2008 Texas Primary

Tall tales coming out of Texas, and some citizens riding to the rescue.

Big Julie's Blank Dice and the Texas Two-Step

Big Julie’s Blank Dice and the Texas Two-Step:
Thoughts on March 4th and Computerized Elections

by Jonathan Simon, Election Defense Alliance

Last week, as I was watching what could be watched of the crucial March 4th Democratic primary elections and downloading for analysis such data as was made publicly available, a hilarious scene kept coming unbidden to mind. The scene is from Guys and Dolls and it takes place somewhere in the sewer system of New York, where Nathan Detroit’s floating crap game has found a temporary and rather sarcastically colorful and well-lit home.

Big Julie, a scar-faced high-roller in from Chicago to “shoot crap,” is down on his luck and out about 10 Gs. Nathan (Frank Sinatra) says it’s time to go home, but Big Julie is not the kind of mug to go gentle into that good night without his 10 Gs, plus interest. So he challenges Nathan to roll him personally for the dough and Nathan (putting up cash to Big Julie’s “marker” and, to narrow down his choices somewhat, at gunpoint) accepts.

Big Julie, to change his luck, is going to use his own dice. The trouble (for Nathan at least) is that the dice don’t have any spots; they’ve worn off. But, not to worry, Big Julie remembers where they were.

The results (“Hah! Seven! I win . . . . Hah! Snake Eyes! You lose”) are, shall we say, predictable—though Nathan does manage to win when Big Julie rolls him for $1—and Nathan kisses off his last few grand with a resignation worthy of Gore, Kerry, a host of other candidates who would not appear on Karl Rove’s A-list, and the Democratic Party as a whole.

The scene is hilarious, but Tuesday night was not. Nor was New Hampshire, nor 2006, nor 2004, nor 2002, nor any election in America since the vote counting went wholesale into the darkness of proprietary cyberspace and the spots were rubbed off the dice, leaving the equipment vendors, with their avowed partisan proclivities and their secret computer code and memory cards, to tell us who won (“Hah! Seven!”) and who lost (“Hah! Snake Eyes!”).

Before proceeding to analyze yet another evening of bizarre numerical happenings, I want to suggest that we look at some of the occurrences in our New Millennium elections as if they hailed not from our own beloved Beacon of Democracy but from Vladimir Putin’s Russia—from a place, that is, where we have learned to discount the official story as the typical cover job of a pretend democracy. We will find—as we might in Russia, or Kenya, or Ukraine—a parade of numbers and patterns that don’t add up, don’t fit the official story. And all we have to assure us that our democracy and our nation are not being subverted are code and memory cards we are never permitted to see, providing us with very shiny and precise-looking vote totals that may or may not have any correspondence to the votes actually cast—in other words, a pair of Big Julie’s blank dice.

The Democratic primaries held Tuesday shared with the New Hampshire primary the distinction of being do-or-die contests for candidate Clinton. It was effectively conceded by the Clinton campaign that losses back in New Hampshire in January and in Texas or Ohio last week would have spelled curtains for her candidacy. There were four primaries held on March 4th and, leaving aside the delegate-poor and noncompetitive contest in Vermont, there were New Hampshiresque—which is to say somewhere between suspicious and stunning—developments in each of them. Taking them in alphabetical order:

Ohio

The numerical story in Ohio was the old familiar one of exit poll-vote count disparity. Without examining any specific incidents or allegations of foul play, we are confronted with an initial exit poll (EP), posted shortly after poll closing, showing a 3% Clinton margin (51.1% to 47.9%) and a final votecount (VC) showing a 10% Clinton margin (54.3% to 44.0%). This disparity is outside the EP’s margin of error (MOE) even allowing for the “cluster effect.” The VC is, moreover, a significant departure from a compendium of pre-election polls (PEP), showing Obama gaining ground and approaching equality (http://www.usatoday.com/news/politics/election2008/poll-tracker.htm [8]).

Viewed in isolation, Ohio could be explained as a late Clinton surge that caught the pre-election pollsters on the hop. Primaries are indeed more fluid and volatile as elections go, and there is the crossover voting phenomenon to be considered. But Ohio takes its place among a parade of contests in important states in the 2008 nomination battle in which a substantial EP-VC disparity worked in Clinton’s favor: New Hampshire obviously, but also Massachusetts, Illinois, New Jersey, Arkansas, Arizona, California, and now Ohio and, as we will see, Texas and Rhode Island. In contrast, we have observed no battleground states with an EP-VC disparity working in the other direction. As anyone who has spent any time in the countryside of Ohio (or NH, MA, IL, NJ, AR, AZ, CA, TX or RI for that matter) can tell you, when all the cows are facing in one direction, there’s a reason for it (it’s going to rain).

