Suppose you are doing a meta-analysis using 20 studies, and 2 of them show a much larger relationship between the variables studied, compared to all the others.Which of the following is an appropriate way to address this issue?
A)Throw those two studies out.
B)Consider those two studies to be the most important and weight them accordingly in the meta-analysis.
C)Conduct one meta-analysis that includes the two studies and another meta-analysis that excludes the two studies and compare the results.
D)None of the above.

Question

Quizplus · Accepted Answer

Conducting two meta-analyses, one with and one without the two studies showing a larger relationship, will allow for a comparison of the results and may reveal the impact of the outlier studies on the overall findings. Simply throwing out the two studies or weighting them heavily could skew the results and undermine the validity of the meta-analysis.

Suppose You Are Doing a Meta-Analysis Using 20 Studies, and 2