統計学輪講（第１１回）

日時	  2001年6月19日(火)    15時00分〜16時40分
場所	  経済学部5階視聴覚室
講演者    Myunghee Cho Paik(Columbia University)
          大橋靖雄(医学部)(コーディネーター)
演題　    Matched Case-Control Data Analysis with Non-ignorably 
          Missing Covariates

概要
We consider methods for analyzing matched case-control data when some
covariates are completely observed but other covariates are missing
for some subjects.  In complete-record (CR) analyses for
matched case-control studies, completely observed subjects are forced
to be discarded, if none of their matching cases or controls are completely
observed.  More improtantly, the CR analysis yields biased estimates unless
the data are missing completely at random.
We propose a method to handle missing covariates in the context of
conditional likelihood.  The method handles the cases when data are missing
at random (MAR) or missing non-ignorably.  The method is applied to the
Northern Manhattan Stroke Study.