Iraq Death Toll
- Thread starter toby
- Start date
Estimates are always used in wars.
Estimates in the civil war in the Congo range from 3 to 8 million people.
No NeoCons ever get their panties in a wad, about the wide probabalistic estimates from the Congo Civil war.
I think these Lancet researches have done more scientifically credible work than anyone out there. They are quite honest about the statisitcal ranges and limits of their methodology. But, this is a serious study, which is internally consistent with their previous study, the trend lines from passive reporting on civilian deaths, and with other conflicts in recent history.
I'm quite sure that they are in the right ball park with their range of estimates. And, as you pointed out, it is a pretty big range.
But, that's the nature of statistical estimates in wars.
Estimates in the civil war in the Congo range from 3 to 8 million people.
No NeoCons ever get their panties in a wad, about the wide probabalistic estimates from the Congo Civil war.
I think these Lancet researches have done more scientifically credible work than anyone out there. They are quite honest about the statisitcal ranges and limits of their methodology. But, this is a serious study, which is internally consistent with their previous study, the trend lines from passive reporting on civilian deaths, and with other conflicts in recent history.
I'm quite sure that they are in the right ball park with their range of estimates. And, as you pointed out, it is a pretty big range.
But, that's the nature of statistical estimates in wars.
The point to all of this is that this methodology even if there is significant duplication still gives a number that is over three times what John McLaughlin has been reporting and no righties on here were willing to even accept his numbers so it doesn't surprise me a bit that these numbers which could range according to the report from 400,000 to over 950,000 is certainly going to make the righties take a real exception to the report. But it sure makes that McLaughlin number look a lot better doesn't it???
Nah, I'm willing to accept such numbers. I think many of the Iraqis have been killed as a result of our invasion, including regular army as well as civilians, the civilians mostly due to Terrorism attacks. As this report states. Methinks you are too willing to say "no right winger". I was very specific, trusting the reported number when there is such a large range as perfectly accurate is problematic.The point to all of this is that this methodology even if there is significant duplication still gives a number that is over three times what John McLaughlin has been reporting and no righties on here were willing to even accept his numbers so it doesn't surprise me a bit that these numbers which could range according to the report from 400,000 to over 950,000 is certainly going to make the righties take a real exception to the report. But it sure makes that McLaughlin number look a lot better doesn't it???
I wasn't aware that you accepted John McLaughlin's numbers when I posted them a couple of weeks ago. In fact, I must have missed that post in the slew of posts that went on and on about how 100,000 or more dead was impossible. So if I missed your post accepting John McLaughlin's numbers without reservation I am completely and extermely sorry.
I don't think I posted in the thread. I warn't playing devil's advocate that day....
Anyway, I'm willing to accept these kinds of numbers as realistic estimates. I'm just anal about people using them as accurate reality. But then I'm an engineer and such terminology just gets to me...
It isn't like I've been all for this "war" to begin with. I've been against it from the beginning based on the fact that it is an undeclared war, and since its inception the War Powers Act has only gotten us into debacles. If it isn't enough to Declare War I don't think we should be messing with it.
T
TRGLDTE
Guest
My premise in making that point was that there are some who would believe that the total death figure cited is the result of U.S. military action as the sole cause. Clarification of that point is not a refutation of the study. Some are so lacking in a basic understanding of that simple idea that they will write copious quantities of gibberish, ramblings and rantings, illustrating nothing but their own ignorance.
Last edited:
T
TRGLDTE
Guest
I provided the correction to the info. Hopkins is credible as I point out and, in terms of medical / health issue research, Hopkins is likely better than MIT.
OrnotBitwise
Watermelon
Fair enough.
Yes, this study attempts to measure deaths from all causes, not merely direct military action. As such, it is certainly going to generate higher figures than any (credible) count of deaths from direct military action.
T
TRGLDTE
Guest
When a death certificate is available the likelyhood of duplication is small. How many certifications for Ali Akbar, died 1/1/2006 do you see before you realize they are the same guy? The centralized report gathering authority would be the Center at Hopkins. I see no real threat to the validity of this study from duplication.
A bigger threat would be the process by which death certificates are issued. If we are counting death certificates to determine a normailzed death rate as a proxy for quality of life, then a change in the method / system by which the certificates are issued changes the basis. (A gross example would be changing from a metric scale to an English scale.) I have no proof that such a change happened, and this is just a speculative example. If such a change did occur, then I am sure that Hopkins would note this in the report.
No, only a copy is sought after hearing of the death. Where do they seek the copy from?
Honestly, there is a reason for such a large +/- margin of error, and it isn't because duplication is non-existent. Dang. I know you really want to slog this off as superaccurate, but it honestly isn't and for reasons I have proscribed. Even the report itself directly states this particular problem.
T
TRGLDTE
Guest
Yes, this is a change in counting method as I discussed above. This data can't be used to extrapolate because the conditions changed as well. Right now, there isn't an aggressive bombing campaign, so if one tried to use this data to back into numbers when that was happening, it has limited application.
Not compared to "shock and awe" the air based attacks are significantly less than at the beginning of the campaign.
T
TRGLDTE
Guest
That is incorrect. A number is meaningless on it's own. What is it we are counting?
T
TRGLDTE
Guest
Ok, but that's not the point of the comment.
T
TRGLDTE
Guest
That's fine, and any good report should note sources of error.
Margin of error is related to sample size versus population. By math, the margin of error decreases as the sample size versus population increases. I can have a margin of error of +/- WOW! and still get a reasonable conclusion. All I need is two samples sized such that the extreme possible values of the two samples do not intersect for the confidence level I want to achieve.
The numbers I have seen cited were 5.5 /1000 (sample one) and 13.3 / 1000 (sample two). As long as the highest value of 'set one' is less than the lowest value of 'set two' or, (5.5 + margin of error#1) < (13.3 - margin of error#2), the populations do not intersect, and I can infer that the two populations that were sampled are different (wrt conf level%). In other words, the study does not have to be superaccurate to be valid.
In nearly every study, collecting data is far more expensive than crunching the data. Why, given the expense and danger associated to collecting this data, would the researchers collect more data than they needed?
Last edited: