Talk:Lancet surveys of Iraq War casualties/Archive 3

From Wikipedia, the free encyclopedia

< Talk:Lancet surveys of Iraq War casualties

This page is an archive. Do not edit the contents of this page. Direct any additional comments to the current talk page.

Talk archives for Lancet surveys of Iraq War casualties (current talk page)
<< 1	< Archive 2	Archive 3	Archive 4 >	3 >>

January - 2007

1 ILCS vs. Lancet
2 The supposed ILCS "baseline rate"
3 Graphs
4 lancet excess mortality rate june 2006
5 images and captions
6 Removed erroneous Lancet article graph
7 Lancet graph problems
8 why is an incorrect graph in the article?
- 8.1 Similarities and Differences in Lancet and IBC trends
- 8.2 Lancet graph in other articles.
9 User-created graph illustrating wide variation in estimates
10 Erroneous Lancet graph
11 Should the erroneous Lancet graph be in the article?
12 Seigfried4220 and 4 previous anonymous IP addresses

ILCS vs. Lancet

I deleted a bunch of info that User Timeshifter added that was taken directly from the Lancet study. This section was already way too long as it is, and the added info presented only one POV (that of the Lancet author) without any response. Benwing 07:31, 5 January 2007 (UTC)

This was discussed on a couple talk pages. Both sides were presented in the article. If you have more info that needs to be presented, then please do so. --Timeshifter 13:28, 5 January 2007 (UTC)

Benwig, all of Timeshifter's *voluminous* edits on every page dealing with this topic seem geared toward shifting them all toward the POV of the Lancet author74.64.60.148 15:44, 5 January 2007 (UTC)

It looks like you are back. An admin almost protected the IBC page due to your refusal to discuss anything at one point. Welcome back. At least you are talking again. --Timeshifter 16:21, 5 January 2007 (UTC)

You seem to also be fond of making things up about other wiki users when you aren't cherry picking one-sided references to pad all these pages toward your POV. —The preceding unsigned comment was added by 74.64.60.148 (talk) 04:41, 6 January 2007 (UTC).

The claim that "ILCS used a baseline rate of 9" is not in ILCS. This seems to be your interpretation of one of Pedersen's comments, but his comment doesn't say that. Timeshifter concocted a bunch of statistics based on his own dubious interpretation of Pedersen, and then gives an argument for Lancet based on this assumption. This is both unsubstantiated and is editorializing argumentation, inappropriate here. —The preceding unsigned comment was added by 74.64.60.148 (talk • contribs) 04:58, 6 January 2007 (UTC)

Please refrain from undoing other people's edits repeatedly, as you are doing in Lancet surveys of mortality before and after the 2003 invasion of Iraq. If you continue, you may be blocked from editing Wikipedia. Note that the three-revert rule prohibits making more than three reversions in a content dispute within a 24 hour period. Additionally, users who perform a large number of reversions in content disputes may be blocked for edit warring, even if they do not technically violate the three-revert rule. Rather than reverting, discuss disputed changes on the talk page. The revision you want is not going to be implemented by edit warring. Thank you.

You are now up to around 3 different anonymous IPs in our discussions. I will probably list them here eventually. They all have links for user contribution lists. You have deleted the same sourced info on the ILCS study 2 times in less than 24 hours. --Timeshifter 06:44, 6 January 2007 (UTC)

Timeshifter, I'm not sure what the history involving 74.64.etc is, but as for your edits, you need to understand that the intent of Wikipedia is to *summarize*, not to construct a thorough report documenting everything out there. just because material is sourced doesn't mean it belongs. this whole ilcs section is a very good indication of what wikipedia articles should *NOT* look like, as it contains lots of large direct quotes and little info tying them together. the two paragraphs you added only made things worse, as well as presenting only one side of the issue; that's why i took them out.

you obviously won't tolerate your prose being removed, and i don't have the energy to fight you over this, but i hope you seriously try to edit this section and remove the large block quotes. in the meantime i've tagged the section with POV and cleanup tags. Benwing 07:45, 6 January 2007 (UTC)

No offense, but many sections of many wikipedia pages are longer. Including other sections on this page. Also, see other political pages. I like some of your edits though, so feel free to have a go. I will try also. The material is complicated, so please don't delete material you don't understand. Read up further, and inform people in this talk section as to which sections are not clear. Some of the material is duplicated in another section in this article.

User:74.64.60.148 in a previous IP incarnation insisted on it. He went on revert wars back then, too, to get his way. The following notice is for him. 74.64.60.148, you are now up to 3 deletions of the same sourced material in less than 24 hours. --Timeshifter 08:07, 6 January 2007 (UTC)

OK, if i have some time over the next couple of days i'll try to clean up the section. Benwing 08:24, 6 January 2007 (UTC)

Your "information" on the ILCS study is not sourced nor substantiated. It's POV editorializing and assumptions. As for myself, I made some edits in the last day or two and haven't edited anything in weeks or months. My internet connection changes IP from time to time on its own. I don't see what this has to do with anything. Contrarily, you've edited this (and most other pages on this topic) more than anybody else, dozens and dozens and dozens of edits padding all these pages with cherry-picked factoids compiled and presented all to support one POV. But just making stuff up, like that "ILCS used a baseline rate of...", and then writing a whole argument of your own based on that, is beyond the pale.74.64.60.148 08:09, 6 January 2007 (UTC)

why don't you just create a user name? you'd almost certainly get more respect that way. it's understandable why wikipedians don't like edits coming from people without accounts. Benwing 08:24, 6 January 2007 (UTC)

Benwing, I don't see how or why making up and posting anonymously under a phony name like "Timeshifter" is any more or less respectable than just posting. I doubt it would make any difference anyway, as I think this is just a diversion. The diversion will just shift to something else.74.64.60.148 09:44, 6 January 2007 (UTC)

74.64.60.148. Due to your 4th reversion in less than 24 hours a 3RR violation report has been made at WP:AN/3RR. This is the second page for which a report has been made. In the space of an hour. The other page is Iraq Body Count project. You were warned of an impending WP:3RR violation on both talk pages, in edit summaries, and on your user talk page. --Timeshifter 10:14, 6 January 2007 (UTC)

The idea of creating a user account is that all of your contributions become clearly identifiable as coming from you, rather than from various IP addresses that can't easily be cross-referenced and may sometimes correspond to others. hence you can build up a reputation. even if you don't agree with this system it's the way things work here, and if you want to argue your case against Timeshifter you'll get a lot more support if you have an actual account. Benwing 01:02, 7 January 2007 (UTC)

The supposed ILCS "baseline rate"

Editor Timeshifter has concocted this whole section and theory about a supposed "baseline rate" of 9 which he claims is used in ILCS. He appears to base this claim on a paraphrased version of something the ILCS author Jon Pedersen said to a journalist, in which Pedersen is paraphrased saying "the UN used" a baseline rate of 9. "The UN" is not the "ILCS" study. "The UN" is a huge body that has many publications, and has estimated and used that rate before in contexts other than ILCS. I believe this is what Pedersen is referring to. Nowhere in ILCS does it ever mention using this "baseline rate", or any other, for its death estimates.

Furthermore, in another quote put in by Timeshifter in the same section, Lancet author Les Roberts is quoted as asserting that "the overall non-violent mortality estimate (in ILCS) was, I am told, very low compared to our 5.0 and 5.5/ 1000 /year estimates for the pre-war period which many critics (above) claim seems too low."

So Timeshifter is asserting ILCS used a "baseline rate" of 9. And he's quoting Roberts asserting that ILCS estimated the natural death rate at less than 5, based on what he'd been "told" by someone or other.

There is nothing in ILCS saying either thing. Timeshifter is making up this fact of an "ILCS baseline rate" of 9 based on his own dubious interpretation of one paraphrased comment by Jon Pedersen, and then quotes Les Roberts giving a contradictory account based on what somebody had supposedly "told" him. This careless invention of "facts" is more than enough reason for their removal.74.64.60.148 08:42, 7 January 2007 (UTC)

You are not reading the material correctly. --Timeshifter 08:52, 7 January 2007 (UTC)

Explain then. Where is the "baseline 2002 mortality rate" of 9 in ILCS? You're making this up based on this comment about the UN in that article. Then you've constructed a whole theory based on this. Then Roberts contradicts your theory (but then his comment is just hearsay and not reliable either). This is just more POV pushing. Usually you just pad the pages with cherry-picked references to do your POV pushing, now you're just making stuff up.74.64.60.148 09:11, 7 January 2007 (UTC)

It was explained until you deleted sourced material. Most of what you are saying is untrue. What you don't understand is WP:NPOV. Wikipedia puts out all significant viewpoints with sources, and does not take a position. The positions are staked out by the sources. The reader decides how much of what to believe from all sides. You have been deleting sourced material to favor a POV. Not allowed in wikipedia. You have been doing this for a few months now. Under several different IPs. Why can't you get a user name. It is very suspicious to do what you do for so long under several IP addresses. --Timeshifter 09:23, 7 January 2007 (UTC)

More diversions from you. Where does this supposed "baseline rate" of 9, supposedly "used in ILCS", come from? That whole part is predicated on this theory you've made up about this "baseline rate". And the "positions", in your case, are staked out with cherry-picking and biased editorializing for one POV.74.64.60.148 09:31, 7 January 2007 (UTC)

I made up nothing. Everything was sourced. You deleted it. End of story. We spent weeks arguing on another page. While you deleted lots of sourced material. You put out a fog of misinformation then, and you continue now. Put back the sourced material. Admins appreciate that in their decision making. I have deleted none of your sourced material on these topics. That is how it works. Put out the info. Let the readers decide. --Timeshifter 09:37, 7 January 2007 (UTC)

Yes, you have deleted my sourced material, on another page on this topic, which I then put back. What I am deleting of yours is a whole argument about a supposed "ILCS baseline rate", some of the material within your outline of this theory of yours is sourced material, but its relevance to a "comparison of Lancet to ILCS" is non-existant. There is no "baseline rate" in ILCS. Thus this whole section is entirely incorrect, irrelevant and misleading. Pedersen's questioning of a baseline rate has nothing to do with a comparison of ILCS to Lancet because the paraphrase you decided to use as the launching pad for your editorial theorizing is not talking about ILCS.74.64.60.148 10:45, 7 January 2007 (UTC)

You, 74.64.60.148, keep deleting this material quoted directly from the supplement to the Lancet study:

"For the purpose of analysis, the 40 months of survey data were divided into three equal periods—March 2003 to April 2004; May 2004 to May 2005, and June 2005 to June 2006. Following the invasion the death rate rose each year."

