|
Different Types of Missing in Longitudinal Data and the Likelihood-base Methods Applied in their Analysis
|
F Zayeri *   , Ar Akbarzadeh baghban , M Kazemzadeh , M Yaseri , Am Abbasi |
|
|
|
Abstract: (15412 Views) |
| Missing values are frequently seen in data sets of research studies conducted in diff-erent sciences such as medicine and esp-ecially in longitudinal studies in which every individual are exposed to the repeated measures over time. In the last few decades, a vast majority of statistical activities has been done in this area, including the areas of concepts, issues, and theoretical and soft-ware methods. Despite the widespread use of the results of these statistical activities, the researchers, in many cases, have been seen to have a vague impression from these concepts which results in inaccurate infer-ences. Therefore, given the importance of the issue and the need for the scientific community to know these issues correctly and accurately, the current study is set to review and compare the concepts such as missing data patterns and mechanism, as well as the existing models in analyzing longitudinal data with missing values. Furt-hermore, their application will be explored in the data obtained from a clinical trial of addiction treatment with a continuous res-ponse variable. |
|
| Keywords: missing completely at random, missing at random, missing not at random, selection model, pattern-mixture model, shared parameters model |
|
|
Full-Text [PDF 403 kb]
(6997 Downloads)
|
Type of Study: Research |
Subject:
biostatistics
Received: 2013/03/16 | Accepted: 2013/03/17 | Published: 2013/03/17
|
|
|
|
|
|
|
| Add your comments about this article |
|
|
|
