In my previous
blog entry, I did an image analysis on the picture of Obama's birth certificate. While I knew this analysis would get me involved with some conspiracy freaks, I never considered the depth of their devotion to defame Barack Obama.
(NOTE: I'm not voicing my own political opinion. If the same accusations were made against McCain, Clinton, or other people, I would be equally offended.)
- They claimed Obama lied about his name and place of birth, so Obama's team released a copy of his birth certificate.
- They claimed the certificate of live birth (COLB) was fraudulent. A group of analysis experts, including myself, demonstrated that it was not digitally manipulated.
- They said the border was not aligned and looked fake. One commentor on my blog pointed out that it looked realistic for 1960-ish typesetting.
- They claimed that the text was a forgery due to a missing comma in the date (I can't make this up). When faced with the obvious answer (a human typed it in), they changed their attack.
- Now they claim that the seal looks fake. It too, appears to be realistic.
I've seen this type of attack in other forums. Trolls,
griefers, and online malfeasants will do anything to keep the argument going, regardless of any truth. Moreover, anti-Obama conspirators know that the longer they keep the questions going, the more people will hear about the questions and it may sway less-educated voters.
While countering false claims of suspicious image manipulation may make people feel better, it won't stop the false claims. Each time they are proven wrong, the conspirators will come up with something different.
Rather than addressing their charges, let's take a closer look at the people making the accusations...
Susan Who?
The main forum for the conspiracy crowd is by a blogger called "
TexasDarlin". The blogger identity "TexasDarlin" has only been around
since 28-April-2008.
"TexasDarlin" uses the email address "
susanunpc@gmail.com". This identity voices very strong, biased, political opinions. While her blog entries are generally well referenced and well written, they are usually not well argued.
The email address is not unique to TexasDarlin; it is also used by "SusanUnPC" -- a political blogger called
Susan Hu. Her postings are also very strong, political biased, researched, well written, yet poorly argued. Although TexasDarlin is a relatively new identity, SusanUnPC has been actively posting
since at least March 2006.
However, this same email address has been associated with one more name:
Harriet Christian. This one posting, from June 26, 2008 at 9:04 AM, links to a
YouTube video where Harriet makes racial statements, voices very strong and biased political opinions, yet supports
her statements with well-researched references. (Well researched but not well argued.)
Although the email address is used by SusanUnPC, TexasDarlin, and Harriet Christian, it could be used by multiple people. I have no reason to believe that they are by different authors; they could all be by Harriet Christian, or written up by Susan Hu.
The decision to go by the aliases "SusanUnPC" and "TexasDarlin" is not startling. Many people write under a nom de plume. Moreover, the support for the false COLB conspiracy is well within her character; Harriet Christian has written in blog entries, yelled in public, and interviewed with news outlets about her desire to vote against Obama. Although the postings are passionate, her expressed opinions are less than logical with regards to Obama. She has been called a
racist (for calling Obama "an inadequate black male who would not have been running had it not been a white woman who was running for president") and a "
crazy, nut feminist". However, if these are different people using the same email address, then they seem to share political opinions.
Polarik
Unlike TexasDarlin, the primary researcher (and I use that term loosely) goes by the name "polarik".
According to Polarik, and repeated by TexasDarlin, Polarik is an
expert with "20 years of experience in computers, printers, and typewriters". Even though he says he is an expert, he focuses on pixel-based qualitative analysis rather than quantitative analysis and technical evaluation methods. Moreover, he uses no terms common to the field. This reminds me of the Wikipedia
scandal where an author claimed to be a professor of theology, yet made it all up.
The "polarik" entity appears to have been created specificly for this controversy. His TownHall persona uses the name "
Ron Polarik". However, I cannot find that name (or Ronald or "R") in any phonebook. The only "Polarik"s I could find are a Jeffrey Polarik in Wisconsin and Penny Polarik in Michigan. However, I doubt that "Ron" is actually Jeff or Penny. (Please don't harass Jeff or Penny. It's not their fault that Ron chose the same last name.)
In addition, Polarik created a number of accounts right before his "analysis" of the COLB:
- Yahoo. Polarik created his email address on 06/16/2008.
- Town Hall. Polarik did his first posting on June 18, 2008.
- Free Republic (forum). His account was created on Jun 22, 2008. However, the profile has not yet been configured.
When Polarik posted a
comment to my blog, he did it from an IP address in Longwood, Florida.
Online Anonymity
People frequently act differently when they believe they are anonymous. Hiding behind pseudonyms, they will make false statements, argue fake claims, and dream up new conspiracies each time the current one is shot down. As TexasDarlin finally
conceded:
I had personally come to the conclusion recently that Obama’s COLB was probably real, although its handling has been incredibly sloppy.
"Incredibly sloppy" is an understatement. TexasDarlin should have never promoted the conspiracy when her "experts" were unverified. And upon learning of the error, she should have recanted the accusation rather than
continue to promote a baseless conspiracy.
Then again, who ever said that racists were rational.
Forgive me for the off-topic comment. I very much appreciated your careful analysis of the Obama COLB last week. I was wondering if you've had the chance to read about the latest development in the COLB saga (http://atlasshrugs2000.typepad.com/atlas_shrugs/2008/07/atlas-exclusive.html).
Apparently, "Techdude" has produced COLBs from 2006 to 2008, all of which have identical border patterns, and none of which match the 2007 Obama COLB border pattern.
Assuming Techdude's 2006-2008 samples are genuine, that seems like evidence that should at least raise an eyebrow. However, it certainly doesn't take any forensic expertise to compare the border patterns, and hence whether Techdude's further analysis is sound or not seems largely immaterial to me. Nevertheless, I'm curious for your reaction.
Thank you.
I've already replied to this:
http://www.hackerfactor.com/blog/index.php?/archives/202-The-Birth-of-a-Conspiracy.html#c282
Basically, I do not agree with any of the work by Techdude or Polarik. Both have an established track record of provably false claims. In the latest incantation, Techdude mistakes JPEG artifacts for "Some random flaws", interprets heat maps incorrectly, identifies scaling artifacts as a kerning issue, and fails to identify many basic issue such as the evidence handling for each of the images.
Image analysis, even based on observation, is a hard problem.
Image analysis from a scanned in document is even more difficult.
Techdude clearly mistakes scanning and compression artifacts as signs of a forgery.
While various algorithms can identify digital manipulation, no algorithm is perfect. (This is why I use a suite of analysis tools. Although it is easy to make a picture that appears real based on one algorithm, it is difficult to make a picture that appears real based on multiple algorithms.)
Techdude's theory that the document is not authentic is based on one analysis method: observation. Just as an "eye witness" is not as reliable as hard evidence, analysis based on "observation" is not as reliable as other analysis methods.
Because observation can be easily fooled, we use analysis algorithms. If any of the tools detected manipulation, then it would be conclusive. However, none of the tools detected manipulation. This is similar to doing a blood test against a crime scene:
- If the blood at a scene is B+ and the suspect is A-, then you can rule out the suspect. This is conclusive.
- However: If the blood at the scene matches the suspect, then the result is inconclusive.
So, here's something to consider:
- None of the tools detected manipulation. This does not mean that the document is authentic, it only means that the document has not been shown to be a forgery and strongly suggests the absence of any form of digital manipulation.
- The visual differences in the border are not significant enough to say that the COLB is a forgery.
- Occam's Razor: Why would they get the seal, text, font, layout, page position, and green background correct, but not the border? If someone is going to create a forgery of the entire document, would they really miss such an obviously large item and yet correctly create a forgery of the minutia? The differences in the border could be due to the scanning process and/or transformation to JPEG.
It is a very hard problem to create a realistic forgery. None of the sample forgeries identified by Polarik and Xenon pass muster. (Polarik claims they do, but they clearly do not. Xenon intentionally used forgeries as examples for how to detect forgeries.) If it were as easy to create a realistic forgery as Polarik and Techdude claim, then why haven't they done it? In contrast, if it is so hard to create a forgery, then why does Obama's COLB pass every test? Back to Occam's Razor again: Obama's COLB passes every analysis test because it is real. And the various forgeries all fail because they are forgeries.
Finally, as I pointed out in this blog entry, the loudest supporter of Polarik and Techdude's work is SusanUnPC. And she has aligned herself with a very biased position: a dislike of Obama based on his race.
I find it interesting how many people are using this same email. They are use it to bash Obama and the DNC.