TY - JOUR
T1 - Highly efficient aggregate unbiased estimating functions approach for correlated data with missing at random
AU - Qu, Annie
AU - Lindsay, Bruce G.
AU - Lu, Lin
N1 - Funding Information:
Annie Qu is Associate Professor, Department of Statistics, University of Illinois at Urbana–Champaign, Champaign, IL 61820 (E-mail: anniequ@ illinois.edu). Bruce G. Lindsay is Willaman Professor, Department of Statistics, The Pennsylvania State University, University Park, PA 16802 (E-mail: [email protected]). Lin Lu is Statistician, Quintiles, Inc., Overland Park, KS 66211 (E-mail: lin_ [email protected]). This work is supported in part by National Science Foundation grants. The authors thank constructive suggestions from the associate editor, two referees, and the editor.
PY - 2010/3
Y1 - 2010/3
N2 - We develop a consistent and highly efficient marginal model for missing at random data using an estimating function approach. Our approach differs from inverse weighted estimating equations (Robins, Rotnitzky, and Zhao 1995) and the imputation method (Paik 1997) in that our approach does not require estimating the probability of missing or imputing the missing response based on assumed models. The proposed method is based on an aggregate unbiased estimating function approach, which does not require the likelihood function; however, it is equivalent to the score equation if the likelihood is known. The aggregate-unbiased approach is based on the best linear approximation of efficient scores from the full dataset. We provide comparisons of the three approaches using simulated data and also a human immunodeficiency virus (HIV) data example.
AB - We develop a consistent and highly efficient marginal model for missing at random data using an estimating function approach. Our approach differs from inverse weighted estimating equations (Robins, Rotnitzky, and Zhao 1995) and the imputation method (Paik 1997) in that our approach does not require estimating the probability of missing or imputing the missing response based on assumed models. The proposed method is based on an aggregate unbiased estimating function approach, which does not require the likelihood function; however, it is equivalent to the score equation if the likelihood is known. The aggregate-unbiased approach is based on the best linear approximation of efficient scores from the full dataset. We provide comparisons of the three approaches using simulated data and also a human immunodeficiency virus (HIV) data example.
KW - Generalized estimating equations
KW - HIV data
KW - Imputation method
KW - Inverse weighted estimating equation
KW - Missing at random
KW - Missing completely at random
UR - http://www.scopus.com/inward/record.url?scp=77952574811&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=77952574811&partnerID=8YFLogxK
U2 - 10.1198/jasa.2009.tm08506
DO - 10.1198/jasa.2009.tm08506
M3 - Article
AN - SCOPUS:77952574811
SN - 0162-1459
VL - 105
SP - 194
EP - 204
JO - Journal of the American Statistical Association
JF - Journal of the American Statistical Association
IS - 489
ER -