이 연구는 사회적 바람직성 반응 정도는 포함된 문항에 따라 차이가 있다는 전제하에 Rasch 평정척도모형을 적용하여 사회적 체형불안척도 문항들의 평정범주별 가치반응 정도를 계량화하고 개개인의 반응편파성 정도를 통계적으로 조정할 수 있는 방법을 제안하는 데 목적이 있다. 이 연구의 참여자는 서울, 경기지역에 소재하고 있는 12개 대학에 재학 중인 무용을 전공하는 학생들과 일반학과 여학생 980명이다. 이 연구를 위해 사용된 척도는 Hart, Leary, & Rejeski (1989)가 개발한 사회적 체형불안척도(Social Physique Anxiety Scale: SPAS)이다. 수집된 자료는 다음과 같은 2국면(응답자, 문항) Rasch 평정척도모형을 이용하여 5단계 범주별 반응빈도와 모형적합도, 범주통계치를 산출하였다. 평정척도 또는 문항에 대한 사회적 바람직성 반응은 산출된 반응범주별 평균능력추정치에 일정한 서열성이 존재하지 않을 경우에 작용한 것으로 정의하였다(Lessard, 2008). 자료분석을 위해 적용된 프로그램은 WINSTEPS (Linacre, 2002)이다. 분석결과 ‘문항 4, 6, 7, 9, 11, 12’의 반응범주는 척도점수가 증가할수록 피험자의 능력측정치도 서열적으로 증가함으로써 응답자의 사회적 체형불안에 대한 적절한 정보를 제공해주고 있는 것으로 나타났다. 반면 ‘문항 1, 2, 3, 5, 8, 10’은 최소한 하나 이상의 반응범주에서 서열성이 무시된 평균능력추정치를 보임으로써 사회적 바람직성 반응이 개입된 것으로 나타났다. 이와같은 사실을 기초로 할 때 각 문항의 반응범주별 능력추정치를 고려한 분석방법은 사회적 바람직성 반응의 영향을 최소화하는 데 필요한 정보를 제공해 줄 수 있음을 시사하고 있다.
Social desirability response, which is defined as participants’ tendency to give ‘desirable’ answers in response to attitudinal questionnaires, affects the validity of attitudinal questionnaires. The resultant value of attitudinal measurement is therefore likely to be distorted by the social desirability response. The present study was to estimate the degree of social desirability bias by each item’s response category of Social Physique Anxiety Scale (SPAS) and to suggest the way of controlling a social desirability bias of each respondent by applying Rasch rating scale model. The 12-item SPAS was administered to 980 college female students across 10 universities located in Seoul and Gyeonggi areas. The category statistics, such as average measure and threshold, of each item, item difficulty and student’s SPAS level (ability) were estimated from the collected data. Model-data fit was determined by Infit and Out fit statistics (>.7 and ≤1.3). Overall, the model fit the data well with fit statistics between 0.71~1.68 except item 3 ("I wish I was n’t so uptight about my physique/figure: Infit=1.60 and Outfit=1.68). The category statistics revealed that the average measures advance monotonically with category from items 4, 6, 7, 9, 11, and 12. However, category disordering occurred from item 1, 2, 3, 5, 8, and 10 which were influenced by social desirability bias . This analysis by using Rasch model indicated that the ability estimates for each item’s rating scale may provide a more accurate level of social physique anxiety that is less influenced by social desirability response.
Social desirability response, which is defined as participants’ tendency to give ‘desirable’ answers in response to attitudinal questionnaires, affects the validity of attitudinal questionnaires. The resultant value of attitudinal measurement is therefore likely to be distorted by the social desirability response. The present study was to estimate the degree of social desirability bias by each item’s response category of Social Physique Anxiety Scale (SPAS) and to suggest the way of controlling a social desirability bias of each respondent by applying Rasch rating scale model. The 12-item SPAS was administered to 980 college female students across 10 universities located in Seoul and Gyeonggi areas. The category statistics, such as average measure and threshold, of each item, item difficulty and student’s SPAS level (ability) were estimated from the collected data. Model-data fit was determined by Infit and Out fit statistics (>.7 and ≤1.3). Overall, the model fit the data well with fit statistics between 0.71~1.68 except item 3 ("I wish I was n’t so uptight about my physique/figure: Infit=1.60 and Outfit=1.68). The category statistics revealed that the average measures advance monotonically with category from items 4, 6, 7, 9, 11, and 12. However, category disordering occurred from item 1, 2, 3, 5, 8, and 10 which were influenced by social desirability bias . This analysis by using Rasch model indicated that the ability estimates for each item’s rating scale may provide a more accurate level of social physique anxiety that is less influenced by social desirability response.