이 예비연구에서는 통합학급 사정 척도를 개발하기 위하여 1차 예비검사와 2차 예비검사를 실시하였다. 먼저, 통합학급 사정 척도의 문항을 개발하기 위해 국내와 국외의 데이터베이스를 사용하여 발달생태학과 사회생태학, 학교 교실적응과 학교생활만족도, 그리고 자기결정과 소외감에 관한 자료를 수집한 뒤, 130개의 문항을 선정하였다. 1차 예비검사에서는 246명의 대학생들이 응답한 자료로 탐색적 요인분석(EFA)의 베리맥스 회전을 실시하여 5요인의 23문항을 추출하였고, 추출된 5개 요인들은 충분한 설명력을 지닌 것으로 나타났다. 23문항의 전체 신뢰도는 높게 나타났고, 타당도는 전반적으로 양호한 것으로 해석되었다. 추출된 5요인을 가족관계, 교사지원, 급우관계, 소외감, 자기결정행동으로 명명하였고, 구성개념 간에 판별타당성이 확보되었다. 예비척도의 5요인모델의 적합도를 추정한 결과, 통합학급 사정 예비척도의 측정모델은 각각의 표본자료에 잘 부합되는 것으로 나타났다. 전문가 검토에서 중학교와 고등학교의 경우에는 진로결정 문항을 추가하기로 하여, 초등학교 척도는 5요인의 23문항, 중학교와 고등학교의 척도는 6요인의 26문항으로 구성된 통합학급 사정 예비척도를 확정하였다. 초등학생 137명과 중학생 114명, 고등학생 119명을 대상으로 2차 예비검사를 실시한 결과, 척도의 전체 반분신뢰도 계수는 전반적으로 양호한 것으로 평가되었고, 문항내적일치도도 만족할 만한 수준인 것으로 나타났으며 측정의 표준오차도 비교적 낮게 나타났다. 타당도 검증에서 초등학교 척도와 중학교 척도는 전반적으로 양호한 집중타당성을 가지고 있는 것으로, 고등학교 척도의 모든 지표들은 권고수준을 충족시키는 것으로 나타나 집중타당성이 충분한 것으로 평가되었다. 척도의 구성개념 간 상관계수도 상대적으로 낮아, 예비척도의 판별타당성이 높은 것으로 확인되었다. 측정모델 적합도 평가에서 초등학교 5요인 모델과 중학교와 고등학교 6요인 측정모델이 각각의 표본자료에 잘 부합되는 것으로 나타났다. 결론에서, 예비연구의 결과는 효율적인 통합학급 사정 척도의 개발을 위한 연구에 기여할 것으로 기대되었다.
1. Pilot Study 1 The purpose of the first pilot study was to examine to factor structure of the Inclusive Classroom Assessment Scale(the Inclusive Classroom Questionnaire ; Hereafter ICQ) using exploratory and confirmatory factor analyses(EFA and CFAs). The subjects were 246 students from University. Participants included 83 University students with disabilities and 163 non-disabilities students who had been experienced inclusive education. 130 items were selected by investigating a large literature. The measurement structure of the ICQ was evaluated in two steps, EFAs and CFAs. Results of the first pilot study were as follows. First, exploratory factor analyses(EFAs) with principal components extraction and Varimax rotation were performed to derive the most optional structural representation of the data in each sample. In the EFAs, factor loadings above 0.50 were considered salient for the purposes of factor definition. And, using the Kaiser criterion (eigenvalue>1.0) for factor extraction, examination of the scree plots suggested five factors, respectively. Five factors were named family, teacher, classmates, alienation and self-determined behavior. The internal consistency reliabilities derived from this sample were .893 for 23 items(.809≦α≦.893). Second, CFAs based on correlated factors were performed to test the fit of the derived EFA solutions. The ICQ measurement model was characterized by adequate convergent and discriminant validity. Third, the ICQ measurement model, the goodness of fit of measurement model were favorable. 2. Pilot Study 2 The second pilot study examined internal measurement properties of the ICQ, including reliability and validity in three samples of elementary, middle and high school students. The ICQ scale evaluated in the second pilot study represent revised versions of the questionnaires tested in the first stage of the pilot study, using data from collective sample of 135 elementary, 114 middle and 119 high school students in June 2014. The elementary version of the ICQ consisted of 23 items intended to assess five constructs: family(5 items), teacher(5 items), classmates(4 items), alienation(4 items) and self-determined behavior(5 items). The middle and high school version of the ICQ consisted of 26 items intended to assess six constructs: family(5 items), teacher(5 items), friends(4 items), alienation (4 items), self-determined behavior (5 items) and career decision(3 items). Results of the second pilot study were as follows. First, the reliability for elementary were .88(.73≦α≦.82), showed a good reliability. The middle school were .92(.69≦α≦.86), showed a good reliability. And, the high school were .90(.76≦α≦.90), showed a good reliability. Second, the ICQ for elementary was characterized by adequate convergent and discriminant validity. The ICQ for middle school was characterized by adequate convergent and discriminant validity, but AVE of alienations showed a low index. And, the ICQ for high school was characterized by adequate convergent and discriminant validity. Third, the goodness of fit of five constructs model and six constructs model to the model reproduced by empirical data were favorable.