close

Вход

Забыли?

вход по аккаунту

код для вставкиСкачать
RELIABILITY AND VALIDITY OF JUDGING ARTISTRY ON BALANCE BEAM AT
WORLD CHAMPIONSHIP IN TOKIO 2011
M.
Introduction
The judging in artistic gymnastics crucially determines
outcome. Currently, final score is a sum of difficulty
score and execution (E) score by the panel of 6 judges.
E score is calculated by subtracting deductions for
various mistakes from 10 points, including the
deductions for flaws in artistry. Evaluation of artistry
could be associated with serious insufficiencies of
reliability and validity. We have analyzed the reliability
and validity of deductions in artistry at one of the top
gymnastics competitions.
Methods
The evaluated categories of artistry are: inappropriate
gesture or mimic, insufficient variation in rhythm,
sureness of performance and lack of creative
choreography. We have used the data from judge
evaluation sheets to examine the deductions for artistry
for each individual contestant. We have calculated the
correlation coefficients (Kendall's tau-b) for association
between the total artistry deduction and final E score.
Same method was used to calculate the correlations in
separate categories of artistry between all possible
judge-pairs. Finally, the special aspect of judging validity
was examined by calculating the Kendall's W statistic this method gives high and statistically significant values
of W, when individual judges systematically give
relatively high or low deductions (systematic bias).
1
Bučar ,
I.
1
Čuk ,
J.
2
Pajek ,
M.
1
Kovač ,
B.
1
Leskovšek
1 Faculty of sport, University of Ljubljana Slovenia
2 University Medical Centre, Ljubljana Slovenia
Judge 1
Results
The number of contestants penalized for other components
of artistry was highly variable between the judges: no
deduction for insufficient variation in rhythm was given to
either 10 contestants (judge 5), 46-88 contestants (judges
1,3 and 4), but even up to 138 contestants (judge 2). High
variability was found in the magnitude of deductions: for the
sureness of performance judge 1 used the 0,3 deduction in
158 contestants and judge 3 in only 9 contestants. The
range of mean total deductions for artistry was 0,18-0,39.
The average Kendall's tau-b correlation coefficient between
the total artistry deduction and final E score for the 5 judges
was -0,61 (range -0,49 to -0,71, p for all <0,001). The
average correlation between judge pairs in same artistry
categories was 0,45 (range 0,01-0,83). When the total
artistry deductions were examined for validity, the Kendall's
W statistic was high: 0,53 (p<0,001). The highest problems
in validity were found in sureness of performance: Kendall's
W 0,53 (p<0.001).
No deduction
Deduction 0,1
Deduction 0,2
Deduction 0,3
Deduction 0,4
Deduction 0,5
Deduction 0,6
Deduction 0,7
Deduction 0,8
Missing
Average deduction
Judge 2
Judge 3
Judge 4
Judge 5
1
17
31
31
1
11
46
53
27
15
13
33
62
26
52
49
57
35
31
22
58
26
3
6
50
46
6
8
18
24
4
2
1
3
1
12
4
1
3
16
0
0
0
3
0
0
0,39
3
0,24
0
0,18
46
0,24
13
0,34
Conclusions
Our results have shown that reliability of judging artistry is
low: there was a high variability between the judges in the
number of penalized contestants in separate categories of
artistry and a very low average correlation between judgepairs for the same categories of artistry. Of highest concern is
our finding of significant systematic bias between judges
implicating a lack of validity. It is clear that we have to
monitor the judging of artistry in future and if such findings
are repetitively found, the justification for including artistry
deductions should be reassessed by the regulating body
(FIG).
[email protected]
1/--страниц
Пожаловаться на содержимое документа