"Pre-invasion: 5.5 deaths/1,000/year
March 2003-April 2004: 7.5 deaths/1,000/year
May 2004-May 2005: 10.9 deaths/1,000/year
June 2005-June 2006: 19.8 deaths/1,000/year
Overall post-invasion: 13.2 deaths/1,000/year"

It is directly relevant to the discussion. --Timeshifter 09:50, 7 January 2007 (UTC)

The discussion of your made-up "ILCS 2002 baseline mortality rate"? Again, you evade the question, posed directly now several times. Where does this supposed "baseline rate" of 9, supposedly "used in ILCS", come from?74.64.60.148 10:45, 7 January 2007 (UTC)

It comes from ILCS. Now that the article is intact again, it is clear. It was always clear. Your deletion attempts were an effort to make it unclear.

User 74.64.60.148, you have been deleting sourced material on a couple pages now for months under several different IP addresses. I finally got fed up and decided to figure out 3RR incident reporting, and he was blocked for 24 hours today. The bottom line in this specific case is that you are deleting the same relevant sourced material, and putting out a fog of misinformation to cover it up. The diffs at the 3RR incident board show the deletions clearly. I have been thanked several times by several editors of various Iraq War casualty pages. Only an anonymous editor (you) seem to have a serious problem with my edits. At times you have been relentless. And why can't you get a user name? I have deleted none of your sourced material. At one point I thought we had come to an agreement. Or it looked that way. We both had been adding material to both pages for awhile. Other editors seemed to be happy with both of our efforts. But then the last couple days you went back to deletion mode. In order to favor a POV. I don't favor any POVs. I like both IBC and Lancet. They both do great work. I am only interested in painting a broad picture with all significant viewpoints on casualty stats in Iraq. It is a complicated subject, and deserving of care, not POV-favoring, slashing, deletion, and misinformation, all while claiming others are POV-pushing. --Timeshifter 18:54, 7 January 2007 (UTC)

Timeshifter, you are lying. It does not come from ILCS, that's why you keep citing this paraphrase from an interview where Pedersen talks about "the UN". If it "comes from ILCS" quote it from ILCS. You can't because it doesn't come from ILCS. You know this and you're just lying about it, as you are about most everythign else above. You're a POV-pushing fraud who is now just lying flat-out.74.64.60.148 17:08, 8 January 2007 (UTC)

"Iraq Living Conditions Survey 2004". 2004 UNDP ILCS. United Nations Development Programme Iraq Living Conditions Survey.

From an October 19, 2006 Washington Post article ^[1] there is this:

"In a telephone interview, Jon Pedersen, research director for the 2004 study, said several factors probably account for researchers' different findings. One key issue is how researchers extrapolate from the deaths identified in their field research to a death toll for the whole country. Pedersen noted that the Lancet study is based on a pre-invasion mortality rate of 5.5 deaths per thousand people [per year]. The U.N., he said, used the figure of 9 deaths per thousand. Extrapolating from the lower pre-invasion mortality rate would yield a greater increase in post-invasion deaths, he noted."

"Is Iraq's Civilian Death Toll 'Horrible' -- Or Worse?". By Jefferson Morley, Washington Post, October 19, 2006.

The above link is the reference used. Quit deleting it. Also, please remove all your personal attacks in this section of the talk page, or I will report them to the Wikipedia:Personal attack intervention noticeboard. I suggest you also read the following wikipedia guidelines:

   * Wikipedia:Assume good faith
   * Wikipedia:Civility
   * Wikipedia:Talk page guidelines
   * Wikipedia:No personal attacks
   * Wikipedia:Personal attack intervention noticeboard

Also, I changed the section heading from

The supposed ILCS "baseline rate"

The ILCS "baseline rate"

From Wikipedia:Talk page guidelines: "Keep headings neutral: A heading should indicate what the topic is, but not communicate a specific view about it."

I suggest people go to my user page for a quick copy-and-paste list of the guideline links listed higher up. --Timeshifter 18:39, 8 January 2007 (UTC)

The source you're using refers to "the UN", not directly to ILCS. Just because the ILCS is one of many UN-based reports does not mean "the UN" therefore means ILCS. "The UN" estimated this rate before ILCS, and the ILCS did not use it for anything. Pedersen is saying that Lancet's "baseline" is much lower than what the UN has estimated (not what ILCS estimated). He is not saying the ILCS estimated it or used it. That is why you can't cite the ILCS itself saying this, and must 'interpret' this paraphrased comment in an interview to fit your theory.74.64.60.148 19:11, 8 January 2007 (UTC)

It's pretty clear: "Pedersen [2004 UN ILCS study author] noted that the Lancet study is based on a pre-invasion mortality rate of 5.5 deaths per thousand people [per year]. The U.N., he said, used the figure of 9 deaths per thousand. Extrapolating from the lower [Lancet] pre-invasion mortality rate would yield a greater increase in post-invasion deaths, he [Pedersen] noted." --Timeshifter 19:18, 8 January 2007 (UTC)

No, your interpretation is that "ILCS used" this in its estimates, and you do your own (again 'interpretative' "comparison" to Lancet based on that interpretation of yours. That's not what he says and that's nowhere in the ILCS itself, which is why you can't cite ILCS saying that anywhere and can only use and "interpret" this paraphrase (not even a direct quote) of Pedersen from an interview. He does not say this figure comes from ILCS and the figure is nowhere in ILCS. "The UN" has published all kinds of estimates regarding Iraq. ILCS is just one, and nowhere in ILCS does it say it estimated or "used" this "baseline rate" for anything at all.74.64.60.148 19:30, 8 January 2007 (UTC)

Another editorial invention of yours is to claim this is a "2002 baseline rate", while Pedersen doesn't say that either (or rather, isn't paraphrased as saying that). He's paraphrased as referring to some "pre-invasion mortality" not from any particular year, and thus not indicating "2002" in particular.74.64.60.148 19:39, 8 January 2007 (UTC)

Furthermore, you quote Les Roberts claiming the ILCS pre-invasion mortality rate is *lower* than Lancet's: "I suspect that Jon's mortality estimate was not complete. I say this because the overall non-violent mortality estimate was, I am told, very low compared to our 5.0 and 5.5/ 1000 /year estimates for the pre-war period". So you're making up that this 9 is an ILCS rate, you're making up that it's for 2002, and then you quote Roberts saying ILCS estimated pre-invasion mortality at less than 5! This is a total mess of editorializing assumptions and 'interpretation'74.64.60.148 19:44, 8 January 2007 (UTC)

Furthermore, in your quoting of what you've chosen to call "ILCS related questions" you quote the following: "The UN has as estimate of 10? Isn't that evidence of inaccuracy in the study?" LR: "The last census in Iraq was a decade ago and I suspect the UN number is somewhat outdated." So LR is saying he "suspects" this "UN" pre-invasion mortality rate is from ten years ago. Why does LR not know this "UN" rate is from ILCS and for 2002, as you claim is "pretty clear"?74.64.60.148 19:55, 8 January 2007 (UTC)

OK. I am taking out the word 2002. So the sentence in question will now say "pre-invasion mortality rate" instead of "2002 pre-invasion mortality rate". I will remove the confusing sub-rate material. --Timeshifter 20:49, 8 January 2007 (UTC)

Better, but this whole section is still garbage because everything you have in here about the "baseline rate" issues is entirely out of context if the topic is "ILCS compared to Lancet". That debate is a legitimate part of criticisms, but has nothing to do with a comparison of these two studies, which is what the section is supposedly about. But that's ok. It's at least only misleading and out of context now. The POV-tag on the article is enough. I'll leave it as is for entertainment purposes too.74.64.60.148 21:58, 8 January 2007 (UTC)

Graphs

The graph at the top of the page has been on this page for many months. It's also been on the Casualties of the Conflict in Iraq page for many months. New material has come to light related to errors in this graph, a graph which everyone has been seeing on these pages for months (and presumbaly believing to be accurate). I've added material showing that it's not, with citations to the Lancet. Timeshifter then suddenly deleted the whole graph.

His first excuse went that my citations were not in conformity with WP:RS because the Lancet website requires an account to access the articles in question. I pointed out that WP:RS doesn't say anything about this and such links are prevalent throughout wiki pages. He dropped this excuse.

Another excuse is that the graph has to be removed because it's erroneous. Or, because Lancet authors have conceded these errors, therefore the graph needs to be removed. The page is about the Lancet study. If parts of the study are erroneous, this information is relevant to the page. And how are people supposed to know the graph is erroneous, and Lancet authors have conceded this to be the case, unless Timeshifter stops deleting and suppressing this information from the page?

His next excuse he used on the Casualties page was that using the graph anywhere but the Lancet page would be a copyright violation. Yet it doesn't say this, and the same rationale which makes it Fair Use for the Lancet wiki page would make it Fair Use for the Casualties page, which both discuss the Lancet study at length. And this graph has been on both pages for months, while Timeshifter has done dozens if not hundreds of edits on both these pages over all these months. All of sudden, when the new material about its errors are added, he decides that the graph is some kind of copyright violation on the Casualties page.

When all these excuses fail, he then starts attacking me, accusing me of violating three-revert rules and then "vandalism" for not going along with his suppression efforts. Again, look at the course of the edits. I added new sourced material to a graph that has been on the page for months. Timeshifter then deletes the whole graph along with my added material using spurious excuses. I put it back. He deletes again with a different excuse. I put it back. He deletes again with another excuse. Who is the "vandal" here?

Timeshifter is now levelling all kinds of false accusations and smears, and trying to get the page 'protected' from so-called "vandalism" all because he is determined to ensure that this new information about these errors is concealed from the readers of these pages.

To add in a dash of hypocrisy, after Timeshifter used his first excuse that he had to delete the new sourced material I added (along with the graphic entirely) because it was in violation of WP:RS guidelines, I noticed that another graphic on this page certainly is in violation of WP:RS, namely this red, black and green graph under the "second study 2006" section. This graphic is purporting to represent data in the Lancet, "Bush/Military", and Iraq Body Count. Yet it's a homemade graphic by somebody not associated with any of these sources, the graphics have not been approved by any of these sources it's supposedly representing, and the graphic is not published in any reputable source. It was just a homemade graphic that somebody decided to upload to wikipedia. This clearly does not conform to WP:RS, yet Timeshifter keeps putting *this* graph back in, while deleting the actual Lancet graph, and the new material I added from letters about this graph in the Lancet, while claiming this doesn't meet WP:RS standards!

It seems to me Timeshifter just wants to make sure nobody sees this information I've contributed, he's run out of excuses, and now his desparation to conceal this information from wiki readers is leading him to just throw around false accusations and threats.74.73.39.219 23:42, 18 January 2007 (UTC)

This is a known multiple-IP vandal. If this vandal wants to paste in the Lancet letters here we can determine whether any of this merits further discussion. Otherwise it is just more of this person's vandalism. If anybody new to this page needs the URLs to the 4 different IP contribution lists and IP user pages of this vandal, just ask.

