{"id":377,"date":"2017-05-12T11:06:55","date_gmt":"2017-05-12T11:06:55","guid":{"rendered":"http:\/\/jsr.isrt.ac.bd\/?post_type=article&p=377"},"modified":"2017-05-12T11:07:02","modified_gmt":"2017-05-12T11:07:02","slug":"likelihood-ratio-test-nonignorable-missingness-incomplete-binary-longitudinal-data","status":"publish","type":"article","link":"http:\/\/jsr.isrt.ac.bd\/article\/likelihood-ratio-test-nonignorable-missingness-incomplete-binary-longitudinal-data\/","title":{"rendered":"A likelihood ratio test for nonignorable missingness in incomplete binary longitudinal data"},"content":{"rendered":"
Missing data are common in many clinical studies. When missingness is non-
\nignorable, a full likelihood analysis of the data requires incorporating a missing
\ndata model into the observed data likelihood function. In this article, we study
\nthe bias of the ML estimator when the corresponding maximum likelihood is ob-
\ntained under a misspeci\fed missing data model. We further explore a likelihood
\nratio statistic for testing the missing data mechanism in binary longitudinal data.
\nThe empirical level and power of the test are investigated in small simulations.
\nWe also present an example using some real data obtained from a longitudinal
\nstudy.<\/p>\n
<\/p>\n