확인적 잠재계층 요인분석을 활용한 응답편파모형 탐색 = Exploring Response Bias Model: Application of a Confirmatory Latent-Class Factor Analysis

국문 초록 (Abstract)

본 연구에서는 평정형 응답척도를 사용한 심리측정에서 체계적인 오차의 유발 요인으로 고려되는 응답편파(response bias) 효과를 측정 및 통제하기 위한 방안으로 확인적 잠재계층 요인분석(confirmatory latent-class factor analysis: LCFA)을 적용하였다. 구체적으로 TIMSS 2011의 8학년 학생 설문 중 수학 및 과학 가치 10개 문항의 응답에서 가장 빈번히 발생하는 응답편파인 묵인반응(ARS) 및 극단반응(ERS) 양식를 탐색하기 위하여, Moors(2003)에 의해 제안된 응답편파모형을 수정한 6개의 LCFA 모형을 설정하고 비교하였다. 연구결과, 응답편파요인의 유무 및 요인의 개수 등에 따른 대안모형 중 최적의 모형으로는 ARS와 ERS 및 2개의 가치모형이 포함된 4요인 LCFA 모형이 선정되었다. 즉, 수학 및 과학 가치 측정에 있어서 응답편파의 효과가 확인되었고, 이는 양적지수에 기반한 ARS 및 ERS 지표와의 타당화 및 모형의 모수치 등을 통하여 타당화 되었다. 또한 국가(한국, 싱가포르 및 미국) 변인을 투입하여, 집단 간 차이를 탐색한 결과, 수학 및 과학가치의 내용요인 뿐 아니라 ARS와 ERS의 응답편파 요인에 있어서도 통계적으로 유의한 집단 차이가 파악되었다. 마지막으로 본 연구는 집단 간 평균 비교 및 변인 간 인과적인 모형을 활용한 사회과학 연구 결과의 타당성 제고를 위하여, 응답편파의 측정 및 통제의 중요성을 강조하며, 또한 LCFA 모형의 추가적인 적용 및 보다 정교화된 응답편파의 탐색방안에 대한 후속연구의 필요성을 논의하였다.

다국어 초록 (Multilingual Abstract)

Many empirical studies confirmed that response biases can seriously distort psychological measurement such as attitudes or personality scales and subsequent causal models. This study would apply a confirmatory latent-class factor analysis(LCFA) as one of the statistical methods for detection of and correction for response biases such as acquiescence response style(ARS) and extreme response style(ERS). The response bias model originally proposed by Moors(2003), which had several advantages especially for dealing with both substantive and response bias factors, was modified for this study. Using data of three countries(Korea, Singapore and US) from TIMSS 2011, response biases such as ARS and ERS in answering Likert-type questions on math and science value were explored. Several LCFA models in terms of the number of factors or including response bias factors were compared and the 4-factor model as a best fitting model was finally selected. The model specifications of the selected model with two substantive and two response bias factors are analyzed and interpreted. Also, the effects of county group on math/science value and response bias factors were examined. The results showed a strong evidence for the presence of response bias(ARS and ERS) and three countries differed not only in their math/science value but also in ARS and ERS. These findings suggested the importance of detecting and controlling for response bias in psychological measurement. Implications for future research are discussed.

목차 (Table of Contents)

• 요약
• Ⅰ. 서론
• Ⅲ. 연구 방법
• Ⅳ. 연구 결과
• Ⅴ. 논의
• 참고문헌

참고문헌 (Reference)

1 신학경, "혼합라쉬모형을 적용한 TIMSS 2011 수학흥미 척도의 응답양식 탐색" 한국교육평가학회 27 (27): 429-448, 2014

2 김석호, "응답자의 성격특성과 응답스타일" 한국조사연구학회 12 (12): 51-76, 2011

3 Bachman, J. G., "Yea-saying, nay-saying, and going to extremes : Black-white differences in response styles" 48 (48): 491-509, 1984

4 Mõttus, R., "The effect of response style on self-reported conscientiousness across 20 countries" 38 (38): 1423-1436, 2012

5 Moors, G., "The effect of response style bias on the measurement of transformational, transactional, and laissez-faire leadership" 21 (21): 271-298, 2012

6 Weijters, B., "The effect of rating scale format on response styles : The number of response categories and response category labels" 27 (27): 236-247, 2010

7 Billiet, J. B., "Testing the stability of an acquiescence style factor behind two interrelated substantive variables in a panel design" 36 (36): 542-562, 2008

8 Baumgartner, H., "Response styles in marketing research : A cross-national investigation" 38 (38): 143-156, 2001

9 Billiet, J. B., "Modeling acquiescence in measurement models for two balanced sets of items" 7 (7): 608-628, 2000

10 Paulhus, D. L., "Measures of personality and social psychological attitudes" Academic Press 17-59, 1991

11 Bolt, D. M., "Measurement and control of response styles using anchoring vignettes : A model-based approach" 19 (19): 528-541, 2014

12 Vermunt, J. K., "LG-Syntax User's Guide: Manual for Latent GOLD 5.0 Syntax Module" Statistical Innovations Inc 2013

13 Austin, E. J., "Individual differences in Response scale use : Mixed Rasch modelling of Responses to NEO-FFI items" 40 (40): 1235-1245, 2006

14 Greenleaf, E. A., "Improving rating scale measures by detecting and correcting Bias components in some Response styles" 29 (29): 176-188, 1992

15 Akaike, H., "Factor analysis and AIC" 52 (52): 317-332, 1987

16 Schwarz, G., "Estimating the dimension of a model" 6 (6): 461-464, 1978

17 Moors, G., "Diagnosing response style behavior by means of a latent-class factor approach. Socio-demographic correlates of gender role attitudes and perceptions of ethnic discrimination reexamined" 37 (37): 277-302, 2003

18 Morren, M., "Dealing with extreme response style in cross-cultural research : A restricted latent class factor analysis approach" 41 : 1-13, 2011

19 Buckley, J., "Cross-national response styles in international educational assessments: Evidence from PISA 2006"

20 Agresti, A., "Categorical data analysis" John Wiley & Sons 2002

21 Raftery, A. E., "Bayesian model selection in social research" 25 : 111-163, 1995

22 Cheung, G. W., "Assessing extreme and Acquiescence Response sets in cross-cultural research using structural equations modeling" 31 (31): 187-212, 2000

23 Meisenberg, G., "Are acquiescent and extreme response styles related to low intelligence and education?" 44 (44): 1539-1550, 2008

24 Lin, T. H., "A comparison of model selection indices for nested latent class models" 12 : 239-259, 2006