655,000

#51
#51
CSpindizzy said:
So you are telling me that every poll being released is just complete garbage? Because your method is never used in polling unless it is some very local race perhaps for dog catcher.
Click to expand...
No, I am saying that most legitimate polls provide their room for error with the poll. It is usually something along the lines of, +/- 2%. That to me is fairly useful. It is not +/- 92,000, as was the 5% room for error in the 2004 survey. Heck, it is not even provided in this survey.

So, yes, this is garbage.
 
#52
#52
therealUT said:
Out of 26,100,000 people, I would like for them to poll more than, at most, 9,600 people. Maybe actually shoot for something higher than 0.03% of the population. Polling that little of the population, is why they end up with confidence intervals that are all over the place. Until they actually poll a greater percentage of the population, then their study will be worthless.
Click to expand...

I'd be less concerned about the polling numbers (the sample size is not really out of whack) and more concerned about 1 - the sampling technique and 2 - the measures used.

They used a cluster technique which is correct but choosing within the cluster and handling refusals can be problematic. There can be a self-selection bias if not handled correctly. They state in the study the great difficulties encountered in conducting the sample.

I'll have to read the report more carefully but I'd also be interested in how they verify a death and attribute cause. Most appears to be direct reporting. In some cases, they had death certificates. What does it mean to "attribute" the deaths to the war. In some cases, it is degraded healthcare or access to food. But they do claim that 600K have violent deaths with the vast majority being gunshot.

I don't know if it's accurate or not but here's where I would question the methodology.
 
#53
#53
CSpindizzy said:
To be considered a household member, the deceased had to have lived in the home at least three months prior to death. When interviewers asked to see a death certificate at households reporting a death, it was presented in 92 percent of instances. The survey recorded 1,474 births and 629 deaths among 12,801 people surveyed. The data were then applied to the 26.1 million Iraqis living in the survey area.
Click to expand...
Thank you for regurgitating the survey. I read it already. That does nothing for the fact that they surveyed less than 0.03% of the population.
 
#54
#54
therealUT said:
Of course you are CSpin...
Click to expand...

Well since you seem incapable of countering with anything factual I see no point in taking anything you say as credible. Like I said, go to Google and there are 5 sources that all have very similar numbers if not exact. You love to argue with me but off nothing to prove me wrong.
 
#55
#55
therealUT said:
Thank you for regurgitating the survey. I read it already. That does nothing for the fact that they surveyed less than 0.03% of the population.
Click to expand...

For the slow, I will reiterate the comment on death certificates presented as well. I'm sure there's a conspiracy of a basement rolling out fake death certificates of Iraqis. That would be paid for by the DNC?
 
#56
#56
CSpindizzy said:
Whoa Bama....I'm not tying my numbers to that study. So let's clear that up before you put more words in my mouth.
Click to expand...

I wasn't tying it directly to you - just using your numbers to raise the larger question of the huge discrepancy in reports.

:peace2:
 
#57
#57
Except for my very much more advanced insight into statistical measures than you apparently possess. As you still will not respond to my remarks concerning the miniscule size of the sample.
 
#58
#58
volinbham said:
I'll have to read the report more carefully but I'd also be interested in how they verify a death and attribute cause. Most appears to be direct reporting. In some cases, they had death certificates. What does it mean to "attribute" the deaths to the war. In some cases, it is degraded healthcare or access to food. But they do claim that 600K have violent deaths with the vast majority being gunshot.

I don't know if it's accurate or not but here's where I would question the methodology.
Click to expand...

I thought we were questioning the numbers. Now it seems the argument is turning to cause of death. Flip the column to non-violent causes and that still gives you 5 times more deaths than before the war.
 
#59
#59
CSpindizzy said:
For the slow, I will reiterate the comment on death certificates presented as well. I'm sure there's a conspiracy of a basement rolling out fake death certificates of Iraqis. That would be paid for by the DNC?
Click to expand...
What does that have to do with it. If 0.03% of the population offered 100% of the death certificates of the deaths of their family members, it would not change my argument.
 
#60
#60
The survey recorded 1,474 births and 629 deaths among 12,801 people surveyed. The data were then applied to the 26.1 million Iraqis living in the survey area.
Click to expand...

Interesting that the birth to death ratio for the US is actually higher. For the equivalent 1474 births, there are 781 deaths!

The basis of this study is arguing that this number is more than twice as high as prior to the war. If that's the case, would we have seen only 300 or so deaths to the same number of births? Any evidence that the birthrate suddenly fell in Iraq?
 
#61
#61
therealUT said:
Except for my very much more advanced insight into statistical measures than you apparently possess. As you still will not respond to my remarks concerning the miniscule size of the sample.
Click to expand...

I guess my company employs a fraud when it comes to polling and statistics....sorry if I deal with this on a daily basis. I guess your spare time knowledge and my work experience and college degree is not comparable.

