If you thought that this was one of the error-free techniques for analysis, then you are wrong.
There are several limitations that are observed in this technique, due to which researchers must be careful while handling data and its analysis.
In this method of research, the outcomes of the experiments largely depend upon the researches that have been included. If the studies have not been selected properly, then you will never be able to come up with a significant result. The three common problems that are faced in the identification and selection process of the studies are as follows:
Most of the renowned research journals only publish positive studies; those which are composed to favor a particular treatment or against it. This creates an overall biasness in research sample which should contribute all researches. This is called publication bias. The biggest limitation in such kind of analysis is that publication bias will tamper the entire results of the research and make it insignificant.
To overcome this flaw, serious effort must be made to identify studies that have not been published.
Just like the justification of publications matter in meta-analysis, the search for the relevant studies should also be conducted in a justified manner, to make sure that the overall result of the study is significant. Relevant key words should be used in the search engine so that the number of searches is not biased.
After the identification stage is the stage where relevant studies are included in the research. You may have identified many studies that are not directly related to your research topic. If irrelevant studies are included in the research, then the outcome shall, again, be unimportant. For that, it is necessary to make a list of the important components of your research like the research objective, sample size, design method, outcomes measured etc. and make sure that only the studies that fulfill this criteria.
Some other limitations include the misapplication of proper methods to the data, or the mishandling of the data and its analysis. In this method another drawback observed is the “apples and oranges” effect where, the method tries to mix together data from two completely different fields. This technique is also often criticized for its lack of insight for qualitative data.