How important should the significance level be in the decision process to include single results in a Meta-Analysis?

In a Meta-Analysis of different studies some of the included studies describe their results as insignificant. The outcomes of the single studies are reported for control and intervention groups separately and should be included in the Meta-Analysis.

  1. Does the principle of a meta-analysis permit the inclusion of single insignificant results?
  2. Has the insignificance of single studies to be accounted for in the calculation of the overall effect of the meta-analysis?

If you mean that they report the effect size but state it was not significant then you can include it.

If you mean that they do not report the effect size but just state that it was not significant then you might consider using the metansue package available from CRAN which uses a form of imputation. Disclaimer: I have not tried it myself.

  • $\begingroup$ Yes, I meant that they report the effect size but write that it is not significant. $\endgroup$ – Sebastian Jan 23 '16 at 8:05
  • $\begingroup$ Well in that case you are fine to go ahead unless you have some other doubt. $\endgroup$ – mdewey Jan 23 '16 at 16:52
  • $\begingroup$ do you have other statistics in the studies ? $\endgroup$ – Subhash C. Davar May 3 '16 at 12:00

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.