MSNBC has emerged as the one major media outlet giving this story consistently respectful treatment. Now David Shuster has posted a good piece on the Green/Libertarian recount effort in Ohio; the article concludes on this intriguing note:
On Friday night, "Hardball with Chris Matthews" (7 p.m. and 11 p.m. ET) will be jumping into the fray. You won't want to miss that. Plus, this weekend here on Hardblogger, I'll be posting some amazing national election numbers and demographic trends that may shock you. I've run some of the numbers past a few MSNBC political analysts, and everybody is calling it "a big deal."That may have some relationship to this:
Science steps in. Buzzflash directs our attention to an important new document by Dr. Steven F. Freeman of the University of Pennsylvania on the exit poll/final tally anomalies that have commanded so much of our attention.
Note the pattern: Debunkers of the "internet-fueled conspiracy theory" so far have not quoted statisticians to support their skeptical reports. Those professionals and academics who have addressed the question within media earshot all agree that something is amiss.
Freeman's paper has not yet received peer review, and his study remains a work in progress. Even so, his work has the impact of a stinger missile.
Freeman demolishes the notion that exit polls were skewed toward Kerry because they oversampled women:
...the issue of male/female ratio is irrelevant. CNN and others released data presenting male and female preferences separately, thus automatically weighting sex appropriately.As for those who claim that the polls were off-kilter because they were taken early in the day:
Regarding time of day variation, this paper does not refer to mid-day reports, but rather end of day data, which happened to be still available at midnight. But even if there were an early voter bias, is there any reason to believe that early votes would be skewed Democratic?Freeman also reminds us that an exit poll/final tally disparity in the former Soviet republic of Georgia led to the storming of that country's parliament and the forced resignation of president Shevardnadze.
According to Freeman, the odds against the statistical anomalies in just three states -- Florida, Ohio, and Pennsylvania -- occurring within the same election are 250 million to one.
I expect this paper to become a battlefield. Don't be surprised if the good professor becomes the target of a vicious right-wing hate campaign.
3 comments:
Jesus, what an article! Why do the conservatives hate the NYT so much? Murdoch could have published this thing. It is a piece of pure invective directed against blogs which at no point addresses the major issues raised here. The date of Professor Freeman's study, which we must now consider the key piece of evidence, is Nov. 10; the date of this Times piece, which does not even mention this key evidence, is Nov. 12. There is no excuse.
Freeman's article is pretty sound, I read it carefully. In addition I should note that stories ran just prior to election that Bush campaign's internal polling showed that his support had "collapsed", matching those polls of voters which showed a decline in predicted outcome for Bush.
It was not only the election day polls that showed a win for Kerry, but a combination of polls done by a number of companies (of voters, not "operatives") showed a sudden surge for Kerry not Bush.
It's often hard to detect election fraud, even in your local county, even with supposedly safer paper ballots counted by Optiscan machines, but occasionally the effort is clumsy, and gets caught, or corrected.
It happens, and I'm sure that going back far enough in time, both parties have been guilty in some counties. It's a constant struggle which "improved technology" can't fix, because it makes the entire process invisible.
Ed Gillespie of the GOP wants to get rid of exit polls altogether, lord help us. During this last election, the GOP candidate for governor was reported on Nov. 2 by one NC county to have received approx approx 5,000 extra votes that were removed/corrected by Tuesday's canvass. I checked on this because it was not quite probable that the governors race attracted more votes than the presidential one in this small county (Johnston County, NC) to this extent.
In Nov. 2 count there were approx 53,000 votes for president, 52,000 votes for US Senate, and 59,000 votes for governor.
That candidate did particularly poorly and the gubernatorial race was declared (correctly) by the local TV news programs within 1 minute of the closing of polls. This was based on exit polls obviously, as they stated.
But what if he had won? In a much closer race for US Senate in NC (before one could access other counties canvassed results online, 10 days after the election) at least 700 more votes than voters in the Senate race votes were reported the morning after the election after an overnight hand count at one precinct with "problems" with either the voting pens or the Optiscan type machine there.
The race involved Jesse Helms and an African American former mayor of Charlotte NC, named Harvey Gantt. This is the closest Jesse came to being defeated and the race was too close to call in polls going into the election, which featured all the kind of GOP voter suppression dirty tricks as Nov. 2004.The initial results were close but just beyond the automatic recount margin I believe.
Believe you me, neither the elections director nor the GOP members of the board wanted to hear from me (even with 2 reporters in tow. They insisted in mailing off the canvass with incorrect vote totals first, and when we persisted and demanded to see the voter book for the precinct, they collapsed and huddled and finally claimed there was an addition error of 1,000.
They said they would send in a corrected return, and we believed them and that is what was reported, but I found out later that actually it doesn't seem they did. I should add that the county party typically pushed for recounts and studied figures only for local county races and so I have to wonder what really happened that year.
Anyhow, that has made me both alert and sceptical about all election results, especially since ever since Reagan so many precinct officials who are registered Democrats in the South have been voting GOP for years wherever races were competitive (and GOP candidate not just some boob up for show in a really hopeless cause, such as safe Dem. seat.)
I tried to sign into this site last week (still under cyber attack?) my browser went wonky. This time I tried and couldn't, so anon I'm afraid.
p.s. to clarify anonymous post re Jesse Helms Race, a propos of exit polls, etc. (which in a fair election would not have shown Helms behind the actual vote result)....
The year of that race was 1990, so not presidential race year. The turnout was incredible in my county, 85%! My county is the one that had the 700 (or 1000) extra votes, Chatham County.
This year we seem to have had a clean election, we have had a different elections director for several years.
Our state board of elections doesn't seem to list actual number of voters per county, though counties record this by precinct, and this is supposed to be verified by poll observers when polls close, and then compared to reported number of voters per precinct, so that there's a check on "extra votes".
This is how LBJ got caught, though in fact in one precinct the voter book had been altered too, but the last voters to vote remembered who they were. Those were the days, though I should note, LBJ kept his seat I think.
This was all covered in a New Yorker article about stealing elections that ran sometime in the 1980s, that's as close as I can come. It also covered the possibility of gaming lever machines.
The New Yorker also ran an article in May 2001 called "the Doomsday click" about internet security (there really simply cannot be any such thing) and I think any GOP readers of this site should take note that we should be concerned about anyone sabotaging election results, no matter who.
Post a Comment