Effects of Reading Ability on Survey Attrition and Survey Participation Bias

국문 초록 (Abstract)

본 연구는 학생과 학교의 읽기능력이 a) 설문완성 반응(attrition)과 b)설문 참여 편향(bias)과 관련이 있는가를 살펴보았다. 먼저 설문에 참여한 학생으로서 읽기 성적이 있는 학생들을 대상으로 제 2 수준 다층모형 (a two-level Hierarchical Generalized Linear Model)을 사용하여, 학생들의 읽기성취도와, 그리고 학교의 평균읽기 성취도 점수가 설문 응답완성 비율과 관련되어 있는지를 탐색하였다. 집단평균 중심점 교정 모형 (Group mean centered model)과 전체 평균 중심점 교정 모형 (Grand mean centered model)의 비교를 통해, 학생들의 읽기능력이 설문완성도에 관련이 있으나, 학교평균 읽기성취도는 설문완성도에 관련이 없었음을 확인할 수 있었다. 또한 설문에 참여하지 않는 학생들을 이전 그룹에 추가하여 두 모델을 비교함으로써 응답완성과 더불어 설문참여도 편향을 살펴보았다. 학생들의 개인의 읽기 성취도는 그들의 설문완성도에 영향을 미쳤으나, 설문 참여도와는 관련이 없었다. 설문개발과 설문결과의 해석에 관한 것에 대한 함축적 의미에 대한 논의가 있었다.

다국어 초록 (Multilingual Abstract)

This study investigated whether student’s or school’s reading achievement level was related to (a) survey completion attrition and (b) survey participation bias. In order to answer the first question, a sample composed of students who had participated in a large scale survey and for whom reading achievement scores (Test of Achievement and Proficiency score) were available were analyzed using a two level hierarchical generalized linear model (HGLM). By comparing group mean centered models with grand mean centered models, this study found that students' reading achievement was related to their survey completion rates, but there was no indication that schools may have selectively encouraged students with higher reading achievement scores to complete the survey.
In addition, by adding those nonsurvey participating students to the first data set, this study investigated whether the survey completion attrition was related to survey participation bias by comparison with the first model. There was no indication that low achieving students are less likely to participate in the survey. The implication of this study for survey development and interpretation of survey findings were considered.

목차 (Table of Contents)

• ABSTRACT
• Ⅰ. Introduction
• Ⅱ. Methods
• Ⅲ. Results
• Ⅳ. Discussion
• References
• 요약
• Appendix

참고문헌 (Reference)

1 "Public use data set: 2001 Survey of students and teachers user’s manual [On-line" 2004

2 "Nonresponse in Household Interview Surveys" John Wiley and Sons 1998

3 "Nonresponse bias form incomplete interview in the national survey of America’s Families [On-line]" 2001

4 "Non-response bias in prospective studies of patients and health care professionals" 2 : 203-212, 1999

5 "Mail Surveys" Sage Publications 1995

6 "Item analysis by the hierarchical generalized linear model" 38 : 79-93, 2001

7 "How generalizable are the effects of smoking prevension programs? Refusal skills training and parent messages in a teacher-administered program Journal of Behavioral Medicine" 613-28, 1987

8 "Hierarchical linear Models" Sage Publications. 2002

9 "HLM 5: Hierarchical Linear and Nonlinear Modeling HLM 5" Scientific Software International illi (illi): -5, 2001

10 "Design and analysis issues for field settings" 1979

11 "Attrition in prevention research Journal of Behavioral Medicine" 261-75, 1985

12 "Assessing the quality of longitudinal surveys" 3-58, 1988

13 "An introduction to sample selection bias in sociological data" 386-98, 1983

14 "Adjusting for attrition in school-based samples" 21 : 554-567, 1997

15 "A multilevel assessment of differential item functioning" Montreal 1999

16 "A User’s Guide to BIGSTEPS" MESA press 1997