But most people reading this talk page know exactly of who I am talking about. And we need waste no more time on this vandal. He is deleting another graph that has no problems in order to make a WP:POINT. Another violation of wikipedia policies. Also he put an accusation and my user name in a talk page section title. I removed it. It is a violation of talk page guidelines.

Read the graph fair-use info to see that it should not be on the other page. I did not bother with it before because it was a minor violation. But now it is an admittedly erroneous graph.

This vandal is also breaking the wikipedia guidelines listed below once again in order to stir up trouble and waste everybody's time.

"Keep headings neutral: A heading should indicate what the topic is, but not communicate a specific view about it." A good quote from the "No personal attacks" guidelines: "Abusive edit summaries are particularly ill-regarded." From the "No personal attacks" guideline is this: "Do not make personal attacks anywhere in Wikipedia. Comment on content, not on the contributor."

   * Wikipedia:Assume good faith
   * Wikipedia:Civility
   * Wikipedia:Talk page guidelines
   * Wikipedia:No personal attacks
   * Wikipedia:Personal attack intervention noticeboard

If anybody knows how to semi-protect the article page from anonymous users, please do so. It will save a lot of time. Also, do not stop making other edits to the article. That is the main purpose of this vandal on several pages. To stop people from editing. We can keep removing his damage until he is blocked due to 3R violations, and eventually permanently blocked. --Timeshifter 09:51, 19 January 2007 (UTC)

I put a 3RR warning on the anonymous user's talk page: User talk:74.73.39.219. Here it is below, too:

These 2 IP contribution lists are his:

Note that one starts 27 minutes after the other ends on January 9, 2007. His old user talk page has the previous 3RR block for 24 hours:

http://en.wikipedia.org/wiki/User_talk:74.64.60.148 --Timeshifter 12:18, 19 January 2007 (UTC)

Timeshifter, you are lying. My IP has changed in the past. It does this by itself. I have no control over that. That does not support your fraudulent accusations of "vandalism". It is very common. Second, look at how you evade any discussion of the facts, and instead turn to ad hominems and smears.

The simple facts are this. I've put new factual sourced material onto a graph that's been on this page for months. You've made up several different bogus excuses to censor this sourced material. You keep deleting it, and I keep putting it back. Then you start lying and accusing me of "vandalism" for not going along with your censorship. This is ridiculous. You really should be banned for this type of behavior.74.73.39.219 13:40, 19 January 2007 (UTC)

I've also added a 3RR warning to Timeshifter. HE is the one repeatedly deleting/censoring other peoples edits. He is also guilty of violating that lists of "civility". He is the one calling me a "vandal" for putting factual sourced material onto the page and not letting him censor it. How do you maintian "civility" and "good faith" with someone who is lying and smearing you in order to repeatedly censor your factual, sourced edits?74.73.39.219 13:50, 19 January 2007 (UTC)

lancet excess mortality rate june 2006

i see that Timeshifter has corrected my figure of 19.8/1000 to 14.2/1000 deaths as of june 2006. i suspect the new figure is correct but if so then this page needs to be fixed, because it says 19.8/1000 near the bottom. Benwing 02:27, 18 January 2007 (UTC)

eh sorry, i meant to say that the corrected figure was in the template. Benwing 02:28, 18 January 2007 (UTC)

I tried to clarify the math a little in the article. 19.8 deaths/1000/year is the mortality rate for the last time period of the survey (June 2005-June 2006). The pre-invasion mortality rate was 5.5 deaths/1000/year. The difference between the 2 rates is listed as 14.2 in Table 3 in the Lancet article. That difference is the excess mortality rate. Why it is 14.2 instead of 14.3 may be due to how they rounded the numbers. See Table 3 in the article:

http://www.thelancet.com/webfiles/images/journals/lancet/s0140673606694919.pdf --Timeshifter 10:12, 18 January 2007 (UTC)

images and captions

Here is a compromise for now. I put back the user-created image. There is no problem with it since it is based on sourced data from the Lancet article. A very common practice on wikipedia. Data is presented on wikipedia in many forms. As long as it is sourced it is usually OK.

I am keeping the Lancet graph and the links for now. I am removing the words "erroneous" and "misleading" since I am not a subscriber to Lancet online. I do not know what is wrong with the graph. This way people who want to learn more, can subscribe to Lancet online. --Timeshifter 23:33, 22 January 2007 (UTC)

I don't believe this is adequate. With the homemade image, it's claiming to represent data from three sources. None of the three sources has approved these graphics as accurate. Furthermore, the graph is just three blobs sort of shaped like Iraq, with one being bigger than the other two. What is the value of this? Also, one of the blobs appears to be representing nothing but an off-hand comment by some general to a reporter talking about numbers he's seen floating around from other sources, and is not very clear in the graphic. Also, Iraq Body Count has no date so we don't even know what period their blob is supposed to represent. This clearly does not meet WP:RS, and is pretty worthless in any case.

With the Lancet graph, you can create an account for free on the Lancet website and read the letters. Just follow the link for "Free Registration". The existing caption is accurate to the letters. Your new cryptic caption just conceals the whole point of the exchange, which is that the graph is erroneous and misleading.74.73.39.219 00:11, 23 January 2007 (UTC)

The user-made image uses data from verfiable sources. At one pixel per death. Click the image to read about the one pixel per death. The sources are linked. You will soon be in violation of WP:3RR if you keep deleting the image again. --Timeshifter

3RR works both ways. It's hypocritical and disingenuous editing to constantly revert to your desired version and accuse other people of violations for doing the same. Furthermore, the blob for Iraq Body Count is not given for any particular date, nor does it say whether it's using the min or max or what. And the reference to to a 'Bush/Military' estimate is just misleading altogether. As said above, it appears to be based solely on a comment given to a reporter by some general who was just referring to numbers he'd heard floating around from other sources. This is not an estimate from Bush or the military. It's one person in the military telling a reporter he's heard other sources say such and such.74.73.39.219 01:50, 23 January 2007 (UTC)

I clarified the user-created image caption. --Timeshifter 01:59, 23 January 2007 (UTC)

Wikipedia does not take the word of editors about info from sites requiring registration or signup. So there can be no description of the info in those letters until you copy those letters here. Even then wikipedia does not normally allow such info to be used as a verifiable source. From Wikipedia:External links:

4.2 Sites requiring registration

"Sites that require registration or a paid subscription should be avoided because they are of limited use to most readers. Many online newspapers require registration to access some or all of their content, while some require a subscription. Online magazines frequently require subscriptions to access their sites or for premium content. If old newspaper and magazines articles are archived, there is usually a fee for accessing them." --Timeshifter

You've presented a truncated version. It continues saying, quote:

A site that requires registration or a subscription should not be linked unless:

   * The web site itself is the topic of the article, or
   * It has relevant content that is of substantially higher quality than that available from any other website.[end quote]

The "unless" clearly applies here as these letters are not available on any other website as far as I know. Thus they are perfectly appropriate. Second, as I said above you can register free and read the letters.74.73.39.219 01:50, 23 January 2007 (UTC)

After reading the letters it may be found that the errors in the graph are due to bad design, and that the solution may be to just jettison the graph from the wikipedia article. Unless there is some intent to deceive on the part of the designer of the graph. Which you have not shown by copying the letters here.

Wikipedia allows editors such as me or others to remove unsourced material. It is up to the editor providing the material to also provide reliable sources. You will soon also be in violation of WP:3RR concerning this if you keep putting back unsourced caption info such as "misleading" and "erroneous." --Timeshifter 01:24, 23 January 2007 (UTC)

These are not unsourced. They are sourced directly to the Lancet journal. You are making up an excuse and distorting wiki guidelines by truncating them, in order to change their meaning to make it appear as if the quoted material confirms your excuse. It doesn't.74.73.39.219 01:50, 23 January 2007 (UTC)

No one has to believe you or me. Just provide a copy of the letters here. I have little reason to believe you, because I have observed how you spin info in the past. If you can prove the words "erroneous" and "misleading" then those words stay in the caption. It is up to you, not me, to make *your* point. You haven't proven that the subscription info meets the higher standard of "relevant content that is of substantially higher quality than that available from any other website." --Timeshifter 01:59, 23 January 2007 (UTC)

Why don't you go read the letters? If you want them quoted here, go create a free account on the Lancet website and copy/paste the letters here if you like. And yes, I have proven it above. Show me the other website to use to cite the letters and I'll use it. There isn't any. You're making up excuses so you can censor facts that are inconvenient for your incessant POV-pushing and spin-doctoring on this and other pages, you're truncating wiki guidelines to distort their meaning in order to fraudulently intimidate other editors, and now you're starting in with insults again.74.73.39.219 02:09, 23 January 2007 (UTC)

What insult? I will give you a day to calm down, and copy the letters here. --Timeshifter 02:14, 23 January 2007 (UTC)

Removed erroneous Lancet article graph

My edit summary when removing this image: "Removed image since it is erroneous. Links in image caption do not meet WP:RS since they require subscription. Term 'misleading' may be unmerited. No way to know without better sources. See talk."

I put a wider version of the image below to aid in seeing it. Reference links made visible.

