Background: Coaches and physical educators must obtain valid data relating to the contribution of each of their players in order to assess their level of performance in team sport competition. This information must also be collected and used in real game situations to be more valid. Developed initially for a physical education class context, the Team Sport Assessment Procedure (TSAP) is precisely an assessment procedure developed to measure the individual performance of each player in a game situation. The procedure has shown appropriate levels of validity for performance measurement in team sports (soccer, handball, volley-ball) taught in physical education classes. However, this procedure was never adapted for ice hockey and all of its specific characteristics. Purpose: Ice hockey has specific characteristics of play which were not considered in the original TSAP model. In order to establish the quality of the procedure in this new context, it was necessary to verify its psychometrical qualities. The purpose of the study was to establish the validity and reliability of the Team Sport Assessment Procedure (TSAP) in the context of ice hockey. Participants and setting: Nineteen young hockey players, between 14 and 17 years of age, participated in this study. These subjects played in different elite teams (CC to AA levels) in the Quebec City area in Canada. All subjects took part in three afternoon training practices per week in the context of a Sports-Studies school programme. Research design: Two correlational studies were used to determine the validity and the inter-rater reliability of the measuring instrument. Data collection: Data were collected using a peer observation strategy, similar to the one used by Grehaigne et al. in their original study. The performance of each player was assessed during three simulated games of two minutes. Two student-observers were assigned for each player during the game. Data analysis: Validity analysis was conducted using the Spearman rank correlation between the average ranks assigned by expert observers to the players performance and scores calculated from student-observers' collected data using the TSAP. Reliability was established using percentage of agreement between student-observers for the same player observed. Findings: A correlation coefficient of -0.77 was obtained between the average ranks assigned by expert observers to players' performance and scores calculated from student-observers' collected data using the TSAP. In addition, peer observation during a simulated game showed that young players can effectively use the TSAP in an objective manner. A mean percentage of agreement of 81.4% was calculated between student-observers. Conclusions: The results from these different analyses showed that TSAP is a valid and reliable instrument for the assessment of ice hockey performance. Conclusions drawn from the study are similar to those obtained by Grehaigne et al. for the original procedure. (Contains 4 tables, 3 figures, and 3 notes.)