Article type
Year
Abstract
Background: Systematic reviews (SRs) often include 'grey literature' (GL), such as unpublished or non-indexed literature. Researchers must balance time and resources without compromising a search for 'all' evidence. A 1993 survey determined that researchers and editors differed in their views about including unpublished data in meta-analyses.
Objectives: To gauge current views of systematic reviewers, methodologists and journal editors concerning several issues relating to GL.
Methods: We identified 273 systematic reviewers from a cohort of published SRs, 89 methodologists from reports in the Cochrane Methodology Register and 172 editors from journals publishing the SRs or methods papers. We administered a survey by email and facsimile to all of these individuals. Data were analysed descriptively.
Results: The survey response rate was 58.3%. Most respondents felt GL should probably or definitely be eligible for inclusion in SRs; editors were less inclusive (68.6%) compared to systematic reviewers or methodologists (85.3%). Respondents felt the quality and accessibility of GL was probably or definitely worse than published literature. Nevertheless, 79.8% stated they always or sometimes sought GL and 66.5% would search for GL in the conduct of a future SR. If they were an editor presented with an SR containing GL for publication, most respondents stated they would request that peer reviewers comment on the inclusion of GL. No respondent would reject the SR due to GL, however editors were more likely (15.9%) than systematic reviewers or methodologists (3.8%) to instruct authors to delete GL reports, even given a favourable review. Most respondents (87.2%) believed that GL data included in an SR should not bear on subsequent full publication of original research.
Conclusions: Editors appear less favourable towards GL than systematic reviewers or methodologists although differences have decreased compared to previous data. These variations relate in part to differences in SR experience or previous familiarity with GL.
Objectives: To gauge current views of systematic reviewers, methodologists and journal editors concerning several issues relating to GL.
Methods: We identified 273 systematic reviewers from a cohort of published SRs, 89 methodologists from reports in the Cochrane Methodology Register and 172 editors from journals publishing the SRs or methods papers. We administered a survey by email and facsimile to all of these individuals. Data were analysed descriptively.
Results: The survey response rate was 58.3%. Most respondents felt GL should probably or definitely be eligible for inclusion in SRs; editors were less inclusive (68.6%) compared to systematic reviewers or methodologists (85.3%). Respondents felt the quality and accessibility of GL was probably or definitely worse than published literature. Nevertheless, 79.8% stated they always or sometimes sought GL and 66.5% would search for GL in the conduct of a future SR. If they were an editor presented with an SR containing GL for publication, most respondents stated they would request that peer reviewers comment on the inclusion of GL. No respondent would reject the SR due to GL, however editors were more likely (15.9%) than systematic reviewers or methodologists (3.8%) to instruct authors to delete GL reports, even given a favourable review. Most respondents (87.2%) believed that GL data included in an SR should not bear on subsequent full publication of original research.
Conclusions: Editors appear less favourable towards GL than systematic reviewers or methodologists although differences have decreased compared to previous data. These variations relate in part to differences in SR experience or previous familiarity with GL.
PDF