Is there a way of investigating publications bias in a meta-analysis of single case studies? Usually one can assess publication bias using funnel plots or Egger's test to assess funnel plot asymmetry. However those methods require some estimate of the individual studies' variance. In case of single case studies such measure of variance is usually not available. So is there any way to assess publication bias in this case?

Here i provide a dummy example in R. In a meta-analysis of single case studies, all studies investigating a change in symptom score following a treatment have been assessed:

# change_score <- c(-3,-2,-1,-7,-8,-11,-1,-10,-4,-11)
  • 3
    $\begingroup$ Welcome to the site, @jokel. This Q was flagged as being a little sparse. Could you expand on your question a little bit? EG, it's common in MA to make a funnel plot to check for publication bias, are you asking if / how that could be done w/ single case studies? $\endgroup$ – gung Feb 22 '13 at 22:01
  • $\begingroup$ jokel, you have attempted to edit your post from a different 'jokel' account than the one you originally posted from. Please read the help about merging accounts $\endgroup$ – Glen_b Aug 8 '13 at 7:57
  • $\begingroup$ Possibly we might come up collectively with something inspired by instrumental variable analysis. For instance by looking at discrepancies between trends in some uninformative features and outcome features. Of course sample size is not at stake here, nor exploring per se for small study effects... $\endgroup$ – Joe_74 Mar 20 '16 at 23:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.