WINETASTER ON 02/16/99 WITH 8 JUDGES AND 12 WINES BASED ON RANKS, IDENT=N
Copyright (c) 1995-98 Richard E. Quandt
FLIGHT 1:
Number of Judges = 8
Number of Wines = 12
Identification of the Wine: The judges' overall ranking:
Wine A is Kesselstadt Scharzhofberger Sptls ........ 12th place
Wine B is Klosterberg, Zeltinger Sonnenuhr S ........ 7th place
Wine C is Selbach-Oster,Bernkastel Badstube tied for 8th place
Wine D is Busch, Puendericher Marienburg Aus ........ 1st place
Wine E is Kartaeuserhof, Eitelsbacher Ausles ........ 3rd place
Wine F is Busch, Puendericher Marienburg Spt ........ 2nd place
Wine G is Christoffel, Uerziger Wuerzgarten ........ 6th place
Wine H is Loewenstein, Schieferterassen 97 tied for 8th place
Wine I is Kallfelz, Merler Koenigslay Sptls ........ 10th place
Wine J is Immich-Batterieberg,Enkircher Batt ........ 11th place
Wine K is Kartaeuserhof, Eitelsbacher Sptls ........ 4th place
Wine L is Kallfelz, Merler Koenigslay Ausles ........ 5th place
The Judges's Rankings
Judge Wine -> A B C D E F G H I J K L
Clemens 12. 11. 4. 1. 6. 2. 9. 3. 10. 8. 7. 5.
Rita 11. 9. 6. 1. 5. 2. 3. 4. 12. 8. 7. 10.
Wolfgang 12. 5. 10. 6. 3. 7. 8. 9. 4. 11. 1. 2.
Kai 12. 10. 6. 11. 3. 1. 2. 9. 8. 7. 4. 5.
Pascale 12. 1. 8. 6. 3. 10. 9. 5. 4. 11. 2. 7.
Karl 12. 3. 7. 1. 4. 6. 2. 5. 11. 8. 9. 10.
Joern 8. 4. 3. 1. 7. 2. 12. 11. 9. 10. 5. 6.
Pit 12. 10. 11. 2. 8. 7. 3. 9. 4. 5. 6. 1.
Table of Votes Against
Wine -> A B C D E F G H I J K L
Group Ranking -> 12 7 8 1 3 2 6 8 10 11 4 5
Votes Against -> 91 53 55 29 39 37 48 55 62 68 41 46
( 8 is the best possible, 96 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 = 0.3269
The probability that random chance could be responsible for this correlation
is quite small, 0.0025. 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
Clemens 0.5219
Rita 0.5114
Karl 0.4904
Wolfgang 0.4168
Joern 0.4098
Kai 0.3684
Pit 0.1611
Pascale 0.1479
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 Busch, Puendericher Marienburg Aus
---------------------------------------------------
2. ........ 2nd place Wine F is Busch, Puendericher Marienburg Spt
3. ........ 3rd place Wine E is Kartaeuserhof, Eitelsbacher Ausles
4. ........ 4th place Wine K is Kartaeuserhof, Eitelsbacher Sptls
5. ........ 5th place Wine L is Kallfelz, Merler Koenigslay Ausles
6. ........ 6th place Wine G is Christoffel, Uerziger Wuerzgarten
7. ........ 7th place Wine B is Klosterberg, Zeltinger Sonnenuhr S
8. tied for 8th place Wine C is Selbach-Oster,Bernkastel Badstube
9. tied for 8th place Wine H is Loewenstein, Schieferterassen 97
10. ........ 10th place Wine I is Kallfelz, Merler Koenigslay Sptls
11. ........ 11th place Wine J is Immich-Batterieberg,Enkircher Batt
---------------------------------------------------
12. ........ 12th place Wine A is Kesselstadt Scharzhofberger Sptls
We now test whether the ranksums AS A WHOLE provide a significant ordering.
The Friedman Chi-square value is 28.7692. The probability that this could
happen by chance is 0.0025
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 0.59 for significance at the 0.05
level and must exceed 0.50 for significance at the 0.1 level
Clemens Rita Wolfgang
Clemens 1.000 0.734 0.098
Rita 0.734 1.000 -0.084
Wolfgang 0.098 -0.084 1.000
Kai 0.252 0.357 0.343
Pascale -0.021 -0.070 0.748
Karl 0.385 0.783 0.028
Joern 0.476 0.266 0.266
Pit 0.252 0.154 0.462
Kai Pascale Karl
Clemens 0.252 -0.021 0.385
Rita 0.357 -0.070 0.783
Wolfgang 0.343 0.748 0.028
Kai 1.000 -0.035 0.119
Pascale -0.035 1.000 0.259
Karl 0.119 0.259 1.000
Joern -0.028 0.175 0.182
Pit 0.273 0.021 0.133
Joern Pit
Clemens 0.476 0.252
Rita 0.266 0.154
Wolfgang 0.266 0.462
Kai -0.028 0.273
Pascale 0.175 0.021
Karl 0.182 0.133
Joern 1.000 -0.063
Pit -0.063 1.000
Pairwise correlations in descending order
0.783 Rita and Karl Significantly positive
0.748 Wolfgang and Pascale Significantly positive
0.734 Clemens and Rita Significantly positive
0.476 Clemens and Joern Not significant
0.462 Wolfgang and Pit Not significant
0.385 Clemens and Karl Not significant
0.357 Rita and Kai Not significant
0.343 Wolfgang and Kai Not significant
0.273 Kai and Pit Not significant
0.266 Wolfgang and Joern Not significant
0.266 Rita and Joern Not significant
0.259 Pascale and Karl Not significant
0.252 Clemens and Kai Not significant
0.252 Clemens and Pit Not significant
0.182 Karl and Joern Not significant
0.175 Pascale and Joern Not significant
0.154 Rita and Pit Not significant
0.133 Karl and Pit Not significant
0.119 Kai and Karl Not significant
0.098 Clemens and Wolfgang Not significant
0.028 Wolfgang and Karl Not significant
0.021 Pascale and Pit Not significant
-0.021 Clemens and Pascale Not significant
-0.028 Kai and Joern Not significant
-0.035 Kai and Pascale Not significant
-0.063 Joern and Pit Not significant
-0.070 Rita and Pascale Not significant
-0.084 Rita and Wolfgang Not significant
COMMENT:
We tested the same wines as the night before. Each bottle had about 20
hours to breathe. We corrected the mistake with the Busch Spaetlese and
inserted a Busch Auslese. It was also opened the night before. Everybody
agreed that it is much easier to the strengths and weaknesses (although
all of us had a hard night). Maybe this is the reason for the higher
correlations among the judges than the night before. The theory was
proposed that the correlation might have been much higher if the wines had
been opened 3 or 4 days before the tasting (see the October tasting).
Return to previous page or
go to Report 11