In advance: The graphic shows all the votings performed so far on our homepage images. In some votings, older images were removed beforehand. As a result, fewer pictures were voted on. These reduced votings are greyed out and are not included in the following analysis.

- On the far right, you can see the current voting using the pairwise comparison method, which is highlighted in yellow. (I have removed the image numbers so that the current process is not affected.) You can see that the entire scale range is well cupped. The unevenness of the curve is due to the fact that only 12 judges took part.

- To the left, you can see all the former scale-based votes for comparison. One immediately recognises the general reluctance to exhaust the entire available scale. The jurors generally rate too positively. As a result, the shape of the curve is relatively flat, which reduces the differentiation.

- It can also be seen that the scale-based method requires more judges (26 to 30) in order to achieve a similarly good differentiation (40 to 43 ranks) as the pairwise comparison already provides with 12 judges (42 ranks).

- In the pairwise comparison (right), only 5 ranks have been awarded more than once by 12 jurors to date. In the scale-based evaluations, the number of multiple awards is significantly higher. The differentiation is worse there.

Of course, these are only initial results from a single example. They would have to be verified by further analyses of other case studies.

For more information, see my article: Pairwise Comparison for Photo Evaluation