953
edits
Changes
Jump to navigation
Jump to search
Line 16:
Line 16: + +
no edit summary
:::It is speculation in this particular case that the difference between the 2 sets of findings is explained by one being more accurate. In general, prospective data are given more weight. That's my point - there were different findings from one design type relative to the other. It is not clear why. As far as the size of the studies, by my math the 13 prospective studies (ie, the ones that show no increased risk) have an average sample size of 12,493. The 39 retrospective studies have an average sample of 2,228, which is about 17% of the size of the prospective studies. There is also the issue that nearly all of the 33 retrospective studies that are quoted here as showing a risk show that the risk is not statistically significant. [[User:Murray|Murray]] 20:58, 22 March 2008 (EDT)
:::It is speculation in this particular case that the difference between the 2 sets of findings is explained by one being more accurate. In general, prospective data are given more weight. That's my point - there were different findings from one design type relative to the other. It is not clear why. As far as the size of the studies, by my math the 13 prospective studies (ie, the ones that show no increased risk) have an average sample size of 12,493. The 39 retrospective studies have an average sample of 2,228, which is about 17% of the size of the prospective studies. There is also the issue that nearly all of the 33 retrospective studies that are quoted here as showing a risk show that the risk is not statistically significant. [[User:Murray|Murray]] 20:58, 22 March 2008 (EDT)
:::Actually I misread the table initially. The 13 prospective studies have, on average, a sample of 3,374, compared with 2,176 for the retrospective. So the latter are actually, on average, about 2/3 the size of the former. [[User:Murray|Murray]] 21:03, 22 March 2008 (EDT)
== #3 ==
== #3 ==