Many thanks to all who answered for such detailed explanations 🙂
In: 0
Meta analysis
1. Come up with a question? – e.g. does tobacco smoking cause lung cancer
2. Do massive literature search via various databases for ever possible research paper about the subject – tobacco + lung cancer = really long list of papers
3. Manually go through all the papers to see which ones have actually looked at your question specifically, are original research and the research is high-quality. Discard those that do not (i.e those that are reviewing research previously done by others, those asking a different question that got picked up by your database search, those that do not give enough sufficient data for analysis)
4. Combine all the data from every remaining paper and analyse this in one go to see if combined it gives a more definite answer to your question and how confident that answer is. So you could have individual papers where the most number of people studies were 2000, but combine lots of good quality research papers and they’ve looked at 50,000 individuals. You can be more confident that the combined result is truest reflective of real life. In the above example you need to work out if the tobacco smoking is the cause of the lung cancer and not something else that the people are doing.
Cochrane Library – mega-analyses and systematic reviews related to medicine and health-care. The analyses are all done people trained and involved in the field (eg oncology research is assessed by oncology specialists), done in a specific systematic method and is peer-reviewed. The people doing the analysis and writing the reviews are not paid for their work, so conflict of interest is reduced
Latest Answers