Estimating the bias in meta analysis estimates for continuous data with non-random missing study variance

When study variances are not reported or ‘missing”, it is common practice in meta analysis to assume that the missing variances are missing completely at random (MCAR). In practice, however, it is possible that the variances are not missing completely at random (NMAR). In this paper, we examine, ana...

Full description

Bibliographic Details
Main Author: Nik Idris, Nik Ruzni
Format: Proceeding Paper
Language:English
Published: 2010
Subjects:
Online Access:http://irep.iium.edu.my/5551/
http://irep.iium.edu.my/5551/1/skskm2010_manu.pdf
_version_ 1848776523441504256
author Nik Idris, Nik Ruzni
author_facet Nik Idris, Nik Ruzni
author_sort Nik Idris, Nik Ruzni
building IIUM Repository
collection Online Access
description When study variances are not reported or ‘missing”, it is common practice in meta analysis to assume that the missing variances are missing completely at random (MCAR). In practice, however, it is possible that the variances are not missing completely at random (NMAR). In this paper, we examine, analytically, the biases introduce in the meta analysis estimates when the missing study variances occur with non-random missing mechanism (MNAR), namely, when the magnitude of the missing variances are mostly larger than those that are reported. In meta analysis, this is more likely to occur in studies which carry larger variances. We looked at two common approaches in handling this problem, namely, the missing variances are imputed using the mean imputation, and the studies with missing study-variances are omitted from the analysis. The results suggest that for the estimate of the variance of the effect size, if the magnitude of the study-variances that are missing are mostly larger than those that are reported, the variance of the effect size will be underestimated. Thus under MNAR, the mean imputation gives false impression of precision as the estimated variance of the overall effect is too small. On the other hand, if the missing variances are mostly smaller, the variance will be overestimated.
first_indexed 2025-11-14T14:31:27Z
format Proceeding Paper
id iium-5551
institution International Islamic University Malaysia
institution_category Local University
language English
last_indexed 2025-11-14T14:31:27Z
publishDate 2010
recordtype eprints
repository_type Digital Repository
spelling iium-55512011-11-22T00:39:18Z http://irep.iium.edu.my/5551/ Estimating the bias in meta analysis estimates for continuous data with non-random missing study variance Nik Idris, Nik Ruzni HA29 Theory and method of social science statistics When study variances are not reported or ‘missing”, it is common practice in meta analysis to assume that the missing variances are missing completely at random (MCAR). In practice, however, it is possible that the variances are not missing completely at random (NMAR). In this paper, we examine, analytically, the biases introduce in the meta analysis estimates when the missing study variances occur with non-random missing mechanism (MNAR), namely, when the magnitude of the missing variances are mostly larger than those that are reported. In meta analysis, this is more likely to occur in studies which carry larger variances. We looked at two common approaches in handling this problem, namely, the missing variances are imputed using the mean imputation, and the studies with missing study-variances are omitted from the analysis. The results suggest that for the estimate of the variance of the effect size, if the magnitude of the study-variances that are missing are mostly larger than those that are reported, the variance of the effect size will be underestimated. Thus under MNAR, the mean imputation gives false impression of precision as the estimated variance of the overall effect is too small. On the other hand, if the missing variances are mostly smaller, the variance will be overestimated. 2010-12-21 Proceeding Paper PeerReviewed application/pdf en http://irep.iium.edu.my/5551/1/skskm2010_manu.pdf Nik Idris, Nik Ruzni (2010) Estimating the bias in meta analysis estimates for continuous data with non-random missing study variance. In: Seminar Kebangsaan Sains Komputer Dan Matematik 2010, 21-22 December 2010, Kota Kinabalu, Sabah. http://www.skskm.net
spellingShingle HA29 Theory and method of social science statistics
Nik Idris, Nik Ruzni
Estimating the bias in meta analysis estimates for continuous data with non-random missing study variance
title Estimating the bias in meta analysis estimates for continuous data with non-random missing study variance
title_full Estimating the bias in meta analysis estimates for continuous data with non-random missing study variance
title_fullStr Estimating the bias in meta analysis estimates for continuous data with non-random missing study variance
title_full_unstemmed Estimating the bias in meta analysis estimates for continuous data with non-random missing study variance
title_short Estimating the bias in meta analysis estimates for continuous data with non-random missing study variance
title_sort estimating the bias in meta analysis estimates for continuous data with non-random missing study variance
topic HA29 Theory and method of social science statistics
url http://irep.iium.edu.my/5551/
http://irep.iium.edu.my/5551/
http://irep.iium.edu.my/5551/1/skskm2010_manu.pdf