This sounds to me like the traditional fun of building a Guy and then
setting fire to him.
On 8 DFómh 2005, at 12:38, Puliyel wrote:
>
> Yet on systematic review we have 2 RTC coming to diametrically
> opposite conclusions.
What systematic review? Or did you invent it for the sake of argument?
> Why is that – if bias and confounding and chance have been
> excluded? One or both the studies may be wrong. Now if we assume
> only one of the two studies is wrong - taking the mean value
> between the correct value and the wrong value will not yield a
> value that is 'more correct'. This is the error of meta-analysis.
Can you give an actual example where
a) the literature contained a number of trials which came to opposite
conclusions and
b) the authors of the review did not perform any test for
heterogeneity but instead simply calculate a pooled estimate?
If such an example exists, then it is in error. But the error is not
inherent in meta-analysis, but in its misuse. No statistical
technique is guaranteed against misuse. But the argument you present
ignores how meta-analysis is actually done.
Ronán Conroy
[log in to unmask]
|