The miniscule size of the sample? Again, miniscule sizes are pulled constantly. That's why it is a sample. Take all of the polls of the camapigns out there now and tell me there is a huge difference. The point is that this is proven scientific, it is a method used constantly by numerous firms and universities. I am not giving accuracy to the 600K but I am only stating that this methodology is not uncommon or far-fetched.
 
#62
#62
CSpindizzy said:
I thought we were questioning the numbers. Now it seems the argument is turning to cause of death. Flip the column to non-violent causes and that still gives you 5 times more deaths than before the war.
Click to expand...

Given the number of births to deaths - what I'm questioning is are they attributing general death to war-related death.
 
#63
#63
Since you seem very stuck on death certificates, let's see if this passes the common sense test, CSpin. If there are death certificates for 92% of these people (as this study claims) then why not take all the officially recognized deaths since 2003, subtract the pre March 2003 levels for normal deaths, and divide with is left by 92%.

We will use your 48 deaths per day for the basis of this little logic test here:

Three and a half years, 48 deaths per day: 61,320 extra death certificates. Divide that by 92%, 66,652. Or, 10 times less than what the Hopkins survey claims.
 
#64
#64
CSpindizzy said:
I guess my company employs a fraud when it comes to polling and statistics....sorry if I deal with this on a daily basis. I guess your spare time knowledge and my work experience and college degree is not comparable.

The miniscule size of the sample? Again, miniscule sizes are pulled constantly. That's why it is a sample. Take all of the polls of the camapigns out there now and tell me there is a huge difference. The point is that this is proven scientific, it is a method used constantly by numerous firms and universities. I am not giving accuracy to the 600K but I am only stating that this methodology is not uncommon or far-fetched.
Click to expand...

I agree that the sample size per se is not the real issue. The core issue in sampling (assuming you meet size requirements) is how the sample is chosen
 
#66
#66
CSpindizzy said:
I guess my company employs a fraud when it comes to polling and statistics....sorry if I deal with this on a daily basis. I guess your spare time knowledge and my work experience and college degree is not comparable.

The miniscule size of the sample? Again, miniscule sizes are pulled constantly. That's why it is a sample. Take all of the polls of the camapigns out there now and tell me there is a huge difference. The point is that this is proven scientific, it is a method used constantly by numerous firms and universities. I am not giving accuracy to the 600K but I am only stating that this methodology is not uncommon or far-fetched.
Click to expand...
So, when the confidence interval for this survey comes out, and there is about a 1,000,000 person discrepancy from the limits, then you just buy into the 655,000 number because why? Oh yes, because Johns Hopkins could pick any number within the limits and be just as accurate.
 
#67
#67
Again....I will say this one more time since you cannot read and comprehend what I am saying here....I AM NOT DEFENDING THE 655K NUMBER HERE. I have stated this multiple times. I have said over and over that I do not support that. I have stated that the methodology for this sample is scientific and has been used even by our current Administration to argue issues before the UN and other NGOs as well. So I guess you have a problem with your CIC choosing to believe and espouse a similar method for his own policies as well?
 
#69
#69
CSpindizzy said:
Again....I will say this one more time since you cannot read and comprehend what I am saying here....I AM NOT DEFENDING THE 655K NUMBER HERE. I have stated this multiple times. I have said over and over that I do not support that. I have stated that the methodology for this sample is scientific and has been used even by our current Administration to argue issues before the UN and other NGOs as well. So I guess you have a problem with your CIC choosing to believe and espouse a similar method for his own policies as well?
Click to expand...
Either you agree with the 655K number or you believe that the survey is a fraud.

Also, I think statistical surveys are very well used, when used properly. However, when they are not, then they are absolute garbage.
 
#70
#70
therealUT said:
Since you opened yourself up to this, I think I would take Anthony Cordesman's figures over those who just have degrees from Johns Hopkins and MIT.
Click to expand...

Funny. What makes him an expert at polling and numbers? What are his credentials? Give me some stats and proof on that. What are his numbers and how did he obtain them?
 
#72
#72
Something I found from Cordesman is that he states the UN's totals of 80 per day during the first six months of 2006 too conservative and was too Baghdad-centric. So care to now discredit Cordesman and his 80 per day death total with who knows how many additional?
 
#73
#73
CSpindizzy said:
Something I found from Cordesman is that he states the UN's totals of 80 per day during the first six months of 2006 too conservative and was too Baghdad-centric. So care to now discredit Cordesman and his 80 per day death total with who knows how many additional?
Click to expand...
I discredited Cordesman???
 
#74
#74
When the tally for civilian deaths in July is added to the Iraqi government numbers for earlier months obtained by the United Nations, the total indicates that at least 17,776 Iraqi civilians died violently in the first seven months of this year, or an average of 2,539 a month.
 
New posts
You must log in or register to reply here.

VN Store



Back
Top