RSS Towards Data Science - Medium

Statistical Analysis on Scoring Bias

The 2024 Argentine Tango World Championship, also known as Tango de Mundial, was held in Buenos Aires, Argentina, with around 500 dance couples competing in the preliminary round. The competition is highly subjective, with judges having their own opinions on what constitutes quality in various judging criteria. The data from the competition is limited, with only dancer scores available, and the judges are all exclusively Argentine, which has led some to question the legitimacy of the competition. A statistical analysis was performed to assess scoring bias between two judging panels, each comprising 10 judges. The analysis found that the proportion of dancers judged by panel 2 who advanced to the semi-final round was significantly higher than those judged by panel 1. A two-tailed z-test was used to test the difference in proportions, and the results showed that the difference was statistically significant, with a p-value of 0.0. The analysis also found that 17% of dance couples judged by panel 1 advanced to the semi-finals, compared to 33% of those judged by panel 2. This suggests that panel 2 was more likely to give higher scores, resulting in more dancers advancing to the semi-finals. The results of the analysis suggest that there is a bias in the scoring between the two panels. The analysis also highlighted the limitations of the data, including the fact that judges don't represent the world, scoring dancers to the 100th decimal place is absurd, and corruption and politics may play a role in the competition. Despite these limitations, the data from Tango Munidal is the largest and most representative dataset of worldwide Argentine tango dancers available. The author of the analysis is a data scientist and a competitive Argentine tango dancer, and their opinion represents a genuine and informed voice on Argentine tango. The analysis provides insights into the subjective nature of judging in the competition and highlights the need for more comprehensive data to better understand the biases and limitations of the competition.
towardsdatascience.com
towardsdatascience.com