By considering bogus criticisms of the Lancet study it is possible to gain an appreciation of the concept of the infinite. No matter how many you've seen, someone can always come up with a new one. I give you Rob, who writes:
In an attempt at firmer confirmation, the interviewers asked for death certificates in 78 households and were provided them 63 times.So out of 7,000+ people questioned they only asked for proof in 78 instances? And only received proof in 63 percent of those instances? Seems like they should have been taking better care to document these deaths.
There is no point in asking for a death certificate if there hasn't been a death, so the 7,000 people number is irrelevant. And, uh, 63/78 is not the same as 63%.
Then, in this thread, Rob tries to defend his criticism. Only read if you enjoy seeing someone repeatedly punch themselves in the face.
Update: Rob has "corrected" his post by removing the word "percent":
And only received proof in 63 of those instances?
This has converted an arithmetic error to a logic error. What is important is not the raw number of confirmations, but the percentage that could be confirmed (which was 63/78 or about 80%).
I would enjoy seeing some people punch themselves in the face.
I seem to recall that one of your initial posts was called "Flypaper for innumerates". I am glad to see that the paper is still active and working!
What is even more astounding than the level of innumeracy is the level of conceit. The various "rob"s of this world who have most likely never read a statistical study paper in their lives suddenly, without any prior training or practice, consider themselves qualified to criticize a paper published in one of the leading medical journals in the world.
Of course, many of the papers in the leading medical journals are biased and do contain substantial errors, but it does take some training and practice to be able to understand these issues.
Another surprising point is the level of denial: Sure, conceivably the paper could be flawed, but these "rob" types approach this issue as if the paper must be flawed. Can't these guys imagine that it could happen that the invasion really did cause tens of thousands of extra deaths?
Hey, thanks for pointing out the typo. Obviously I didn't mean 63 <em>percent</em> of those instances, I only meant 63 instances. I typed percent by accident.
... and Rob has already corrected the error in his post, without acknowledging it or making the edit evident. One commenter has already been confused by this, asking why Tim said that "apparently" Rob does not know how to calculate a percentage.
[Note to self: always take a snapshot of the page before correcting a blogger.]
While we're at it, let's also mention that Rob has publicly admitted that he does not understand the difference between accuracy and precision, that he has no grasp of how characteristics of a sample can reflect those of a population, either did not read or did not absorb why and when the death certificates were requested, and did not bother to read the Lancet study's excruciatingly detailed description of how the sampled households were selected.
And yet he has the colossal arrogance to call a study by some of the best epidemiologists in the world "crap."
Uhh...moron? The commenter who was "confused" posted <em>yesterday</em>...before I even made the change.
And I acknowledged that the change had been made here. Right where the critics are. I didn't make an update at the post because I don't typically make updates when I change typos.
As for the other alleged admissions that I've made...would you like to provide links to them, because I certainly don't remember saying any such thing.
And before you go shooting your mouth off about changes I've made to the post, here's a cached snapshot of it from Google:
http://64.233.167.104/search?q=cache:5slYPPUvPU4J:www.sayanythingblog.c…
Thanks for deleting my comments.
Talk about people who should have taken a screenshot...
Whoops, never mind. I thought they had disappeared for a moment.
Hi Rob,
In answer to your request directed to jre on 3/3/2005 at 9:01:35 regarding the alleged admissions please refer to the following quote. This is from here:
http://sayanythingblog.com/2004/10/28/more-iraqis-dead-during-war-than-…
"Limited precision? Isn't that another way of saying, oh I don't know..inaccurate? Shouldn't the point of any study be accuracy? Without accuracy what good is a study? Why would you do a study if it couldn't be accurate?"
Quite clearly Rob you are saying that limited precision equates with inaccuracy. In a statistical sense limited precision does not equal inaccurate. You assertion is not true (dare I say it is crap?)
Hee, hee! I love it.
For someone who doesn't hesitate to call the work of Johns Hopkins and the Lancet "crap", you seem a little thin-skinned, Rob.
If you can't stand the heat, perhaps you should stay out of the kitchen.
Regarding your comments on statistics, there is nothing I or anyone else can do to prevent you from continuing to punch yourself in the face -- but feel free to stop at any time.
Making a substantive change evident in the post is considered good practice among bloggers. A good starting point for learning about blogging ethics is the BLAP.
It looks as if Rob has shut down comments, at least the link to post a comment did not work. But since he is here...
Rob, you said:
I am curious. Would you share how you came to the conclusion that the methodology was suspect. Please avoid circular reasoning and the use of other people's words. I want to know how you came to the decision that the methodology was suspect.
Funny how a group with an agenda managed to come up with the nice round figure of 100,000. Almost as funny as Tim Lambert trying to stretch his teeny Tim Blair envying dick far enough to take a piss without dribbling all over his apparel.