Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.524804
Title: Sensitivity methods for publication bias in a meta-analysis
Author: Malley, Paul
Awarding Body: University of Warwick
Current Institution: University of Warwick
Date of Award: 2009
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
Meta-analysis is the statistical part of a systematic review. Many researchers have used selection functions to model publication bias in a meta-analysis. The main problem with this approach is that it is impossible to verify that the selection function truly represents the selection process, and so the use of selection functions can only be seen as part of a sensitivity analysis. In this thesis we present new methods that involve selection functions that aim to make as few strong assumptions about selection as possible, including the use of a non-parametric permutation test, and the use of a step selection function. We also investigate the use of parametric selection functions and suggest how researchers could use these as part of a sensitivity analysis, by looking at a range of plausible values for the overall selection probability. As part of this sensitivity analysis, we assess the effectiveness of the Bounds method as presented by Henmi et al. Throughout the thesis we illustrate all methods with numerical examples, including a meta-analysis investigating the effects of environmental tobacco smoke on the risk of lung cancer in non-smokers.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.524804  DOI: Not available
Keywords: QA Mathematics
Share: