Résumé
Abstract Aim: This study aimed (i) to test the inter-rater reliability of swimming teachers, (ii) to test the swimming teacher's discussion effect on inter-rater reliability and (iii) to verify the intra-rater swimming teacher's reliability. Method: Twenty-one learning swimmers (14.1 ± 5.1 years old) performed two 25-m front crawl courses at a comfortable speed without breathing between the sixth and 20-m, and had their displacements captured on film. Three swimming teachers with different academic backgrounds and skills evaluated the swimmer right upper limb using a 20-items checklist. In the 1st-step, teachers assessed 20-items and in 2nd-step discussed their particular evaluating criteria - selecting five items considered as the most relevant. The inter- and intra-rater reliability were tested using the Fleiss Kappa Coefficient. Results: In the 1st-step substantial reliability was found for item three and in movement descriptor for items three and 20. Nearly perfect reliability was found in the movement descriptor for item 13. In 2nd-step, moderate reliability was found only in the movement descriptor for item 20. Only the most experienced evaluator showed substantial intra-rater reliability for items four and 10 and moderate for item 20. Conclusion: The proposed discussion method did not cause the expected effect on inter-rater reliability. The swimming teacher with a higher degree and swimming skills showed better intra-rater reliability. Some items and movement descriptors proposed at the 20-items checklist can be used in practical settings.