[Later note. Image:2nd Lancet Iraqi death count Figure 4.gif is a link to the graph image. The wikipedia non-free content policy now prevents fair-use images from being on talk pages. A bot removed the image. --Timeshifter 04:10, 15 May 2007 (UTC)]

--Caption begins: --

Figure 4 from the second The Lancet survey of Iraqi mortality, showing a comparison with two other mortality surveys. Two letters subsequently published in the Lancet journal challenged this graph as erroneous and misleading

and the authors of the study conceded these errors in their 'Authors' reply'

http://www.thelancet.com/journals/lancet/article/PIIS0140673607600634/fulltext

--end of caption--

Since the links are not reliable sources (due to need for subscription to see them), and since the image is erroneous for whatever reason, I don't see a reason to keep the image in the article. What would be the point? To show that people make mistakes when making graphs? --Timeshifter 19:04, 17 January 2007 (UTC)

WP:RS does not prohibit sources on the basis of the links requiring an account on the website. Such sources are prevalent throughout wiki pages. The citations are to the Lancet. And the point of showing this is, for one, so that when people read the Lancet report in which the graph is presented as accurate and informative, they might have some chance of knowing that it's not. There are plenty of other reasons, but it's for readers to make up their own minds about what relevance these errors might or might not have. As such, I've replaced this graph along with the citations and request you cease removing sourced material (and also stop pretending this is motivated by technicalties you've made up, rather than a desire to supress facts which you find inconvenient and unflattering).74.73.39.219 21:46, 17 January 2007 (UTC)

Put some quotes here on the talk page to show the relevance. Then we can put back up the graph if it merits it. I have found other minor errors in the article, but they don't merit being discussed in a wikipedia article with limited space. Do you want us to point out typos too? --Timeshifter 09:23, 18 January 2007 (UTC)

Also, please do not put the image on other wikipedia pages. Please go to the image page and read the fair-use rules:

http://en.wikipedia.org/wiki/Image:NonFreeImageRemoved.svg

It says that the image can only be used on the Lancet study wikipedia page. "Any other uses of this image, on Wikipedia or elsewhere, may be copyright infringement." --Timeshifter 10:21, 18 January 2007 (UTC)

First, it does not say the image can only be used on the Lancet study wiki page. I think the same rationale applies to the casualties page, in which the Lancet study is again the "object in question". Second, I did not "put the image on other wikipedia pages". That image has been on this and other pages for many months. I've simply updated the caption with new sourced information about it to where it already was. And you've been the most prolific editor on these pages for all these months and this graph has been there the whole time. It's only now that you - suddenly - want to supress and conceal it from wiki readers, grasping at these lame excuses.74.73.39.219 18:06, 18 January 2007 (UTC)

You are obviously the same anonymous vandalizing editor that has previously been blocked for 3R violation. You are in danger of it again. See WP:3R. I only left the image on the other page as a courtesy, and until someone else noticed it was violating the fair-use rule. But now that you say the image is erroneous it can not remain on any wikipedia pages. You haven't quoted from the letters to the editor to show why we should put erroneous graphs on the Lancet page. Still waiting. --Timeshifter 21:05, 18 January 2007 (UTC)

Timeshifter, you seem to be flailing about here making up excuses. This graph has been here and on the casualties page for months. I've added new sourced material about it. You've now made up several different lame excuses to try to supress this material. Putting source material back which you've deleted is no violation of any revert rule. You blanking out sourced material, and continuing to revert to your blanked out versions could be though. I think you should stop making up excuses and attacking other editors in order to control these pages and supress facts you happen not to like.74.73.39.219 21:30, 18 January 2007 (UTC)

I think this is around your 4th different anonymous IP address. It's boring. --Timeshifter 22:09, 18 January 2007 (UTC)

Lancet graph problems

The 2006 Lancet study household survey ~~and the Department of Defense (DoD)~~ estimate uses the mortality rate scale to the right. The Iraq Body Count uses the total-deaths scale to the left. The purpose of the graph according to the Lancet article is to illustrate the 3 studies' agreement concerning the increasing rate of deaths over time, even though all 3 studies come up with different death totals. The 3 studies use different methods, and they measure different types of deaths. {Later correction: The DoD graph line uses the vertical scale to the left. I struck out the incorrect info. --Timeshifter 17:41, 24 January 2007 (UTC)}

The problem I see with the graph is that the designer of the graph is using the vertical axis to show 2 things. It is an interesting way to show the trends, but it is confusing to most readers. For that reason I think the graph should be removed from the wikipedia article. I do not see it as being intentionally misleading, and so it does not merit discussion in a wikipedia article with limited space. Unless there is some logical objection I am going to remove the graph from the article.

Feel free to leave some relevant quotes here from the Lancet letters. The whole letters would be preferable so there is no suspicion of spinning the issue through the selective use of quotes. --Timeshifter 18:05, 23 January 2007 (UTC)

The problems with the graph appear to be (at least) two-fold. The first you discuss above, but it's difficult to see why you describe this as an "interesting way to show the trends". It isn't really showing the trends. The Department of Defense and Iraq Body Count "trend" lines are simply showing the rise in their total count as they've added deaths to their total count over the course of the war. The Lancet trend line is plotted as per-year rates. Thus, the Lancet line could go up or down or stay the same. The other two *have to* show a rising line, unless people start coming back to life and these two sources start lowering the cumulative counts. The Lancet report claims "corroboration" for its rate trends by comparing it to a rise in two cumulative counts, which have to rise. And since Lancet is independent of the other two, it can also be placed anywhere on the graph. In this case, the authors chose to show it with the DofD line overtaking Lancet, making Lancet look like the moderate source between the two extremes. Therefore, the Guha-Sapir letter says this is "misleading", which it clearly is. From this there's two things wrong in the peer-reviewed report, the graph and the claim in the text about 'corroboration' of Lancet's trends.

The second problem is that the DofD source is a count of casualties (deaths and injuries) so it shouldn't even be on this graph that's claiming to show deaths. In any case, it's way too high on the graph because it's assuming all the casualties to be deaths. From this again, there's two things wrong with the peer-reviewed report, the graph again, and another claim in the text where they say that the DofD put civilian deaths at 117 per day in the most recent period, when the data is both deaths and injuries and it's not just civilian either.

A third problem I see, not mentioned in the letters, is that they seem to have got the DofD figures wrong in the graph too. They're too high even if you assume they're all deaths and not 'casualties'. The Lancet graph says the DofD goes up to 60,000, but in the cited source it looks more like 50,000 to me. Also, the first-year period in the graph is way too high. It puts the DofD at the same spot as IBC, at around 15,000, but the DofD only has about 5,000 in that time period.

So that's three misleading errors in the graph, and at least two in the text of the report. This should remain on the page for several reasons. First, people reading the report will not see or hear of these errors. They are not corrected in the original report, just in these much more obscure letters that most people won't see. Thus the information on this page might help prevent more people from being misled by the report. Second, this information may raise some questions about how and why all these simple and obvious errors were not spotted by any of that rigorous peer review we read about on the page. Third, it may inspire further investigation into other factual claims and arguments in the report, which may also turn out to be erroneous and misleading like these.

For these and other reasons, the graph and this info should be available to readers here. Also, there is no claim one way or the other about all these errors being "intentionally" misleading. The letters don't guess at 'intent' and nor does the description here.74.73.39.219 01:52, 24 January 2007 (UTC)

In the Lancet article in reference to the graph the authors were discussing the escalation in the rate of deaths per year. The increase in the slope of the curves is the point of the graph. Not the cumulative number of deaths, nor the placement of the graph lines. It is a subtle point which I don't think you are getting. I don't think this particular graph is the best tool to express this point. --Timeshifter

Timeshifter, you're missing the point. The graph only shows an "escalation in the rate of deaths per year" for the Lancet study. That's the error Guha-Sapir is talking about and that's why it's misleading. And you're proving the point. It doesn't show rate trends for IBC or the DofD. What it shows is the number their count was by each year. So by April 2004 IBC's count was about 15,000. And by May 2005 its count was about 25,000. This *has to* show a rising line. The Lancet report claims "corroboration" for its rate trends by illustrating this rise in rates alongside a rise in two cumulative counts as they rise over time.74.73.39.219 08:00, 24 January 2007 (UTC)

~~The DoD graph uses the vertical scale on the right, not the one on the left.~~ We are both proving my point that the graph is subject to misinterpretation. I also mistakenly thought at first that the DoD graph line was using the vertical scale on the left.--Timeshifter {Later correction: looks like I had it right the first time. --Timeshifter 18:57, 24 January 2007 (UTC)}

The DoD one *is* using the axis on the left. The IBC and DoD are using "deaths" on the left. The Lancet is alone using the "deaths per 1,000 per year" rates on the right. It can't be using the axis on the right because the DoD source has about 50,000 "deaths" total. If it was using the axis on the right it would mean DoD had hundreds of thousands of deaths.74.73.39.219 08:00, 24 January 2007 (UTC)

You are correct on that one. --Timeshifter 18:57, 24 January 2007 (UTC)

There is a typo in the Lancet article. In the DoD sentence they start by saying "casualties" in the sentence and then end with "deaths". I don't see that same mistake in the graph though. It looks like the authors are using a ratio of one out of 4 casualties being deaths. The DoD casualty rate numbers are from this PDF:

http://www.defenselink.mil/pubs/pdfs/Security-Stabilty-ReportAug29r1.pdf --Timeshifter

It is not a typo. It's a misinterpretation of this source as being deaths. They say the error in the report and they illustrate the same error in the graph. And they did not use any ratio to weed out the deaths from injuries. That would mean DoD should be much lower, but they have it going up to about 60,000 "deaths" by the last period. If they used a ratio of 1/4 there'd only be 15,000 deaths in the DoD (more like 13,000 actually, if you correct their other error of inflating the DoD number by about 10,000). 74.73.39.219

The DoD numbers are discussed and referenced in the Lancet article:

"The US Department of Defence keeps some records of Iraqi deaths, despite initially denying that they did.4 Recently, Iraqi casualty data from the Multi-National Corps-Iraq (MNC-I) Significant Activities database were released.5 These data estimated the civilian casuality rate at 117 deaths per day between May, 2005, and June, 2006, on the basis of deaths that occurred in events to which the coalition responded."

Reference 4: "Tavernise S. US quietly issues estimate of Iraqi civilian casualties. New York Times, Oct 30, 2005."

article link The NY Times article mentions that the IBC uses a ratio of one out of 4 casualties being deaths.

Reference 5: "Department of Defence. Measuring stability and security in Iraq. Washington: Department of Defense, 2006."

http://www.defenselink.mil/home/features/Iraq_Reports/Index.html

It does not matter what percentage of the casualties the authors used to figure out the number of deaths in the DoD stats. As long as they consistently used that same percentage for the whole time period. The purpose of the graph is to show the trend, not the totals. The trend being the increase in the rate of deaths over time. The increasing steepness in all the graph lines. --Timeshifter

They didn't use any percentage. They misinterpreted the source as deaths. And the DoD line is *not* showing rates, as discussed above. To show rises in rates is what the graph appears to show at first, and what it claims to show. But that's not what it shows.74.73.39.219 08:00, 24 January 2007 (UTC)

Your selective info from the Guha-Sapir letter is questionable. The theory about the upper or lower placement of the graph lines is just speculation by you and the letter writers. And I am taking your word for it anyway which means that it is all speculation since I have not read the Lancet letters. --Timeshifter

Why don't you create a free account on the Lancet website and read them? 74.73.39.219

In particular that the placement of the graph lines is misleading. What did the authors say in reply? That there was no intention to mislead? Or that the placement of the graph lines meant nothing? That is my understanding of the graph. The point of the graph lines was to show an increase in slope, not the end point, nor which line was above the other. The end point is irrelevant to the purpose of the graph. --Timeshifter

That's why it's misleading. The two cumulative counts have to show a rising slope, unless people stop dying altogether (in which case the slope would go flat), or start coming back to life (in which case the slope would start to tilt downward). Even if *rates* go down the two counts would still show a rising slope. That's why the thing is so misleading (in addition to just getting DoD wrong).74.73.39.219 08:00, 24 January 2007 (UTC)

Maybe you can copy the complete letters here. You still have not shown any intentional deception on the part of the authors. In fact there are no errors shown by you in the graph. The only thing shown is what I have already noticed. That the graph tried to do too much with the vertical axis. So it is very confusing and subject to misinterpretation. We both misinterpreted the graph as concerns which vertical scale to use with the 3 estimates of deaths.

So if you want to keep the graph and the letter links, then fine. But I will be removing the wording of "misleading" and "erroneous". There are no errors in the graph. There is no proven misleading by the authors. There is no proven misleading by the graph. It is just a graph that is subject to misinterpretation.

So if you want we can put this: "The graph is subject to misinterpretation. See these letters to the Lancet editors." --Timeshifter 03:58, 24 January 2007 (UTC)

The graph is misleading and it's also erroneous. If one somehow manages to correctly interpret the graph they would have to do so against the descriptions and its appearance. IOW, they'd have to interpret the graph to be not showing what the report claims it's showing and what its looks suggest it's showing. It's subject to misinterpretation because it's presented as something it's not, which (mis)leads readers to the wrong interpretation. In addition, even if you interpret what's being shown for DoD correctly, you'd still be misled, because what it's claiming for DoD is just wrong.74.73.39.219 08:00, 24 January 2007 (UTC)

It looks like I had it right the first time. The DoD graph line is using the vertical scale to the left. Proves my point that the casual reader can easily misinterpret the graph unless they have personally crunched the numbers. The point of the graph is the increasing slope of the graph anyway. I will rewrite the caption and explain the graph lines more clearly. There are no errors in the graph itself. The error was just in my changing interpretation of which graph lines are assigned to which vertical axis. I will explain this fully in the caption and I will remove any implication that this was intentionally misleading by the authors. --Timeshifter 17:35, 24 January 2007 (UTC)

Here is a quote from page 34 of the August 2006 DoD report:

http://www.defenselink.mil/pubs/pdfs/Security-Stabilty-ReportAug29r1.pdf

"According to MNC-I reporting, civilian casualties increased by approximately 1,000 per month since the previous quarter. Executions in particular reached new highs in the month of July [2006]. The Baghdad Coroner’s Office reported 1,600 bodies arrived in June and more than 1,800 bodies in July, 90% of which were assessed to be the result of executions. This is due to increased targeting of civilians by al-Qaeda in Iraq and the increase in death squad activity."

Do the math and you will see that the chart of average daily casualties on page 32 is for deaths alone. --Timeshifter 19:07, 24 January 2007 (UTC)

I have done the math. I suggest you read the letters. It explains all this. I suggest you also read the article cited as footnote 4 in the Lancet report where they first claim the DoD estimated "deaths per day". You can read the article here: http://www.michaelmoore.com/words/latestnews/index.php?id=4703.

Some quotes: "BAGHDAD, Iraq, Oct. 29 - In the first public disclosure that the United States military is tracking some of the deaths of Iraqi civilians, the military has released rough figures for Iraqis who have been killed or wounded by insurgents since Jan. 1 last year. The estimate of dead and wounded Iraqi civilians and security forces was provided by the Pentagon in a report to Congress this month."74.73.39.219 19:51, 24 January 2007 (UTC)

That 2005 New York Times article made the same mistake we did. That is why that NY Times article also wrote:

"Indeed, the tally is lower than the 11,163 deaths of Iraqi civilians in the war during the same period counted by Mr. Dardagan's group [Iraq Body Count], which draws its data from reports of deaths and injuries by news services, newspapers and other news outlets. It is also lower than figures released by Iraq's Interior Ministry showing that 8,175 Iraqi civilians and police officers had been killed by insurgents from August 2004 through May 2005." --Timeshifter 19:59, 24 January 2007 (UTC)

"We" didn't make a mistake. The Lancet authors made the mistake, and they concede this in the letters. You're making the mistake all over again, even with the evidence in front of you. But you're not reading it right. What you quote above is after Tavernise applies a 1/3 ratio of dead/injured to arrive at a smaller figure out of the "casualties" which refers to deaths and injuries in the DoD chart. The DoD chart refers to deaths and injuries, while the Lancet report says they're all deaths, and illustrates them as all deaths, while also inflating the numbers on top of it.74.73.39.219 20:21, 24 January 2007 (UTC)

See my replies in the section below. --Timeshifter 21:12, 24 January 2007 (UTC)

why is an incorrect graph in the article?

i can't really follow what the above discussion is about, but everyone seems to agree that the graph is wrong. so why is it at the top of the article? either [a] it shouldn't be present at all, or [b] it should go down in the criticism section as an example of something that the lancet authors screwed up, with specific discussion of this (if this is indeed the point of including the graph; i'm not sure about this). wrong information should not be presented as correct! Benwing 19:19, 24 January 2007 (UTC)

well, Benwing that's why i added the captions explaining that it's wrong information, so that it is no longer being presented as correct (as it has been for the last few months). Your (a) isn't appropriate because the page is about the Lancet study. If the Lancet study presents wrong information this should be explained on the page about the Lancet study.74.73.39.219 20:09, 24 January 2007 (UTC)

The info in the graph is not wrong. It is just confusing due to its presentation. The most damaging thing that could be said about it is that it is unintentionally misleading. The info in the new subsection for the graph clears up this problem, and it links to the letters that pointed out the problem. --Timeshifter 20:36, 24 January 2007 (UTC)

The DoD info in the graph is flat wrong. Also the claim that the graph is a comparison of trends is wrong. Your new 'info' does not clear up the problem. It adds new errors on top of the old.74.73.39.219 20:45, 24 January 2007 (UTC)

OK. Now that you finally backed up your claims with actual sourced info consisting of quotes from the letters, the issue is resolved. Why could you not have done that to begin with? And the DoD info is wrong in both places; the graph, and in the DoD report itself. The DoD's own numbers are inconsistent. --Timeshifter 21:05, 24 January 2007 (UTC)

I don't know why you chose not to read the letters for yourself. I also do not see why you say the DoD's report is wrong or inconsistent.74.73.39.219 21:24, 24 January 2007 (UTC)

I don't know why you choose to ignore wikipedia guidelines about sourced info. It is up to you to source the info you want to put in an article. Not me. --Timeshifter 21:57, 24 January 2007 (UTC)

btw why don't we have a graph comparing the total deaths given by the lancet surveys and the various other sources (ibc, recent un, etc.)? afaik this is the most relevant info, and it clearly makes the point about why the lancet surveys are controversial: that their numbers are much greater than the numbers of the other sources. if there's no existing graph of this nature, make one up; that's perfectly acceptable as long as the info in the graph is clearly sourced and the graph presents the info neutrally. Benwing 19:19, 24 January 2007 (UTC)

I agree. I will move the confusing graph image to the criticism section. I will move the other image to the top since it illustrates your point about the wide variation in estimates. --Timeshifter 19:36, 24 January 2007 (UTC)

The graph has been at the top of this article for many months. A few days ago, I just added the new material from these Lancet letters which point out these errors. If you guys want it moved to criticism, that seems ok to me.74.73.39.219 20:03, 24 January 2007 (UTC)

As for Benwing's suggestion about another graph, it seems to me the huge disparity is apparent just looking at the numbers. You don't really need a graph. But if there were a graph it would need to be done properly, unlike the other image Timeshifter is referring to, which is illustrating some dateless IBC figure from who knows when, and is illustrating a comment to a press reporter by a general mentioning figures he's seen in other sources. It claims the latter is a Bush/Military estimate when it's probably IBC or LAT and/or UN. And UN doesn't appear, nor do the LAT figures.74.73.39.219 20:03, 24 January 2007 (UTC)

I've added quotes from the letters to the page. The Guha-Sapir letter also provides a corrected graph of IBC compared to Lancet using 'deaths per 1,000 per year' using three points like the original, but without mixing rates and counts. The new graph suggests no similarity between the trends of IBC and Lancet. However, the Lancet authors claim that a scale different than the one used in the original and then in Guha-Sapir's corrected version would show some similarity, if looking at it from a 'per month' level. This graph should probably be added and discussed, but I'm not sure how to add images to wiki. The image is here: http://www.thelancet.com/journals/lancet/article/PIIS0140673607600610/fulltext 74.73.39.219 20:51, 24 January 2007 (UTC)

Well finally you put some quotes from the letters instead of just your interpretations. That is why wikipedia is based on sourced material, and not on original research. And my main point is still true. The purpose of the original graph was to show an increase in the rate of deaths over the years. Few question that anymore. Here is recent IBC info below. For the week ending Dec. 31, 2006 the IBC reports [3]:

"It was a truly violent year, as around 24,000 civilians lost their lives in Iraq. This was a massive rise in violence: 14,000 had been killed in 2005, 10,500 in 2004 and just under 12,000 in 2003 (7,000 of them killed during the actual war, while only 5,000 killed during the ‘peace’ that followed in May 2003). In December 2006 alone around 2,800 civilians were reported killed. This week there were over 560 civilian deaths reported." --Timeshifter 21:05, 24 January 2007 (UTC)

Yes, look at IBC's per year rates. The 2nd year is *lower* than the first. If IBC's per-year "rates" were shown in the graph they would show a trend line that is higher on the left and right, and which dips down in the middle. It would not show a steadily rising line from left to right like the Lancet. This is the point that the Guha-Sapir letter illustrates in their graph. When rates are compared to rates they don't really "corroborate" the Lancet findings. There are inconsistencies between the two. 74.73.39.219 21:24, 24 January 2007 (UTC)

There is no dip in the the Lancet graph for IBC because it starts at the total for the first year (which includes the invasion). This way the invasion deaths do not mess up the curve. There is a fairly steady increase in the yearly and monthly rates of death after "Mission Accomplished". So the increasing steepness of the slope of the DoD and IBC curves corroborate the Lancet results. That was the whole point of the graph to begin with, and even with the correction of the DoD numbers, it is still true. --Timeshifter 21:55, 24 January 2007 (UTC)

You are still completely confused by this graph. There is no dip in the Lancet graph of IBC because it is not illustrating IBC rates. It is illustrating the rise in the sum total of IBC's count from year to year. The first period shown for IBC shows about 15,000. The second period for IBC shows about 25,000. IBC did not have 25,000 in the second year alone. Year 2 in the graph is showing year 1 *combined* with year 2 (both years together in IBC make about 25,000). And year 3 is showing year 3 combined with year 1 and year 2 (all three years together in about make about 40,000). IOW, it's illustrating the rise in IBC's total, cumulative, count by each year. It's not illustrating IBC rates.

Only the Lancet line is illustrating rates. According to the Lancet line, the death rate from lowest to highest is: Year 1, Year 2, Year 3. That's not what IBC rates would show. IBC's data would show, lowest to highest: Year 2, Year 1, Year 3. And IBC's line would therefore not look like the Lancet line. It would have a dip in the middle. Showing this correctly does not "mess up the curve". It messes up the "corroboration" claim because that claim is messed up.

This shows clearly that there is inconsistency between the trends in the two sources. Lancet suggests that year 2's death rates surpassed year 1's by a very substantial margin. IBC's rates suggest the opposite.74.73.39.219 22:31, 24 January 2007 (UTC)

You still don't understand how the increasing steepness of the slope of the IBC graph shows the increasing mortality rate. Yes, the IBC graph line is showing cumulative totals. It is supposed to. Research how graphs work, and you will see what I am talking about. Lancet articles are written mostly for a specialist audience who would understand the math involved.

Timeshifter, you are completely confused by this graph. What you say here is gobbledygook.74.73.39.219 23:35, 24 January 2007 (UTC)

For the week ending Dec. 31, 2006 the IBC reports [4]:

IBC yearly death totals (not counting the initial invasion deaths):

2003: 5,000
2004: 10,500
2005: 14,000
2006: 24,000

The mortality rate increased each year. --Timeshifter 22:59, 24 January 2007 (UTC)

No, the mortality rate increased each year if you cherry pick from IBC data and pretend the invasion didn't happen. Even doing it that way, the degree of similarity there might be is unclear. It looks like Lancet's curve would rise much faster than IBC's. The IBC and Lancet lines would not closely track each others' slopes the way the do in the original. Lancet would be much steeper. It would hardly be reasonable or honest to claim that IBC data "corroborate our ﬁndings about the trends in mortality over time." by a cherry-picked comparison that still isn't a very similar rate of increase.

IBC yearly totals should be: 12,000 - 10,500 - 14,000 - 24,000. The mortality rate *did not* increase each year. Year 1 has higher mortality in IBC than year 2. Lancet shows year 2 being much higher than year 1. That's a big inconsistency and their trend lines should not look anything like they do on the graph. There seems to be some similarities and some differences with their trends, but what is clear is that the Lancet graph, by mixing rates and counts, creates the impression that IBC trends closely track Lancet trends, and that is not the case.74.73.39.219 23:35, 24 January 2007 (UTC)

Also, recall that the invasion did not happen until March 2003, so really the figures given in your IBC quote are comparing 12 months (2004, 2005 and 2006) to 9 months (2003). So 2003 is only for 75% of a year. If you account for this, that would mean IBC 2003 *rates* are equivalent to about 15,000 over a full 12 months. That means IBC's 2003 rates are higher than both their 2004 *and* 2005 rates. Only 2006 eclipses IBC's rates in 2003, making it even less consistent with Lancet trends.74.73.39.219 23:45, 24 January 2007 (UTC)

Timeshifter is now trying to put his argumentation above as the encyclopedia entry. He is employing a new truncated and selective comparison (not illustrated in the graph) to try to give an argument of his own about the trends. He slips in the phrase "not including invasion phase deaths" just for IBC, but this new cherry-picked version of the trends is not what's illustrated in the graph. Both Lancet and IBC include the invasion. His editorial argument is misleading and selective, and not appropriate for the page. Let the letters speak for themselves.74.73.39.219 23:54, 24 January 2007 (UTC)

You are the one cherrypicking. You yourself pointed out that the author's reply stated that a month-to-month comparison of rates or death totals would show the increase in the mortality rate. I am going to delete the whole section if you keep reverting selectively in order to favor a certain POV. Stop deleting sourced data. If you put in a quote from the authors about the month-to-month comparison then it would be an NPOV presentation. I am now agreeing with Benwing that the whole section should be deleted. I am going to delete it all now. The section serves no purpose. It was a simple error on the part of the authors. No deception involved. --Timeshifter 00:11, 25 January 2007 (UTC)

I am not cherry-picking. I'm excluding your own new editorial argument which is cherry-picking from IBC data to try to give your own argument. If you want to quote from the letters go ahead. There is a part in the Guha-Sapir letter showing a corrected graph, showing the yearly rates for IBC alongside Lancet, and they don't corroborate at all. They show exactly what I said. The Lancet authors reply suggests that if you did a different comparison - which they didn't do - there would be some similarity with parts of the data. But they don't illustrate this, and this doesn't negate the fact that the yearly rates are not similar. As the page stands now it doesn't give an argument about this. It just points out the errors.

I did not include the other graph or this part of the exchange because I don't know how to upload the corrected graph image in the Guha-Sapir letter. It would be too unclear to explain without this. Without this, the quote from the Lancet authors doesn't make sense because it references this graph. Guha-Sapir claims the two are not similar. The Lancet authors claim they have some similarity if you do an entirely different comparison than the one they did - one not looking at per-year rates, but looking at certain months while excluding others. I didn't include either argument. I didn't include Guha-Sapir's claim, nor the Lancet author's claim. You would need the corrected graph on the page for that to make sense.

What you want to did was introduce editorialized apologetics (POV) which tries to shift the goal-posts of the comparison to something else to give an argument that it still shows some kind of similarity in some way. Maybe that could be done, but only if outlining the differences too. But then all this would be original research by wiki editors: NOT sourced material.74.73.39.219 00:34, 25 January 2007 (UTC)

I'm getting rather sick of Timeshifter's deceitful editing. He's now removed the Lancet graph and all the discussion, citing an "agreement" with Benwing to do this. Benwing wrote: "either [a] it shouldn't be present at all, or [b] it should go down in the criticism section as an example of something that the lancet authors screwed up, with specific discussion of this"

Timeshifter is pretending Benwing said (a) and not (b). Option (b) is what we did. And he doesn't like the fact that i deleted his cherry-picked original research.74.73.39.219 00:48, 25 January 2007 (UTC)

PS. I'm also going to use an account. I will be Seigfried.74.73.39.219 00:48, 25 January 2007 (UTC)

I may have assumed too much on Benwing's part. Ease up on the personal attacks or I will report you to the personal attacks board. After Seigfried4220 returned part of the section to the article I also returned the sourced parts of the section that you, Seigfried4220, had previously deleted under your anonymous IP address. Glad you finally got a user name:

http://en.wikipedia.org/wiki/Special:Contributions/Seigfried4220

You first contributed with that user name on January 19, 2007. Why haven't you used it here since then? --Timeshifter 01:35, 25 January 2007 (UTC)

Similarities and Differences in Lancet and IBC trends

Timeshifter is trying to put in a new argument illustrating a similarity in IBC and Lancet trends. If you choose to exclude the invasion (March-April 2003) from IBC, then (and only then) does IBC show a steady rise year on year. The claim is that the two sources show the same trends, but Timeshifter is trying to hold this claim together by just chopping stuff out of IBC which doesn't conform to the claim. While it may be easy to prove claims if you're allowed to just conceal any data which contradicts the claims, it hardly seems credible or honest.

His comparison conceals the fact that that IBC's data shows 2003 rates are higher than 2004 or 2005 rates, while Lancet's full data suggests 2003 rates are the *lowest*. That's a big difference in the trends concealed by Timeshifter's version.

The Lancet graph compares per-year rates. If doing that, IBC's trends are not the same. They are very different. If you start taking stuff selectively out of IBC and doing a different comparison - one not done in the Lancet, then yes, you can see some similarity too. But there's still the differences as described above.

And even if doing it Timeshifter's way there seems to be a pretty big difference in the *degree* by which the sources rise. Just because both sources rise to some degree or other doesn't mean they "corroborate" each other's findings about trends over time. It looks to me like Lancet's trends increase much faster than IBC's. Thus, if plotted on a graph, and even if using Timeshifter's cherry-picked version of IBC data, IBC's line would still not closely track the slope of Lancet's line the way it appears to do in the original Lancet graph. The two lines would both rise to some degree, but they would diverge sharply in terms of the degree of rise and suggest much different rates of increase.

Also, comparisons need to be consistent. The year by year periods Timeshifter is using from IBC are not the same periods illustrated in the Lancet graph. The comparison in the Lancet graph uses 13 month periods starting in March 2003. The IBC figures Timeshifter is using use a 9 month period (2003) and then three 12 month periods (2004, 2005 and 2006). And the last six months of the 2006 period in IBC is not in the Lancet study, because it was done in June 2006.

If a wiki editor were going to try to tackle all this and try to give original comparisons it would be difficult and a lot of calculations would be required to show the proper trend lines. And It would be all original research on our part. It would be 'interpretation' by wiki editors, not sourced facts.74.73.39.219 01:23, 25 January 2007 (UTC)

You yourself said that the Lancet authors said in their letter that a month-by-month graph would show their points. But you haven't put that quote here yet. You only selectively quote. Why not just copy the whole letters here? And why are you using a user name and an anonymous IP to revert the article? You are now close to another WP:3RR violation. --Timeshifter 01:40, 25 January 2007 (UTC)

I've created an account. My connection went down earlier and re-set, which seems to reset my IP address. I'll try from now on to post using this account, but i sometimes seem to get logged out by mistake if my browser gets locked up and I have to close it. Regardless, I've already explained why I didn't quote all the comments about this from the letters. I did not quote this part from the Guha-Sapir letter nor the Authors reply. Nor did I say that the Authors said that a "month-by-month graph would show" anything. I'd be happy to do so but in order to do so I'd need to upload the graph in the Guha-Sapir letter, which I've not figured out how to do. If you can do that, please do. Follow the links, create a free account on the Lancet website and read them. You're pretending that I have access to secret material. I don't. You're just *choosing* not to go read it in order to be deliberately obtuse and create confusion.Seigfried4220 03:02, 25 January 2007 (UTC)

Lancet graph in other articles.

As Benwing wrote previously about placing the graph in this article: "[a] it shouldn't be present at all, or [b] it should go down in the criticism section as an example of something that the lancet authors screwed up,"

The same logic holds true for putting the graph in Casualties of the conflict in Iraq since 2003. There is no devoted criticism section in that article. Criticisms are blended into the article. Putting that graph in that article without the explanatory info will just confuse people. And it would be a needless duplication of the same material. --Timeshifter 22:16, 24 January 2007 (UTC)

I don't think there was anything there to "confuse" people, aside from the graph itself, which the new caption correctly explained was erroneous and misleading, which would help prevent the confusion. But I'm ok with taking it off that page.Seigfried4220 03:42, 25 January 2007 (UTC)

User-created graph illustrating wide variation in estimates

Independently-created graphical representation of the October 2006 Lancet survey estimate on the left. Two other October 2006 estimates are on the right; Iraq Body Count project (IBC) and Bush/General Casey. Deaths are represented at one death per pixel. Click the image to enlarge it, and to read the text on the image. The Lancet chart colors represent men (red), women (gray), children (green), and elderly (black). An October 12, 2006 San Francisco Chronicle article [1] reported: "Asked at the news conference what he thinks the number is now, Bush said: 'I stand by the figure a lot of innocent people have lost their life.' At a separate Pentagon briefing, Gen. George Casey, the top U.S. commander in Iraq, said that the [Lancet] figure 'seems way, way beyond any number that I have seen. I've not seen a number higher than 50,000. And so I don't give it that much credibility at all'." The maximum IBC estimate of Iraqi civilian deaths as of October 2006 was 48,693 [2]. That is 50,000 when rounded off. Thus the 2 graphs on the right are exactly the same.

I want to continue the discussion about whether this graph to the right should be at the top of the article. --Timeshifter 22:20, 24 January 2007 (UTC)

My vote would be that this graph should not be on the page at all. It's a sloppy graph and it's misleading. There's no date for IBC so we don't even know what period it's illustrating. And the "Bush/Military" estimate is referring to a comment by a general to a reporter, in which he said he's seen figures like 50,000 floating around from other sources, and not seen higher ones. He didn't create any estimate, he's just telling a reporter about other peoples' estimates he's seen and saying he thinks the Lancet one is wrong. Graphing this as if it was an independent estimate coming from Bush/Military or whomever is misleading. Furthermore, the caption is misleading about Bush. What "figure" is Bush standing by? The only figure he ever gave was 30,000. The graph seems to be illustrating a figure of 50,000. So what relevance does this have to Bush? It doesn't seem to me there needs to be any graph, but certainly not like that one.74.73.39.219 22:43, 24 January 2007 (UTC)

The IBC graph is the exact same graph as the Bush/Casey estimate. So both graphs on the right side represent around 50,000 deaths at one pixel per death. The IBC estimate in October 2006 was a maximum of 48,693. See:

http://news.bbc.co.uk/2/hi/middle_east/6045112.stm

So that is 50,000 when rounded off. --Timeshifter 00:18, 25 January 2007 (UTC)

i don't like this graph. for one thing, it's out of date, and (as 74.etc said) is mixing off-the-cuff statements with real data. for another, it is clearly trying to substantiate the lancet figures over the others, by claiming other estimates as "previous" (implying no longer valid) and placing the lancet estimates in the middle, all nicely colored and such. also, it is clearly trying to appeal to people's emotions by using the shape and colors of iraq and splitting out children and elderly. if you want to make a graph, it should include all and only the substantive, current estimates (the ones in the template of iraqi casualties) and be of the straightforward type with bars or with points connected by lines. if you want to split out any subfigures, use civilians vs. soldiers and use type of death (violent vs. non-violent, crime vs. military attack, etc.). Benwing 00:32, 25 January 2007 (UTC)

I see your points. Even though the graph is accurate in my opinion, it does POV-favor the Lancet graph in the ways you describe. I am deleting the graph from the article. --Timeshifter 00:34, 25 January 2007 (UTC)

Benwing, one of the problems with such a graph is the issue of "current". Different figures have come out at different times and cover different periods. The main ones I know are:

IBC - 60,000 (March 03 - Jan 07 - 'current')
LAT - 50,000 (March 03 - Jul 06 - 'out of date', may not update)
UN - 34,000 (2006 only - using Iraqi gov figures, may have other figures for some earlier periods buried in the reports somewhere - 'current')
ILCS - 24,000 (March 03 - May 04 - 'out of date', may not update)
Lancet - 655,000 (March 03 - Jun 06 - 'out of date', may not update)
Iraq Interior Ministry - 14,000 (2006 only, may not update)

The problem is that it's hard to get consistent comparisons in terms of the different time-frames. It's also hard in that some sources include or exclude certain kinds of deaths while others don't. This stuff would be difficult to get into a graph that maintains some kind of consistency in the comparisons. And inconsistent comparisons are just misleading. Better would be just a list like above. From that people can see the disparity.Seigfried4220 03:30, 25 January 2007 (UTC)

I am the author of this graphic. It appears that sometime since I posted it to the Commons, it's embarked on an interesting failed marriage with this article. May I ask a few questions about my poor child?

The use of the word "Previous" seems to be the sharpest criticism of the graphic, as it implies a superior estimate rather than concurrent estimates. Perhaps "Contemporary" would have been a better word choice.
It would be silly to reject a comparison because the data was obtained by different means. To take a Pentagon estimate given by a General in the United States Army at an official press conference as inaccurate opens a different can of worms. The Iraq Body Count number used is consistent with the timeframe of the period covered in Lancet II.
I'm sorry, but the Point of View argument simply doesn't hold water. The graphic is an illustration of the data in Lancet II. Lancet II is naturally biased in favor of Lancet II. Lancet II attempts to estimate the increased rate of human death in Iraq since the onset of the second Iraq War. I don't like its conclusions either, but the graphic is an accurate representation of the study's data, and an accurate comparison with competing estimates from Iraq Body Count and the Pentagon. This is an article which benefits from comparisons to multiple estimates but it is not the article about the plurality of estimates -- it's the article about Lancet II.
This is a graphical representation of the data with a pictographic key. Embodying my main criticism of the infographics now popular in periodicals, textbooks and television news, nothing within the graphic is superfluous -- no texturing, no glass effects, no perspective lettering, etc. This is about as stripped-down as anything short of an Excel graph will ever get. Is the very idea of an infographic (as opposed to, again, an Excel graph) antithetical to Wikipedia?
I've just read this article for the first time today, and I can't help but notice that it devotes more lines to criticism of the subject than it does to describing the subject. Not even the Fascism article can claim that. Needs editing. SpaceToast 06:55, 4 February 2007 (UTC)

Infographics are welcome in wikipedia. It is just another method of presenting sourced info. The key is that the info sources must be referenced. That can be done in the caption. I had removed the graph from the article due to the points mainly raised by Benwing. In my mind you have adequately answered all of those points. I have no problem with the graph being put back in the article. I would suggest clarifying as much as you can in the caption of the image in order to address Benwing's points. For example; you could say, "All estimates are highly disputed". That should overcome some readers' inclinations to read that the Lancet estimate is superior due to the wording "previous estimates" for the graphs on the right side of the image. Also, feel free to upload revised versions of the image. That is allowed, and people often upload revised versions of their images as improvements are made. --Timeshifter 18:09, 4 February 2007 (UTC)

Erroneous Lancet graph

Here below is the section in the article as it stands now. Go to the article to follow the reference links. --Timeshifter 01:19, 25 January 2007 (UTC)

Figure 4 from the second The Lancet survey of Iraqi mortality, showing a comparison with two other mortality estimates. Two letters subsequently published in the Lancet journal challenged this graph as erroneous and misleading ^[2] ^[3], and the authors of the study conceded these problems in their 'Authors' reply'^[4]

Figure 4 from the October 2006 Lancet survey of Iraq War mortality, showing a comparison of 3 mortality estimates. Two letters subsequently published in the Lancet journal challenged this graph. ^[2] ^[3]

The authors of the study replied ^[4]: "Josh Dougherty and Debarati Guha-Sapir and colleagues all point out that figure 4 of our report mixes rates and counts, creating a confusing image. We find this criticism valid and accept this as an error on our part. Moreover, Dougherty rightly points out that the data in the US Department of Defense source were casualties, not deaths alone."

The purpose of the graph according to the Lancet article is to show that these 3 mortality estimates all indicate an increasing rate of deaths over time. For example; for the week ending Dec. 31, 2006 the IBC reports ^[5]:

IBC yearly death totals (not counting the initial invasion deaths):

2003: 5,000
2004: 10,500
2005: 14,000
2006: 24,000

The mortality rate increased each year.

For both mortality rate graphs, and for graphs of cumulative death totals, the increasing mortality rate is shown by the increasing steepness in the slopes of all 3 graph lines. Accompanying the graph is a claim that "the similar patterns of mortality over time documented in our survey and by other sources corroborate our ﬁndings about the trends in mortality over time." This is difficult math even for the specialist readers of The Lancet journal, because it is not the individual data points in the graph that are being compared. It is the increasing steepness of the graph lines that are being compared.

Confusion is caused because the Iraq Body Count project (IBC) and Department of Defense (DoD) slopes are not illustrating "rates", but rather the rise in their cumulative totals over time, and are plotted along the "Deaths" axis on the left. Contrarily, rates for the Lancet are plotted independently using the "Deaths per 1,000 per year" axis on the right. It is not very clear in the image caption as to which vertical axis is assigned to which graph line.

A letter by Debarati Guha-Sapir, Olivier Degomme and Jon Pedersen explains: "Burnham and colleagues' figure 4, in which cumulated Iraq Body Count deaths parallel their study's mortality rates, is misleading. Rates cannot be compared with numbers, much less with cumulative numbers."

A second letter by Josh Dougherty explains that the DoD figure is misrepresented: "Burnham and colleagues' assertion that the DoD 'estimated the civilian casualty rate at 117 deaths per day' is mistaken, as is their figure 4, which repeats this error in graphic form. These data refer to Iraqi civilians and security-force personnel, not just to civilians, and to casualties (ie, deaths or injuries), not just deaths."

The last DoD report before the October 2006 Lancet study came out was the August 2006 DoD report.^[6]^[7] There is a timeline chart on page 32 of "Average Daily Casualties."

--end of section in the article-- --Timeshifter 01:20, 25 January 2007 (UTC)

As you know. I've disputed your inclusion of selective, cherry-picked comparisons which conceal the differences in the trends and constitute POV and original research. I've deleted your editorializing and interpretation and returned the section purely to a description of the letters.74.73.39.219 01:26, 25 January 2007 (UTC)

Everything is sourced. You are the one concealing info. Why haven't you put the Lancet authors' letter here? Or even just the quote you refer to about month-to-month IBC totals as a graph? --Timeshifter 01:43, 25 January 2007 (UTC)

You're cherry-picking from sources and concealing the full IBC data. I have already explained why i did not put the full quotes about this, because it requires uploading a corrected graph from the Guha-Sapir letter which I don't know how to do here.

The above description from you is misleading. You're trying to make the argument that IBC shows that the "mortality rate increased each year", but you're doing this by concealing and truncating data from IBC which doesn't conform to that claim. It's easy to prove a hypothesis when you just selectively exclude data that contradicts it Timeshifter! That's what you're doing with IBC in 2003.

