상세검색
최근 검색어 전체 삭제
다국어입력
즐겨찾기0
153775.jpg
KCI등재 학술저널

문화간 감수성 측정도구 단축판 구성

Rasch 평정척도모형의 적용

  • 25

본 연구의 목적은 문화간 감수성 함양을 위해 설계된 대학 중국문화수업에 적합한 검사도구를 제안하고자 Rasch의 평정척도모형을 사용하여 양호한 문항을 선별하고 이 문항들로만 구성된 단축판을 구성하는데 있다. 이를 위해 대학생 225명의 응답자료를 분석하여, 응답범주의 적절성, 문항의 내·외적합도, 문항난이도, 피험자 능력 분포 등을 분석하였다. 연구결과, 응답범주는 5점 척도보다 4점 척도로 일 때, 문항 및 피험자 분리지수, 검사 신뢰도가 가장 높았으며, 문항난이도 분포 및 피험자 능력 분포가 적합하였다. 범주확률곡선 분석에서도 응답범주별 경계점수의 차이가 뚜렷하였다. 최종문항 선정과정에서 과적합보다 부적합 문항을 우선 삭제하였으며 해당 문항의 난이도가 그 영역의 다른 문항에 의해 측정되지 않는 문항을 재검토하였다. 최종적으로 선정된 18문항을 대상으로 단축판을 구성하였고 문항난이도를 고려하여 문항의 순서를 재배열하였다. 끝으로 단축판 검사의 활용방안과 본 연구의 제한점을 제시하였다.

The Intercultural Sensitivity Scale is a 5-point Likert scale consisting of 24 items for the assessment of the following five factors: relational engagement, regard for cultural diversity, relational certainty, relational satisfaction, and relational carefulness. With the aim of constructing a shortened version of the Intercultural Sensitivity Scale, this study examined which response category among 5-point, 4-point, and 3-point scales can best distinguish the responses of examinees and assess them with reliability, based on the Intercultural Sensitivity Scale test response data collected from 225 individuals and using the Rasch Rating Scale Model. The results are as follows. Item and person separation indices and reliability were highest with a 4-point response category (strongly disagree-uncertain-agree-strongly agree). In terms of the infit and outfit of items, the number of overfitting or underfitting items decreased from 10 to 8 on the 4-point scale. Overall, items did not significantly deviate from item difficulty and examinee ability distributions on the 4-point scale. On the category probability curve as well, the boundary scores for each response category according to examinee ability levels showed distinct differences on the 4-point scale. The selection of deleted items was made following the principle of excluding underfitting items first before overfitting items. However, if the difficulty of underfitting items are not measured by other items in the same area, the items were not deleted. With regard to overfitting items as well, if the difficulty of these items was not measured by other questions in the same area, they were not deleted. The shortened version was constructed with 18 finally selected items, and the order of these items was reset in consideration of difficulty. Lastly, this study presented suggestions regarding the use of this shortened version while clarifying its limitations.

1. 서론

2. 연구방법

3. 연구결과

4. 논의

로딩중