Hi all,
I want to conduct a meta-analysis review. I found that it is very common to adopt some strict criteria such as choosing four and three stars journals. However, the topic I want to discuss does not have enough studies in the top journals. Can I include some studies from two stars journal?
Hi Saleh,
I agree with Weibin that you need to consider as many studies as possible when performing meta-analysis to avoid publication bias.
However , the thing that you need to be cautious of is that since meta-analysis integrates the effects sizes from these studies, so if some of the poor quality studies are included, they may provide inaccurate results which then will affect the overall results and give you misleading conclusions. i.e. you get bias overall treatment effect with good precision.
Hi,
You might even include some grey literature (unpublised data o congress posters, by requesting it from their owners) when you have little indexexed papers on your subject
Best regards
Start with as much information as you can, i.e., all of the studies you can find, and then you can discard/weight based on study strengths/weakness, including the granularity of reported results. The process always involves some bias. But as long as you clearly identify where the decisions regarding study quality/includability were made, you should be good. People can criticize you for making different decisions than they would, but I don't think they will be able to criticize the rest of your methodology.
Sounds very good advises regarding inclusion of studies in meta-analysis. if possible, also include studies in other languages other than English. if you can the results translated to you by some one. it will increase sample size and generalizability of your research work.
As Phil Khoury has pointed out: the more information you have (studies, including grey literature) the better. You can code study quality or publication form (i.e. journal article, thesis, book chapter) and conduct respective moderator analyses in order to be able to say whether the effect size depends on study quality. Without conducting such analyses one may not assume that articles published in "top journals" have better quality than other manuscripts (Conn et al., 2003)
Article Grey Literature in Meta-Analyses
Excluding studies published in lower impact journals will bias your results. Higher impact journals tend to publish more "surprising" results and / or larger effects compared to lower impact journals so this would inflate your overall effect size. To reiterate some of the answers above, you should also try to include unpublished studies to get a sense of the true effect as unpublished studies are more likely to have non significant results than published studies and may still be high methodological quality.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis