volume 21, issue 12, P1804-1804 2002
DOI: 10.1002/sim.1163
View full text
|
|
Share
Petra Macaskill, Les Irwig, Stephen D. Walter

Abstract: We agree with the comments made by Dr Begg and thank him for highlighting a number of important points. Our discussion section supports his view that funnel plot asymmetry is a useful criterion for detecting publication bias, and concludes that the available 'tests do provide an objective means of identifying the presence of publication bias'.As noted by Dr Begg, our results for test size re ect increasing asymmetry in the distribution for the log-odds ratio as the mean departs from zero. We expect that such …

expand abstract