유아의 문제행동 검사도구 개발

Abstract

ABSTRACT

Development of a Problem Behavior Rating Scale for Young Children

Go, Hyun Advisor : Prof. Yoon, Cho-Hee, Ph.D. Department of Education Graduate School Chosun University

The purpose of the present study was to develop a problem behavior rating scale for young children and to verify its adequacy. Research questions to be answered for this purpose are as follows. Question 1: How valid is the rating scale for young children?fs problem behavior developed in this study? Question 2: How discriminative are the questions in the rating scale for young children?fs problem behavior developed in this study? Question 3: How reliable is the rating scale for young children?fs problem behavior developed in this study? In order to answer these questions, we conducted three preliminary studies for constructing and drafting questions and a validation study for verifying the adequacy of the rating scale. The preliminary studies for drafting questions were conducted from January 7 to April 9, 2006. Before the preliminary studies, we reviewed previous research on young children?fs problem behavior and examined existing rating scales for young children?fs problem behavior. Based on the review, a total of 216 preliminary questions were developed. The 1st preliminary study for testing face validity was conducted with 10 field teachers, the 2nd one with 46 specialists in early childhood education, and the 3rd one with 50 teachers and 50 parents. Through these procedures, 44 questions were adopted finally for testing their validity and reliability. The validation study was conducted from May 3 to June 24, 2006. The subjects of the study were 316 3?`5 year old children attending one of 11 early childhood education institutions in large cities, small/medium cities, towns and villages. The concurrent validity and the test-retest reliability of the scale were verified with 127 children selected out of the 316 children, and the discriminative validity of the questions was tested with 158 children who belonged to the top 25% or the bottom 25%, based on the mean score of the scale. The methods of data collection and analysis were as follows. In order to answer Question 1 ?gHow valid is the rating scale for young children?fs problem behavior developed in this study,?h the construct validity and concurrent validity of the scale were analyzed. The construct validity was tested through exploratory factor analyses using SPSS 15.0 and confirmatory factor analyses using AMOS 7.0. In order to test concurrent validity, Pearson?fs product moment correlation coefficient r was calculated between the problem behavior rating scale for young children developed in this study and K-CBCL which was developed by Achenbach and Edelbrock (1983) and translated by Oh, Lee, Hong, and Ha (1997). In order to answer Question 2 ?gHow discriminative are the questions in the rating scale for young children?fs problem behavior developed in this study,?h t tests were performed on the behavior scores of the children belonging to the top 25% or the bottom 25%, based on the mean score of the scale. In order to answer Question 3 ?gHow reliable is the rating scale for young children?fs problem behavior developed in this study,?h Pearson?fs product moment correlation coefficient r was calculated between each question item and the total score of the rating scale. In addition, Cronbach ?? was calculated for verifying the reliability of the scale based on internal consistency among the question items. Finally, test-retest reliability was analyzed. The results of this study are summarized as follows. First, the results of the exploratory factor analysis for testing the construct validity of the rating scale showed that the factors related to problem behavior were peer relation, development, basic life habit, emotion, morality and sex, and that the factor loading of each factor ranged between .419 and .819. The six factors explained around 51% of the whole variance. In addition, the results of the confirmatory factor analysis showed that the goodness of fit of the measurement model represented by a variety of indexes was acceptable. As the indicator of concurrent validity, the correlation of the rating scale developed in this study and K-CBCL was .688, which indicated a relatively high correlation between the two measures. Consequently, the rating scale developed in this study was found to have construct and concurrent validity. Second, the standard deviation for each question in the scale was found to range from .35 to .95, showing that the distribution of the children?fs responses was not too narrow. Also the results of the analysis for the discriminative validity of the questions showed that the top 25% and bottom 25% groups showed significant differences in each question item and each factor at ?? <.001. This suggests that the rating scale can discriminate between a group in which problem behaviors are observed frequently and one in which problem behaviors are observed infrequently. Third, the results of the reliability analyses showed that the correlation between the score of each question and the total score of the scale ranged between .336?`.687, and the correlation between the score of each sub?]factor and the total score ranged between .530?`.874. This suggests that the questions and the sub factors of the rating scale are homogeneous in the internal structure. In addition, the internal agreement of the scale ranged between .69?`.90, showing that the questions of the rating scale are homogeneous and have highly reliable internal consistency. Furthermore, the results of the correlational analysis of the test and retest scores showed that the correlation coefficient between the test and retest scores in each sub?]factor and the total score ranged between .864?`.897, proving that the rating scale is reliable. In conclusion, the problem behavior rating scale for young children developed in this study was found to be valid and reliable with questions that can discriminate problem behaviors observed in children?fs everyday life. In addition, unlike other problem behavior rating scales for young children, the current rating scale provides teachers and parents with a preliminary diagnostic tool composed of the same question items and a much smaller number of items. Therefore the current rating scale is considered to enable both teachers and parents to collect objective and reliable data on children?fs problem behavior for educational uses.