Adjusting for publication bias: modelling the selection process

Abstract

Weighted distributions offer a flexible approach that allows the potential to modify the selection function to incorporate other factors. Methods that adjust combined estimates should not be used to provide an alternative answer but to consider the robustness of the combined estimate to publication bias.

Statistical methods for dealing with publication bias in meta‐analysis

“…Preston et al . used the logistic weight function to adjust publication bias . They assumed the f ( x ; θ ) followed a normal distribution.…”

“…The new weight function could be denoted as Preston et al . also suggested to use the half normal distribution w ( x i )= exp{− β [1−Φ( x i / σ i )] 2 } and negative exponential distribution w ( x i )= exp{− β [1−Φ( x i / σ i )]} as the weight function . The procedures were similar as logistic weight function. Dear and Begg's selection functions Dear and Begg developed a semi‐parametric weight function to model the selection procedure .…”

Developments in the evidence‐based health care debate – 2004

“…The fourth paper in this section on debate and development of methodologies relevant to the EBM debate has been contributed by Preston and her co‐workers (Preston et al . 2004) and is focused on methodologies for adjusting for publication bias through a modelling of the selection process.…”

“…The selection mechanism that causes publication bias is, as they point out, complex, yet despite an extensive literature of empirical research identifying risk factors for publication bias, little work has been carried out to improve the models for selection. Using a systematic review of oral rehydration solution in the treatment of dehydration, Preston and her colleagues (Preston et al . 2004) compare methods that adjust combined meta‐analytic estimates for publication bias and consider the impact of extending selection models within a weighted distributions framework.…”

Synthesizing medical evidence: systematic reviews and metaanalyses

“…109 Many formal methods for estimating the amount of publication bias in a systematic review, and its effect on the conclusions of the review, have been published. 65,257,265,295,313,326,340,352 One simple method is to calculate the Rosenthal "fail-safe N," which is the number of unpublished studies with a null effect that would need to be added to a metaanalysis to render a treatment effect estimate nonsignificant. 277 This method assumes that unpublished trials would, on average, show a null effect for the experimental treatment.…”