透過您的圖書館登入
IP:18.117.183.172
  • 期刊
  • OpenAccess

配對數據分析模式之應用探討

An Application of Model Matched-Paired Data Analysis

摘要


在病例對照研究中,通常是以McNemar檢定來討論隨機配對變數間的關聯性,惟若引入共變數進行配對數據模式分析,為了避免參數估計值偏差的結果發生,須利用條件最大概似法進行模式分析,而非一般所指的最大概似法。本研究旨在探討配對研究資料在公共衛生領域之分析應用,主要陳述如何進行配對數據資料模式分析,擬就Cox存活模式、分層邏輯斯模式及無截距邏輯斯模式分別詮釋1:1及m:n配對資料。其中Cox存活模式、分層邏輯斯模式均可就1:1及m:n配對資料進行分析,然不同模式在分析時所需資料格式有所不同。另一方面,無截距邏輯斯模式可就共變數之差相對應的組合配對進行分析,惟其僅適用於1:1配對資料。本文以實例配合SAS程式進行相關資料分析,並提供有關Cox存活模式、分層邏輯斯模式及無截距邏輯斯模式配對資料SAS程式,以深入簡出的方式詳細探討各類配對數據模式之異同及其適用範圍,並提供SAS程式便於相關領域研究人員參考。

並列摘要


In matched case-control study, McNemar test is used to measure the association between match paired random variables. On the other hand, conditional maximum likelihood is applied for analyzing the match-paired related model once the covariates are introduced. This study uses 2 examples describing how to apply Cox proportional hazard model, stratified logistic regression model and logistic regression model without intercept to fit 1:1 and m:n matched-paired data. The 1:1 and m:n matched-paired data could be analyzed by Cox proportional hazard model and stratified logistic regression model, while the logistic regression model without intercept transforms each matched pair into a single observation and its corresponded explanatory variables given by the difference for the value of each pair of case and control could be used for 1:1 matched-paired data only. The main purpose of this study is to interpret how to use the method of conditional maximum likelihood analyzing match-paired data. Comparisons of SAS programs among Cox proportional hazard model, stratified logistic regression model and logistic regression model without intercept are also provided. It would be useful for people to handle the problem about matched paired data analysis efficiently.

參考文獻


沈葆聖(2002)。SAS 統計軟體與資料分析。台中:滄海書局。
Agresti, A.(2002).Categorical Data Analysis.New York:John Wiley & Sons.
Cox, D. R.(1970).The Analysis of Binary Data.London:Methuen.
Hosmer, D. W.,Lemeshow, S.(1989).Applied Logistic Regression.New York:John Wiley & Sons.
Kleinbaum, D. G.(1994).Logistic Regression.New York:Springer.

延伸閱讀