Grant 0.829 1.000 -0.543 Bob E. -0.314 -0.543 1.000 Frank 0.600 0.314 0.086 Bob A. 0.771 0.600 -0.086 Burt 0.086 -0.314 -0.086 Bill -0.371 -0.771 0.771 Frank Bob A. Burt John 0.600 0.771 0.086 Grant 0.314 0.600 -0.314 Bob E. 0.086 -0.086 -0.086 Frank 1.000 0.657 0.486 Bob A. 0.657 1.000 0.314 Burt 0.486 0.314 1.000 Bill 0.257 -0.029 0.543 Bill John -0.371 Grant -0.771 Bob E. 0.771 Frank 0.257 Bob A. -0.029 Burt 0.543 Bill 1.000 Pairwise correlations in descending order 0.829 John and Grant Not significant 0.771 John and Bob A. Not significant 0.771 Bob E. and Bill Not significant 0.657 Frank and Bob A. Not significant 0.600 John and Frank Not significant 0.600 Grant and Bob A. Not significant 0.543 Burt and Bill Not significant 0.486 Frank and Burt Not significant 0.314 Grant and Frank Not significant 0.314 Bob A. and Burt Not significant 0.257 Frank and Bill Not significant 0.086 John and Burt Not significant 0.086 Bob E. and Frank Not significant -0.029 Bob A. and Bill Not significant -0.086 Bob E. and Bob A. Not significant -0.086 Bob E. and Burt Not significant -0.314 John and Bob E. Not significant -0.314 Grant and Burt Not significant -0.371 John and Bill Not significant -0.543 Grant and Bob E. Not significant -0.771 Grant and Bill Not significant
COMMENT: These were all Cabernet Franc wines, and their prices all ranged between $25 and $35. The l'Ecosse and the Stonefly may have been "underpriced" because they were obtained directly from the vineyards.Report
WINETASTER ON 11/09/99 WITH 7 JUDGES AND 3 WINES BASED ON RANKS, IDENT=N Copyright (c) 1995-98 Richard E. Quandt
FLIGHT 2: Number of Judges = 7 Number of Wines = 3
Identification of the Wine: The judges' overall ranking:
Wine A is Viader '97, Napa ........ 1st place Wine B is Cheval Blanc '94 ........ 3rd place Wine C is L'Angelus '94 ........ 2nd place
The Judges's Rankings
Judge Wine -> A B C John 1. 3. 2. Grant 1. 2. 3. Bob E. 1. 3. 2. Frank 1. 3. 2. Bob A. 1. 2. 3. Burt 1. 3. 2. Bill 2. 3. 1.
Table of Votes Against Wine -> A B C
Group Ranking -> 1 3 2 Votes Against -> 8 19 15
( 7 is the best possible, 21 is the worst)
Here is a measure of the correlation in the preferences of the judges which ranges between 1.0 (perfect correlation) and 0.0 (no correlation):
W = 0.6327
The probability that random chance could be responsible for this correlation is quite small, 0.0119. Most analysts would say that unless this probability is less than 0.1, the judges' preferences are not strongly related. We now analyze how each taster's preferences are correlated with the group preference. A correlation of 1.0 means that the taster's preferences are a perfect predictor of the group's preferences. A 0.0 means no correlation, while a -1.0 means that the taster has the reverse ranking of the group. This is measured by the correlation R.
Correlation Between the Ranks of Each Person With the Average Ranking of Others
Name of Person Correlation R John 1.0000 Bob E. 1.0000 Frank 1.0000 Burt 1.0000 Grant 0.5000 Bob A. 0.5000 Bill 0.5000
The wines were preferred by the judges in the following order. When the preferences of the judges are strong enough to permit meaningful differentiation among the wines, they are separated by -------------------- and are judged to be significantly different.
1. ........ 1st place Wine A is Viader '97, Napa --------------------------------------------------- 2. ........ 2nd place Wine C is L'Angelus '94 --------------------------------------------------- 3. ........ 3rd place Wine B is Cheval Blanc '94 We now test whether the ranksums AS A WHOLE provide a significant ordering. The Friedman Chi-square value is 8.8571. The probability that this could happen by chance is 0.0119 We now undertake a more detailed examination of the pair-wise rank correla- tions that exist between pairs of judges. First, we present a table in which you can find the correlation for any pair of judges, by finding one of the names in the left hand margin and the other name on top of a column. A second table arranges these correlations in descending order and marks which is significantly positive significantly negative, or not significant. This may allow you to find clusters of judges whose rankings were particularly similar or particularly dissimilar. Pairwise Rank Correlations Correlations must exceed in absolute value 1.00 for significance at the 0.05 level and must exceed 1.00 for significance at the 0.1 level John Grant Bob E. John 1.000 0.500 1.000 Grant 0.500 1.000 0.500 Bob E. 1.000 0.500 1.000 Frank 1.000 0.500 1.000 Bob A. 0.500 1.000 0.500 Burt 1.000 0.500 1.000 Bill 0.500 -0.500 0.500 Frank Bob A. Burt John 1.000 0.500 1.000 Grant 0.500 1.000 0.500 Bob E. 1.000 0.500 1.000 Frank 1.000 0.500 1.000 Bob A. 0.500 1.000 0.500 Burt 1.000 0.500 1.000 Bill 0.500 -0.500 0.500 Bill John 0.500 Grant -0.500 Bob E. 0.500 Frank 0.500 Bob A. -0.500 Burt 0.500 Bill 1.000 Pairwise correlations in descending order 1.000 John and Frank Significantly positive 1.000 John and Bob E. Significantly positive 1.000 John and Burt Significantly positive 1.000 Bob E. and Burt Significantly positive 1.000 Frank and Burt Significantly positive 1.000 Bob E. and Frank Significantly positive 1.000 Grant and Bob A. Significantly positive 0.500 John and Bill Not significant 0.500 John and Grant Not significant 0.500 Grant and Frank Not significant 0.500 Burt and Bill Not significant 0.500 Grant and Burt Not significant 0.500 Bob E. and Bob A. Not significant 0.500 John and Bob A. Not significant 0.500 Bob E. and Bill Not significant 0.500 Frank and Bob A. Not significant 0.500 Grant and Bob E. Not significant 0.500 Frank and Bill Not significant 0.500 Bob A. and Burt Not significant -0.500 Bob A. and Bill Not significant -0.500 Grant and Bill Not significant
COMMENT: The second flight of wines consisted of mixtures in which the Cabernet Franc was mixed with Cabernet Sauvignon or Merlot or Malbec. The percentages were Viader: Cabernet Franc 38%, Cabernet Sauvignon 62%; Cheval Blanc: Cabernet Franc 66%, Merlot 33%, Malbec 1%;, L'Angelus: Cabernet Franc 50%, Merlot 45%, Cabernet Sauvignon 5%. The Viader was a surprise winner over Cheval Blanc and L'Angelus. All but two of the judges thought that the L'Angelus was the American wine; one taster thought that the Cheval Blanc was the American wine because it was more tannic that the Viader.Return to previous page