The actual figures in IBC are given as "around":

12,000 - 2003 (9 months - beginning in mid-March 2003)
10,500 - 2004 (12 months)
14,000 - 2005 (12 months)
24,000 - 2006 (12 months)

Thus, mortality did not increase each year in IBC. Further, You are repeating the Lancet authors' error of mixing rates and counts. 12,000 over 9 months gives a *rate* of about 45 per day in 2003. The *rate* for 2004 suggests about 30 per day. The *rate* for 2005 suggests about 40 per day. And the rate for 2006 suggests about 70 per day.

Thus IBC's mortality *rates* would be:

2003 - 45 per day
2004 - 30 per day
2005 - 40 per day
2006 - 70 per day

This of course shows that, according to IBC data, it is not true that "mortality rate increased each year" as you state is a fact in your POV-distorted version that cherry-picks from IBC data to show only the similarities and obscure the differences.

It does seem correct that if you exclude the invasion altogether from IBC then you have a steady rise from beginning to end, but why exclude the invasion from war-related mortality figures? You could do that, as it's another comparison, but to do so by itself is misleading because it obscures some very big differences in the trends. Also, it creates an inconsistent comparison, because Lancet is not only measure rates "not counting the initial invasion deaths". That's just a selective and inconsistent truncation of IBC data to get the hypothesis to still hold when it does not hold.74.73.39.219 01:49, 25 January 2007 (UTC)

74.73.39.219 wrote: "It does seem correct that if you exclude the invasion altogether from IBC then you have a steady rise from beginning to end, but why exclude the invasion from war-related mortality figures?"

Because that is what the Lancet authors did on the original graph. For me to make another comparison would be original research. --Timeshifter 02:01, 25 January 2007 (UTC)

No, you have your facts wrong again Timeshifter. The first point in the IBC line covers the period March 2003-April 2004. That's including the invasion. You can also see that it obviously includes the invasion because the total *count* by April 2004 is listed as being about 15,000. If you exclude the invasion IBC wouldn't have nearly that many by April 2004. It seems like you're determined to confuse yourself, and others.74.73.39.219 02:09, 25 January 2007 (UTC)

BTW..you can replicate the calculations to see what I'm talking about using the rates. We know from the IBC quote that the 2003 total (for 9 months of 2003) is about 12,000. But the time frame in the Lancet graph for that first period goes from March 2003 to April 2004. That means we have to add in four months of 2004. Well, the 2004 rate is about 30 per day, as I showed above. 30 per day over 4 months gives you about 3,000. So 12,000 (March-December 2003) plus 3,000 (Jan to April 2004) gives you the roughly 15,000 that the Lancet authors illustrate for the first period in IBC. If they were excluding the invasion (subtracting about 7,000), that first point for IBC would be below the 10,000 mark. 74.73.39.219 02:13, 25 January 2007 (UTC)

Also, if the Lancet authors *had* removed the invasion from IBC in their graph it would have been misleading again because they don't say anywhere that they were doing that. But they don't say so because they didn't.74.73.39.219 02:17, 25 January 2007 (UTC)

You are missing the point. Both the DoD and IBC graph lines start *after* the invasion. Therefore the slope starts there. It is a tricky math concept to grasp. The increasing steepness of the slope *after* the invasion is what is being measured by the DoD and IBC graphs for cumulative deaths. --Timeshifter 02:18, 25 January 2007 (UTC)

I'm not missing the point. You're misinterpreting the graph. I've proven that IBC has to include the invasion. If it doesn't they have IBC plotted by wrong numbers. If IBC didn't include the invasion it would only have about 8,000 by April 2004. If you can explain how they get IBC to 15,000 by April 2004 while excluding the invasion, go ahead and try. 74.73.39.219 02:27, 25 January 2007 (UTC)

Perhaps it would help if you could explain *why* you believe the original graph excludes the invasion for IBC. That would make it easier for me to explain where you're going wrong. But so far you're just asserting it.74.73.39.219 02:35, 25 January 2007 (UTC)

Just look at the graph. It is plain as day. The graph starts at the platform level that includes the invasion deaths and the deaths through April 2004. --Timeshifter 02:40, 25 January 2007 (UTC)

More gobbledygook. Your interpretation is not "plain as day". It's obviously wrong. The IBC figure has to include the invasion. The only way the period March2003-April2004 gets anywhere near to 15,000 by April 2004 is if the invasion is included.74.73.39.219 02:43, 25 January 2007 (UTC)

Should the erroneous Lancet graph be in the article?

Please vote "support" or "oppose". Also please leave a comment if desired. Please don't forget to sign your vote. --Timeshifter 01:19, 25 January 2007 (UTC)

By the way Seigfried4220 and 74.73.39.219 are the same person as admitted by 74.73.39.219 higher up in another section. So that is only one vote. --Timeshifter 01:51, 25 January 2007 (UTC)

Also, please make your vote bold. I made the vote of 74.73.39.219 bold. To see an example of a poll see the one with the green background on this talk page here: Talk:Iraq War --Timeshifter 01:55, 25 January 2007 (UTC)

Oppose. The graph does not add any useful info to the article. All editors discussing this so far are in agreement that the graph is erroneous now that the anonymous editor has selectively quoted from the subscription-only Lancet letters. From those letters nothing has been quoted claiming deception on the part of the Lancet authors. Just errors. Errors the Lancet authors acknowledged. --Timeshifter 01:06, 25 January 2007 (UTC)
Support - I've made my comments above. The article is about the Lancet study. If the Lancet report contains erroneous stuff this info should be available for many reasons, among those i discussed above.74.73.39.219 01:24, 25 January 2007 (UTC)

given that there are only two people editing this article, a poll is not going to help. the purpose of a poll is to establish consensus when it isn't already known. in this case, everyone's opinions have already been expressed and there clearly isn't any consensus.

imo this entire article is way too long. it would be much better if you two could concentrate on shortening the article, e.g. by deleting most of the block quotes and replacing them by short summaries. Benwing 01:43, 25 January 2007 (UTC)

Well how about deleting the whole section on this erroneous Lancet graph? What purpose does it serve in the article?

Polls are polls. If there were consensus there would be no need for a poll. See the polls on the Iraq War talk page for example. You say everyone's opinion has already been expressed. I am not clear on your opinion. I may have made a mistake. --Timeshifter 01:51, 25 January 2007 (UTC)

I agree Benwing, but the problem is that "summaries" require interpretation, and you know where that's going to go. In my view, factual errors in the report are important. And in my view, Timeshifter wants them concealed because his whole purpose here and elsewhere on wiki has been to plaster the pages with biased edits to POV-push for this Lancet study as the truth above other sources.74.73.39.219 01:57, 25 January 2007 (UTC)

My view is that the section claiming to compare UNDP to Lancet should go altogether, because it's wrong, factoids are presented out of context and misleadingly, and it's just Timeshifter trying to throw mud at UNDP because it contradicts Lancet. It's mostly his own inaccurate editorializing peppered with cherry-picked quotations that create the POV 'narrative' he's trying to get across.74.73.39.219 01:57, 25 January 2007 (UTC)

I have blocked no sourced criticisms of any study. The section on the erroneous Lancet graph is not even a real criticism of the Lancet study methods or results. It is a criticism of a simple error. It is a complete waste of wikipedia space. --Timeshifter 02:02, 25 January 2007 (UTC)

It's actually a series of complicated errors, which all conveniently lead to an illusory "corroboration" for the Lancet study. The errors mislead readers about DoD, IBC and Lancet, and about the trends. People should know this and should be able to make up their own minds about what it amounts to. They should at least have the chance to know that when they read the Lancet report this graph and the claims that accompany it are bogus. You'd prefer people not to know this, but I don't think that should matter.74.73.39.219 02:24, 25 January 2007 (UTC)

After discussion with you I see the lack of math understanding that you have. This graph section in the article will only add to the confusion of readers. From discussing with you (under 4 different IP addresses) in several article talk pages and seeing your edits in those articles I am convinced that is what you want. --Timeshifter 02:36, 25 January 2007 (UTC)

It is the Lancet graph which creates the confusion because it's erroneous and misleading about the trends. The section explains that they got DoD wrong, and that they're mixing counts and rates in the graph, which means the similar slope of the three lines does not indicate any "corroboration". If people can get that from the section, they're much better prepared for understanding this part of the report than if I allow you to conceal all this information from them in order to aid in your POV pushing for this study.74.73.39.219 02:48, 25 January 2007 (UTC)

Seigfried4220 and 4 previous anonymous IP addresses

Seigfried4220 (or 74.73.39.219) is currently on a binge inserting POV, and/or deleting sourced info, on several Iraq War casualty pages.

http://en.wikipedia.org/wiki/Special:Contributions/Seigfried4220

Good lord. Talk about pot, kettle, black. (Sorry for interrupting your important contributions about IP addresses)Seigfried4220 05:34, 25 January 2007 (UTC)

I operate openly. Not behind multiple anonymous IP addresses so that I can come back and create more problems by deleting the same sourced info without being noticed as much. --Timeshifter 06:24, 25 January 2007 (UTC)

You operate deceitfully. You've repeatedly lied about what I've said and now this. I've explained several times now that my IP changes by itself. Yet you're circulating this form letter around to try to imply that I'm doing something underhanded with IP shifting. That's just a deceitful smear campaign.Seigfried4220 07:36, 25 January 2007 (UTC)

And btw..posting under "Timeshifter" is still operating *anonymously*.Seigfried4220 07:38, 25 January 2007 (UTC)

Stop with the personal attacks. I may report you to the appropriate incident board. See my user page for links to the wikipedia policies and boards. You have been asked numerous times to get a user name. Now finally after your IP changes 4 times (that I know of) you finally get a user name. --Timeshifter 08:34, 25 January 2007 (UTC)

---

IP contribution lists with beginning and end dates:

http://en.wikipedia.org/wiki/Special:Contributions/72.68.212.175 - 14:41, 6 November 2006 to 02:36, 10 November 2006
http://en.wikipedia.org/wiki/Special:Contributions/71.246.104.28 - 10:12, 10 November 2006 to 10:47, 16 November 2006

The 71.246.104.28 account linked just above made frequent use of the word "Truthshifter" in comments on this talk page:

http://en.wikipedia.org/wiki/Talk:Iraq_Body_Count_project

The 74.64.60.148 account linked below made this comment below using the word "Truthshifter." Thus helping to tie all 4 IP address accounts to the same person.

"The first paragraph above is relevant to IBC, but is misrepresented by Jamail and now worse by Truthshifter as a 'criticism' of IBC, while it's not at all like those of Jamail and the others."

Here is the revision difference link below showing the addition of the above statement to the talk page for the Iraq Body Count project.