[Caveat: I have no opinion on circumcision status. Not my area of clinical expertise. However I can scan a statistical paper and am here to argue about its methodology]
The paper is freely available online for anyone to read. I have yet to see any convincing argument in this thread as to why the methodology of this paper is flawed. This comes the closest:
if you actually read the study, they have a laughably small number of children who are circumcised
in otherwise, bullshit study
you get skewed results due to the sheer miniscule portion of these "300,000" kids who are actually circumcised.
also most of their results are not significant
To which I reply to GiJoccin, if you read line 5 of page 8 in the paper you'll discover that over 3300 boys who underwent circumcision were included for this study. Although this represents a small proportion, no researcher in the world is going to call a 3300 sample size a "laughably small number." The main argument of the paper comes from Table 2, 0-4 age group. The hazard ratio in the non-Muslim comparison is 4.23 with a 95% CI of 1.90-9.44. That is statistically significant. The Muslim comparison was also significant but less so.
The
cases are rare, I'm not arguing that. But if you know basic biostatistics that's also the main reason why case-control studies are done in the first place.
You should also realize that the researchers also performed regression analysis on all the pertinent confounding factors as listed in Table 1 and still found statistical significance. They did their homework.
My personal criticisms? No fucking p-values within the tables. I have to read every confidence interval and see if it crosses the value 1. That's bullshit. Moreover, the authors do not state whether their p-values stem from one-tail or two-tailed analysis. Finally, one author has disclosed a conflict of interest. Edit: Probably what I believe is the weakest part of the study: I'm willing to bet that the 0-4 age group versus 5-9 age group was a post-hoc analysis because they never mention in their background/introduction section nor in their discussion section about what's so magical about the 4th year of life that decreases the association of autism with circumcision. Unless I missed it, they do not make this designation a priori. The authors also include 2-4 year-olds as "infants," which is not what most pediatricians would call them.
Just because a study is peer review doesn't make you couldn't play around with figures to get the result you were looking for.
Just like all those 'peer reviewed' studies that prove vaccines cause Autism.
You mean the
Andrew Wakefield paper? The paper with a sample size of 12?
There is a heck of a lot wrong with a study unless there is much more than the newspaper is letting on.
Elaborate. What was wrong with their methodology?
I'm not arguing that circumcision causes autism. But the fact that researchers found a statistical significance after controlling for confounding variables should make others ask why and everyone here is just spouting their personal biases without even looking at the methodology. I'm not sure why these types of threads are even allowed if no one can even freaking read the paper.
