My main concern, my underlying point, is that the methodology they used would only have been acceptable if they could have established that their pairings were in fact similar. I now have two sets of data that indicate this not to be true for at least half of the pairs, if not five out of six.

Seeing as how they are different, this automatically spells out problems because any variance between the pairs would not make it into the study and the confidence interval – obviously since half of the pairs have not been sampled, and the sampled member of the pair being projected as the same rate for the other member.

The entire clustering of clusters procedure is one that I cannot find done anywhere else, and which seems to be at odds with the definitions of a random sample, and also at odds with the entire concept of sampling. That is unless you assume that the paired up governates had similar death rates…. which I think two sets of data now put to shame.

But, rereading a previous thread, I suggest you polish your definitions of terms like stratified sampling and cluster sampling before submitting (Hint: they are not related in any significant way, except that the population is divided into groups before the sample is drawn.) Stats journal referees tend to be biased against people who get this kind of thing wrong.

quiggin,

I already have a dialog going with Les Roberts himself, which I will finish after the holidays. Obviously the Lancet people would want to defend a study they have put their credibility behind in publishing, just like Lambert incessantly defends it with every ounce of effort in his body. Roberts told me himself that it had been reviewed by so and so. That’s great – I’ll never know what they said about it anyways. Their assumption that the governates they paired up were similar might have been taken unchallenged by the peer-reviewers.

If I don’t get anywhere with Roberts, then I might be persuaded to submit my critique to the Lancet, although I’d expect the same results as I have gotten on Lambert’s blog. When people have invested too much of their own credibility into something, they will never admit to being wrong in any way.

Which is also why Lambert censors my comments instead of letting the truth withstand the scrutiny it deserves.

Seixon wrote: “I just think it is quite telling that Lambert censors every comment of mine that uncovers his dishonesty about the Lancet study, or any other comment about the study at all.”

That statement is 1. untrue and 2. Seixon knows that is is untrue. Seixon lied.

But that wasn’t your initial argument was it? Your initial accusation was that he claimed he was banned when he wasn’t, hence he lied. Look at your entire post 70.

Now you change your tune, claiming – contrary to your previous assertion that you DID require him to shunt his comments into a moribund thread – that you do not, in fact, engage in that mode of censorship. Not only did you change your argument, in changing your argument, you contradict yourself. Which is it?

That statement is 1. untrue and 2. Seixon knows that is is untrue. Seixon lied.

As for marginalizing him as a buffoon. Look at the post. I just presented his own words.

Publication would make it clear that you have, at least, a credible point to make rather than, as it appears to most of us, a basic lack of understanding of statistics. Or do you think the editors and referees of the Lancet are in on the conspiracy?

As my “sockpuppet” has pointed out, I never said I was banned. In fact, I said the direct opposite. I did write an e-mail to Lambert asking WTF was going on since none of my comments were being allowed on his blog, and I’m guessing he has some type of keyword block on me so that all comments that mention the Lancet are immediately sent to moderation or deleted. I don’t know, but I certainly haven’t said here that I am banned.

Lambert’s habit of ignoring simple details persists, you know, like actually reading what I said, and not what others were saying about me.

Tsk, tsk.

As for you brendan, why you are setting up strawmen about me not agreeing with heliocentrism and a bunch of other topics I have never even discussed on Lambert’s blog (or anywhere else – I’m not a science geek after all), I do not understand. Oh, other than to continue to try and marginalize me as a complete buffoon. Yes, I believe that the world is flat. Ya happy? LOL.

I know you aren’t interested in debating people who can show clear faults in the Lancet study, and neither is Lambert. He has invested around 50 posts in defending the damn thing, it’s not like he is going to turn around on that record.

Anyone who will describe the Lancet study methodology as “robust”, well, I’ve got some beach front property to sell them on the North Pole.

LOL. Tim, how are you not agreeing wth Seixon’s depiction of your censorious antics?

Seixon: “[Lambert] started censoring all my comments on the Lancet study and forced me to post in a dead thread about it so no one would read it. . . . I haven’t been banned, although I am strictly forbidden from saying anything about the Laneet study [in new threads about the study] because Lambert can’t bear it.”

Lambert: “He has not been banned from commenting on the Lancet study [except in new threads and “every other discussion on the Lancet”] and he knows it. He has posted hundreds of innumerate comments criticizing the Lancet study in my blog. All I have done is require him to post his comments in his own threads [i.e. dead threads] and not clutter up every other discussion on the Lancet with his nonsense. . . . He is well aware of this because I explained it to him again in email and I allowed his comment once he posted it to the correct thread.

Seixon sez: you have not banned me outright.
Tim sez, ridiculously: he has not been banned so he is lying.

Seixon sez: he forces me to post in moribund threads only, and I can’t post in any other thread even when it discusses the Lancet study.
Tim sez: Well he is required to post in these threads and not others, so he lies.

In other words, Seixon’s comments aren’t allowed unless they are posted to so-called ‘dead’ threads.

So, isn’t Lambert just saying what Seixon is saying? So how is Seixon “lying” then?

Tim Lambert cred = nil.

Shame.

It’s sad that you have to maintain a consistent censorious regime in solidarity with Tim ‘no cred’ Lambert in order to avoid being taken to task by your critics. As it is, I did not express an opinion as to the state of the Lancet debate. I did however, come to the defence of someone who has been unfairly maligned here, judging from only the posts made here. And that conclusion is sound: Lambert is clearly dissembling by ostensibly agreeing with what Seixon wrote, and then ridiculously accusing him of “lying”.

Unless of course you knew all this already. In which case: shame.

LL. Tm, hw r y nt grng wth Sxn’s dpctn f yr cnsrs ntcs? Sxn: “[Lmbrt] strtd cnsrng ll m cmmnts n th Lnct std nd frcd m t pst n dd thrd bt t s n n wld rd t. . . . hvnt bn bnnd, lthgh m strctl frbddn frm syng nythng bt th Lnt std [n nw thrds bt th std] bcs Lmbrt cnt br t. Lmbrt: H hs nt bn bnnd frm cmmntng n th Lnct std nd h knws t. H hs pstd hndrds f nnmrt cmmnts crtczng th Lnct std n m blg. ll hv dn s rqr hm t pst hs cmmnts n hs wn thrds [.. dd thrds] nd nt clttr p vr thr dscssn n th Lnct wth hs nnsns. . . . H s wll wr f ths bcs xplnd t t hm gn n ml nd llwd hs cmmnt nc h pstd t t th crrct thrd. rn’t y jst syng wht Sxn s syng? Tm Lmbrt crd = nl. Shm.

The rest of my statements stand.

He is well aware of this because I explained it to him again in email and I allowed his comment once he posted it to the correct thread.

How do you reconcile that with the recent all-provinces survey that had a majority saying their lives were better post-invasion?

http://news.bbc.co.uk/1/shared/bsp/hi/pdfs/12_12_05_iraq_data.pdf

It would be interesting to take that survey, discard the data from the provinces that the Lancet study decided to skip, and see just how different the result was.

soru

Electricity production in population centers fell dramatically, quality/availability of drinking water fell too, plus the crime and bombings—and 100K out of 25M is less than 0.5%, one excess death for every 200 people.

Imagine your town with little or no electricity (and 110 all summer), no drinking water or sewage system, plus rampant crime and occasional explosion here and there – is this going to kill just 1 out 200 over 18 months? No way, man. It’ll be at least a dozen, probably more. So, I have to agree with Seixon here: the Lancet study is crap, not even close. We should be talking millions here.