서브메뉴
검색
Causes of Nonlinear Metrics in Item Response Theory Models and Implications for Educational Research
Causes of Nonlinear Metrics in Item Response Theory Models and Implications for Educational Research
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20250211153130
- ISBN
- 9798346875741
- DDC
- 371
- 저자명
- Liao, Xiangyi.
- 서명/저자
- Causes of Nonlinear Metrics in Item Response Theory Models and Implications for Educational Research
- 발행사항
- [Sl] : The University of Wisconsin - Madison, 2024
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2024
- 형태사항
- 125 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 86-06, Section: A.
- 주기사항
- Advisor: Bolt, Daniel M.
- 학위논문주기
- Thesis (Ph.D.)--The University of Wisconsin - Madison, 2024.
- 초록/해제
- 요약Educational research outcomes frequently rely on an assumption that measurement metrics have interval-level properties. While most investigators know enough to be suspicious of interval-level claims, and in some cases even question their findings given such doubts, there is a lack of understanding regarding the measurement conditions that create nonlinear metric distortions. This dissertation aims to provide a framework that can assist researchers in thinking about the scalar properties of educational test score metrics. Following an overview of educational research studies and measurement applications that depend heavily on the interval level metric assumption, along with associated commentaries that have noted this dependence, as well as reasons to question the latent metrics produced by item response theory (IRT) models, this dissertation presents three studies: (1) an exploration of the psychological response process, particularly the proficiency-related guesses and slips in multiple-choice items, as a source of item characteristic curve (ICC) asymmetry in IRT models and thus a source of metric distortion; (2) an investigation of conditions of test multidimensionality, specifically where dimensionality correlates with item difficulty, which similarly causes nonlinear distortions of measurement metrics; and (3) an analysis of the practical impacts on vertical scaling when nonlinear metrics are not well attended to. These studies extend IRT models to incorporate nonlinear latent metrics under various conditions. The findings aim to inform the development of measurement instruments, or in some cases alternative measurement practices, that can come closer to achieving the desired interval-level properties of the measurement scale.
- 일반주제명
- Educational evaluation
- 일반주제명
- Educational philosophy
- 키워드
- Dimensionality
- 기타저자
- The University of Wisconsin - Madison Educational Psychology
- 기본자료저록
- Dissertations Abstracts International. 86-06A.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008250123s2024 us c eng d■001000017165156
■00520250211153130
■006m o d
■007cr#unu||||||||
■020 ▼a9798346875741
■035 ▼a(MiAaPQ)AAI31767741
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a371
■1001 ▼aLiao, Xiangyi.
■24510▼aCauses of Nonlinear Metrics in Item Response Theory Models and Implications for Educational Research
■260 ▼a[Sl]▼bThe University of Wisconsin - Madison▼c2024
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2024
■300 ▼a125 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 86-06, Section: A.
■500 ▼aAdvisor: Bolt, Daniel M.
■5021 ▼aThesis (Ph.D.)--The University of Wisconsin - Madison, 2024.
■520 ▼aEducational research outcomes frequently rely on an assumption that measurement metrics have interval-level properties. While most investigators know enough to be suspicious of interval-level claims, and in some cases even question their findings given such doubts, there is a lack of understanding regarding the measurement conditions that create nonlinear metric distortions. This dissertation aims to provide a framework that can assist researchers in thinking about the scalar properties of educational test score metrics. Following an overview of educational research studies and measurement applications that depend heavily on the interval level metric assumption, along with associated commentaries that have noted this dependence, as well as reasons to question the latent metrics produced by item response theory (IRT) models, this dissertation presents three studies: (1) an exploration of the psychological response process, particularly the proficiency-related guesses and slips in multiple-choice items, as a source of item characteristic curve (ICC) asymmetry in IRT models and thus a source of metric distortion; (2) an investigation of conditions of test multidimensionality, specifically where dimensionality correlates with item difficulty, which similarly causes nonlinear distortions of measurement metrics; and (3) an analysis of the practical impacts on vertical scaling when nonlinear metrics are not well attended to. These studies extend IRT models to incorporate nonlinear latent metrics under various conditions. The findings aim to inform the development of measurement instruments, or in some cases alternative measurement practices, that can come closer to achieving the desired interval-level properties of the measurement scale.
■590 ▼aSchool code: 0262.
■650 4▼aEducational tests & measurements
■650 4▼aEducational evaluation
■650 4▼aEducational philosophy
■653 ▼aDimensionality
■653 ▼aItem characteristic curve
■653 ▼aEducational research
■653 ▼aMeasurement metrics
■653 ▼aInterval-level claims
■690 ▼a0288
■690 ▼a0443
■690 ▼a0998
■71020▼aThe University of Wisconsin - Madison▼bEducational Psychology.
■7730 ▼tDissertations Abstracts International▼g86-06A.
■790 ▼a0262
■791 ▼aPh.D.
■792 ▼a2024
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17165156▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
