| Abstract [eng] |
Summary The purpose of this research is to evaluate WAIS-III psychometric characteristics. In total 75 subjects participated in the research; they were provided with WAIS-III tasks by the qualified researchers. When selecting the participants for the research, a target selection was carried out based on the following criteria: age, place of residence and education. 36 men and 39 women participated in this research. The age of the participants was from 16 to 79 years. While evaluating WAIS-III psychometric characteristics suitability of separate tasks of the subtests was estimated, as well as availability of new tasks, reliability of verbal and non-verbal subtests by the odd/even method and also the validity of the construct of the test. Suitability of the tasks of the verbal and non-verbal subtests was evaluated based on the indexes of the complexity of the task and the differential power, coefficient of the adjusted correlation. The Vocabulary subtest was transformed the most, in it 21 words from the original version were maintained and 12 new words added. In the Similarities subtest 19 tasks from the original version were maintained and 3 new tasks added. In the Information subtest 24 questions from the original version were maintained and 4 new questions added. In the Comprehension subtests 18 tasks were maintained from the original version and 2 new tasks were added. All the tasks of verbal subtests were set out in a new order, except for the Digit Span and Letter – number sequencing. The tasks of Picture Completion, Block Design, Matrices and Picture Arrangement were set out in a new order. Object Assembly subtest remained the same as is provided in the original version. When evaluating the reliability of verbal subtests it was determined that the most reliable subtests are those of the Vocabulary and Information, on the average reliable are the subtests of Similarities, Arithmetic and Comprehension, the least reliable are the subtests of Digit Span and Letter – number sequencing. When evaluating the reliability of the non-verbal subtests it was determined that the most reliable subtests are those of Block Design and Matrices, on the average reliable are the subtests of Picture Completion, Picture Arrangement and Object Assembly. All verbal subtests are in better correlation with the verbal subtests than with non-verbal ones. The best correlation is among Vocabulary, Similarities and Information subtests. All non-verbal subtests are in better correlation with non-verbal subtests than with the verbal ones. The best correlation is among Picture Completion, Block Design, Digit Symbol and Matrices subtests. Assessment of the verbal subtests performed by the participants of different age did not differ, except for the Digit Span and Letter – number sequencing. Assessments of non-verbal tasks of participants of different age differ significantly. |
