Previous Next Index Thread
Previous Next Index Thread
Don Doff wrote: > I believe it is possible to evaluate how meaningful the conclusions of > a regular double blind study are by examining methodology. Is it not > also possible to evaluate meta analyses in a some similar manner, so > as to determine whether conclusions are meaningful, that would > compensate for very understandable biases in experimental methodology > by researchers desiring their studies yield positive findings for > reasons of continued tenure, funding, status in their field, etc? Yes, it is possible to evaluate meta-analysis and give a realistic appraisal of its value, but almost no-one ever does. It's very time-consuming and brain-straining to go through, say, six different studies, all with different methodologies and different analytical techniques, and then squeeze them into a meta-analysis. It's a hard enough job when you're doind your meta-analysis. No-one is keen on repeating the procedure just to evaluate the study (you're not likely to get your name on a paper because of it, and it's a LOT of work). The other problem is that many in the medical field are not au fait with statistics, and soon as they see a large sample size and a decent statistical significance figure, they assume the study is a good one. The original researchers may know their stuff, but most medicos (who are the largest recipients of meta-analysis studies) do not more than the basics. -- _____________________________________ Chris Lawson claw@ozemail.com.au _____________________________________