Saturday, October 21, 2006

The plausibility of the Lancet study

Echoing Eli a bit, a few thoughts are in order about the recent Johns Hopkins study published in Lancet that estimated approximately 655,000 Iraqi deaths as consequence of Bu$hCo's war. My guess is that Eli is largely correct in noting that most people who encounter a news report on the study (or a talk show opinion on the study) will not care one way or another as to the nature of peer review, probability sampling methods (and their preferability to nonprobability sampling methods), scientific methodology, the rigors involved in data collection, etc. Even though I find such information important and worth pointing out, for a vast proportion of our population it's likely to go in one ear and out the other.

We can discuss I suppose the reason why that is the case - an historical anti-intellectualism that makes up the American Zeitgeist, a fundamental hostility toward the sciences that has developed in the US over the course of my lifetime (which is starting to haunt us - but that's a story for another time), the neglect of science education at all levels of education from k-12 up through higher ed, or some combination of the above. Heck it wasn't all that long ago that expert witnesses testified in the OJ Simpson case on the DNA evidence (the consensus among my grad school colleagues at the time was that the testimony and evidence regarding the DNA samples collected was very well-done), only for that testimony to be roundly ignored by the jurors in that particular trial.

Eli points out to a creative, street level means of gauging the plausibility of the Johns Hopkins study that I'd like to pass along to y'all:
All of which is a long introduction to one more piece of evidence in that effort. With a hat tip to Cursor, this effort by a blogger to calculate the number of bullets being used by American forces in Iraq. The answer? 275,000 bullets per day. Now, as he points out, some of those are used in training, some may be stolen, and most miss their targets. But if a mere 1% of that number actually hits an Iraqi, and let's say 10% of those shots are fatal, that would be 275 people a day (100,000/year) being killed, just by bullets (that is, not including tank shells, missiles, bombs, etc.). Now of course these are no more than "back of the envelope" calculations. But they certainly provide one more piece of evidence that the Johns Hopkins result is very much in the right ballpark.
Hardly scientific, but at least gives us one more reason to consider the results from the Johns Hopkins study as plausible.

No comments:

Post a Comment