Reliability of Scores Obtained with the Objectives-Based Motor Skill Assessment Instrument

This study was designed to investigate the reliability of individual and composite scores obtained with the Objectives-Based Motor Skill Assessment Instrument. Generalizability theory was used to determine if the test scores were reliable across raters and occasions. Two D studies were conducted to ascertain whether an increase in the number of raters from 10 to 20 would increase the reliability. Twenty raters were required to evaluate the fundamental motor skills of 10 subjects across two occasions from videotapes. In 92% of the individual skill scores, the between-subject variance component contributed the most to total variance. Therefore, it can be concluded that the test is extremely reliable across raters and occasions. Examination of the generalizability coefficients for the two D studies indicates that the reliability is not largely affected by increasing the number of raters.