Facts, science and linear extrapolation

The quote below is from Craig Murray's blog and wins the Quote of the Day prize simply because it calls an estimate a fact in the same sentence. I think that's pretty special for someone that was an Ambassador. After all, you don't get into the Diplomatic Service if you're a bit thick, yet for a man who is by no means that to make such a self-contradictory statement I can't deny that it's pretty damn impressive.

Today, we are calling the fact that, around now, on our best estimate, a million people have died in Iraq as a result of the chaos launched by the US and UK led invasion

When you click one of the links from Murray's site you get taken to here which explains the figure is "based on the only scientifically valid study of violent Iraqi deaths caused by the U.S.-led invasion of March 2003". They're referring to the Lancet survey that was released in March 2006 which, from a sample of thousand or so interviews extrapolated the data across the country and provided an estimate that has been (understandable of course) seized upon by every anti-war critic going as a "fact", not only that it's a "scientific" fact too.

It matters not whether you're anti-war or pro-war; nor does it matter whether one questions the political motivation and thereby interpretation of the data by those carrying out the statistical study. They are irrelevant to reality that the figure remains an estimate based on extrapolation from a small sample and bedrock assumptions about cause of death that remain entirely subjective. The figure is unquantifiable and throughly unscientific because there is no method of testing that could disprove it.

Of course the original figure was not one million, the original figure was 600,000 and was derived from an estimate that 1000 death per day were caused as a direct result of invasion from the interviews of the original limited sample set. The one million figure now exists as a result of campaigners taking the assumed starting point that the Lancet's daily extrapolation was correct, and simply adding a 1000 to it every day.

Now don't get me wrong here, I'm not saying that the figure is, by necessity, wrong. What I am saying is that it is wrong to treat it as truth. In fact it requires a startling poor intellectual sleight of hand to make the statements (a) that the figure is a fact, and (b) that it is a "scientific fact".

It is not. It is a statistical estimate based on extrapolation, and the second new figure of one million is based on an even greater degree extrapolation which completely ignores external influencing factors during the period for which it's extrapolation has been assumed correct and continued.

One can, if one chooses to, accept the Lancet's original number as a best guess but you cannot call it a fact or scientific. To be a fact it would need to have actually counted and detailed all 600,000 deaths. It didn't. To be scientific the figure would need to be falsifiable, because of its extrapolated nature it is isn't.

You cannot falsify something that is made up by mathematical playfulness that at the same times asserts an assumed subjective view of the causality of death. What's more, even if one accepts the starting figure of 600,000, it would be intellectually bankrupt to assume that the 1000 deaths per day was, is and remains valid, simply on the basis of mathematical addition per day.

Having said all this, given that I supported the war, I imagine that fact (for that is a real fact) would be enough for some to dismiss such criticism out of hand. I am, after all, simply trying to discredit something in order to handle my own guilt at having supported something that killed so many death, right?

Alternatively I might just be someone that quite literally despises the way terms like "scientific" are bandied around in order to make things sound like they have far greater validity, and also when people who I would expect to be pretty damn clever go and conflate estimation with truth.