Search
Search

연구성과물 검색 타이틀 이미지

HOME ICON HOME > Search by Achievements Type > Reports View

Reports Detailed Information

https://www.krm.or.kr/krmts/link.html?dbGubun=SD&m201_id=10007560&local_id=10012799
이분변수자료의 분석을 위한 로지스틱 이원다층모형
Reports NRF is supported by Research Projects( 이분변수자료의 분석을 위한 로지스틱 이원다층모형 | 2004 Year | 강상진(연세대학교) ) data is submitted to the NRF Project Results
Researcher who has been awarded a research grant by Humanities and Social Studies Support Program of NRF has to submit an end product within 6 months(* depend on the form of business)
사업별 신청요강보기
  • Researchers have entered the information directly to the NRF of Korea research support system
Project Number B00414
Year(selected) 2004 Year
the present condition of Project 종료
State of proposition 재단승인
Completion Date 2006년 05월 13일
Year type 결과보고
Year(final report) 2006년
Research Summary
  • Korean
  • 다층모형은 지난 20여년간 교육연구방법의 중핵적인 통계방법으로 발전한 보편적인 통계방법으로서, 다층모형 방법론자들은 다양한 다층구조를 모형에 반영하고, 동시에 다양한 유형의 종속변수를 분석하기 위한 다층모형들을 개발하여 왔다. 이 연구에서 제시하는 로지스틱 이원다층모형은 이러한 경향을 반영한 결과로서, 전통적인 다층모형과 달리 수준-2(macro-level)에서 무선효과를 갖는 집단단위가 하나가 아니고 두 가지이며, 수준-1(micro-level)의 관찰단위는 두 가지 수준-2 요인이 교차하여 생성하는 칸에 내재하는 구조의 자료를 분석하기에 적절한 다층모형으로서 이원다층모형이라고도 한다. 동시에 이 연구에서 제시하는 다층모형은 수준-1에서의 종속변수가 연속변수가 아닌 이분변수인 경우의 자료분석을 위한 로지스틱 이원 다층모형이다. 이 논문은 모형의 개념을 일반화 선형모형(generalized linear model)의 관점에서 로짓 연결함수 사용한 다층모형의 개념으로 설명하였다. 추정방법은 PQL 방법과 경험적 베이시안 관점에서 제시하였고, 모형의 모수 추정을 위한 계산 알고리즘은 EM- 알고리즘을 적용하였으며, 예시자료는 서울시에 소재한 54개 중학교의 교사들을 대상으로 수집한 학교의 사회적 환경자료를 분석하였다.
  • English
  • Multielvel models have become the core of eductional research methods in the past two decades. Researchers studying the multilevel modeling have challenged to account for various nested sturctures of educational data and for the various type of outcome variables. This study is to present a new multilevel model to account for more complex data structure where level-1 units are nested within the cells cross-classified by two macro units having random effects. On the other hand, the model addresses the estimation method when the data have binary outcomes given the cross-classified multilevel structure. The conceptualization of the model is drescribed in the perspective of generalized linear model by using logit link function and the multilevel model with crossed random effects at level-2. Estimation method is briefly presented from the Penalized Quasi Likelihood method and from an empirical Bayses view point. The computation is implemented via the EM algorithm and an illustrative application of the model was performed by analyzing the data of 1708 teachers from 54 middle schools.
Research result report
  • Abstract
  • This study is to present a new multilevel model to account for more complex data structure where level-1 units are nested within the cells cross-classified by two macro units having random effects. On the other hand, the model addresses the estimation method when the data have binary outcomes given the cross-classified multilevel structure. The conceptualization on the model is described in the perspective of generalized linear model by using logit link function and the multilevel model with crossed random effects at level-2. Estimation method is briefly presented from the Penalized Quasi Likelihood method and from an empirical Bayes viewpoint. The computation is implemented via the EM algorithm and an illustrative applicatoin of the model was performed by analyzing the data of 1708 teachers from 54 middle schools.
  • Research result and Utilization method
  • The Logistic Crossed Multilevel Model (here after LCMM) has been developed with the computing program via the Gauss software system that uses matrix language. The newly developed LCMM can be estimated via the author's program only. This computing program needs to adopt macro program so that other researchers can use the LCMM in a more comfortable way. The LCMM can be used for any studies that use binary data given the cross-classified multilevel structure.
  • Index terms
  • multilevel model, cross-classified multilevel model, logistic multilevel model, Penalized Quasi Likelihood method, EM algorithm
  • List of digital content of this reports
데이터를 로딩중 입니다.
  • This document, it is necessary to display the original author and you do not have permission
    to use copyrighted material for-profit
  • In addition , it does not allow the change or secondary writings of work
데이터 이용 만족도
자료이용후 의견
입력
트위터 페이스북
NRF Daejeon
(34113) 201, Gajeong-ro, Yuseong-gu, Daejeon, Korea
Tel: 82-42-869-6114 / Fax: 82-42-869-6777
NRF Seoul
(06792) 25, Heonreung-ro, Seocho-gu, Seoul, Korea
Tel: 82-2-3460-5500 / Fax: 82-2-3460-5759
KRM Help Center
Tel : 042-869-6086 Fax : 042-869-6580
E-mail : krmcenter@nrf.re.kr