I will allow your comments so far to stand, but since this post has scrolled off the blog front page and no one else is commenting, I’m turning off comments on this post.

The issue is thus not so much presence of death certs as the absence of a reliable central registry.

(Were certs issued at the high rate of 501/545, and were they reliably registered centrally, then simple passive surveillance of the central registry would have given the results sought without need for a survey.)

“In their Article on mortality before and after the 2003 invasion of Iraq (Nov 20, p 1857),1 Les Roberts and colleagues use several questionable sampling techniques that should have been more thoroughly examined before publication.

Although sampling of 988 households randomly selected from a list of all households in a country would be routinely acceptable for a survey, this was far from the method actually used–a point basically lost in the news releases such a report inevitably engenders. The survey actually only included 33 randomised selections, with 30 households interviewed surrounding each selected cluster point. Again, this technique would be adequate for rough estimates of variables expected to be fairly homogeneous within a geographic region, such as political opinion or even natural mortality, but it is wholly inadequate for variables (such as violent death) that can be expected to show extreme local variation within each geographic region. In such a situation, multiple random sample points are required within each geographic region, not one per 739 000 individuals.

In my opinion, such a flaw by itself is fatal, and should have precluded publication in a peer-reviewed journal. However, the authors’ sampling technique is also questionable in other ways.”

(For the record, I oppose the Iraq war but simply find the study poorly done. It is sometimes important to divorce one’s agenda from the science.)

Now for the study:

You are correct that the study (usually) selected its household start point randomly — but from then on selected contiguous households. That means the 1800 households are NOT independent but highly correlated.

Nor did you read the study well. Even the start point “random selection” had a flaw: The report admits, “Decisions on sampling sites were made by the field manager. The interview team were given the responsibility and authority to change to an alternate
location if they perceived the level of insecurity or risk to
be unacceptable.” This means even the starting points were not chosen by a fully random process but instead by human intervention which is likely to carry bias.

The study also states, “By confining the survey to a cluster of
houses close to one another it was felt the benign purpose
of the survey would spread quickly by word of mouth
among households, thus lessening risk to interviewers.”

Communication among households in a survey is even worse — it’s likely to increase the household-to-household correlation already present due to contiguity.

I assert again that cluster sampling is a valid technique — but the Hopkins study applied it poorly.

By the way, Pederson, experienced in such procedures and an author of the UN Iraq survey, has also made clear scepticism about the Hopkins study. (The Hopkins team praised him.)

Since you didn’t read the study, I can’t take your opinion seriously.

Either the ratio of certs to deaths is low (unreported), or it is high (92%). But not both.

If the Lancet study is correct, then passively counting certs is sufficient. The study didn’t explain this discrepancy.

Reasons for study over-estimation:

1. Substantial uncertainty in pre-invasion mortality baseline. Could vary from 5 to 6.8 which would remove many of the “excess deaths.”

2. Households within clusters were NOT randomly chosen. If they were self-selecting then the whole study is bogus.

Yes, I have stat background. Yes, cluster sampling is a valid technique. No, the Hopkins study did not apply it well.

Are you sure? I skimmed through the report itself and found a couple of spots in which they said death certificates are not issued in many cases.

Here, you can look it up yourself.

I think you’re makin’ stuff up. Try again.

Ignoring methodology, either the Iraqi claims are accurate because of the death certificates, and if so where are the (400,000, 600,000, 900,000 - you take your pick) certificates? or there are no death certificates and so spomething must be horribly wrong with the research.

Suffice it to say that the extra deaths in question, which have gone completely unnoticed by the world media - most of it salivating to say something anti-Bush, come out at massively more than the German civilian death toll from the allied bombing campaign in WW2.

The Top 10 GOP Excuses Regarding the Casualty Estimates:
http://www.balloon-juice.com/?p=7448