Abstract
This study aimed to develop a one-dimensional scale for the difficulty of test items and to examine the validity of the ability score based on Item Response Theory, the logistic model with three parameters, ‘difficulty of item’, ‘discriminating power’, and ‘guessing’. Subjects were 3518 young children aged from 2.5-years-old to under 6.5-years-old, and the pass-or-fail tests included 14 items representing motor patterns of the locomotion type. The difficulty score of the items properly reflected the difficulty of performances and was a stable value independent of the constitution of the subjects. Therefore, the score was considered to be an effective one-dimensional scale. The ability score, based on the data pattern of pass-or-fail, was superior in diversity and was stable among differentially constituted groups. In addition, that score also reflected the total of passed items which is widely used for estimating the motor skill and showed a significantly higher mean in older age groups in all comparisons among age groups. The ability score was supposed as an effective scale to estimate the motor skill.