Given the directionality of the disparities, it is also worth noting that we have received no assurance that the first posted EP of the evening has not in fact already been partially adjusted toward conformity with the incoming VC, a process which continues in several steps throughout the evening until virtual full conformity with the final VC is achieved. Edison/Mitofsky, which performs and processes the EPs for the media consortium known as the National Election Pool (NEP), acknowledges that the adjustment process begins with “Quick Counts,” which are available from selected precincts and early voting tabulations immediately upon poll closing. Especially in instances where the first EP posting is delayed by more than a few minutes after poll closing, there exists ample opportunity to begin the process of adjustment, which of course has the effect of minimizing the observable EP-VC disparity.

Rhode Island

There’s not a whole lot to say about Rhode Island other than if exit polls are this far off, why bother exit polling? And if vote counts are this far off, why bother voting?

The EP-VC disparity in RI was 14.1%; that is, the exit poll posted after poll closing had Clinton up 4.1% (51.6% to 47.5%) over Obama, and the official vote count had Clinton up 18.2% (58.8% to 40.6%). This is far outside the most generous calculation of the EP's MOE, and on a par with the similarly perplexing 15.5% disparity favoring Clinton in Massachusetts on Super Tuesday.

Now since EP-VC disparities of suspect, if not outright stunning, magnitude have become commonplace in the era of computerized vote tabulation, it is clear enough that something is not happening according to Hoyle. What that something is has been settled on by the mainstream media and all analysts under contract to such: Since we dare not question the vote counts, the exit polls must be off again. . .and again . . .and again. In fact it is now established that the exit polls are always off (recently joined by the pre-election polls, especially in the wake of New Hampshire 2008) and no longer worthy of our attention, because they just keep on disagreeing with the vote counts—pretty much always in the same direction—and we dare not question the vote counts. . . .

And so the circular argument goes, by now repeated with enough reassuring smiles to take on the polished finish of fact. Except if they really cared to find the truth, the logic of the denialists would puncture, like an overinflated balloon, with one prick of a pin, and all would agree that without first investigating and verifying the vote counts—without, that is, putting the spots back on the dice—no one can conclude that all those polls are “off.”

Texas

The striking phenomenon in Texas was the magnitude not of the EP-VC disparity (it was a relatively modest 4%, in the usual direction, but withheld from the public until more than an hour after poll closing, allowing ample opportunity for extensive adjustment toward conformity with the in-coming vote count) but of the early voting (EV) vs. at-precinct voting (APV) disparity, which was of staggering proportions that at first seemed to defy explanation.

The earliest returns posted on network websites showed a total of approximately 740,000 votes cast in the Democratic primary with 0% of precincts reporting. This then was the early/absentee vote tally, which in most states is pre-counted and available for release immediately upon poll closing. Obama’s margin at that point was 436,034 to 303,276 for Clinton, or 59% to 41%. By the time the counting was done the next morning, Clinton had a 51% to 48% victory, a whopping 21% margin reversal.

What was even more stunning, however, was that Clinton had caught up to Obama before even a quarter of the precincts had reported: with 23% of the precincts reporting (and almost exactly as many APVs as EVs counted), the count stood at Obama 711,759, Clinton 711,183 (49%-49%), a dead heat. To catch up so quickly and produce those numbers, Clinton had to win the APV in that quarter of Texas precincts by 59% to 41%, an exact reversal of the EV Obama landslide. Judging by the county-level results posted, that APV Clinton landslide came predominantly from the rural areas of the state.

So what we saw until that point were essentially equal and opposite landslides, as if we were observing two not only separate but radically divergent electorates, one that chose to vote early and one that chose to go to the polls. Ordinarily explanations for a divergence of such magnitude, particularly in intra-party contests, would be found only in such time-specific phenomena as late-breaking gaffes, scandals, debate blowouts and the like. But there was no such occurrence. The early voting period inTexas extends from 17 days to 4 days prior to the election. During this period the average of 13 pre-election polls was Clinton 45.6%, Obama 46.7%. In the three days before the election, after the early voting period had ended, the average of eight polls was Clinton 46.8%, Obama 46.1%, a very modest change and certainly not the 21% mega-reversal displayed by the EV and APV vote counts.

Since ordinary political dynamics fail to explain the bizarre Texas numbers, we look to the extraordinary. There has been much made in the March 4th post-mortem period of the impact of crossover voting, specifically Republican voters exhorted by Rush Limbaugh and other lesser-known leaders, to hold their noses and vote in the Democratic primary for Clinton.

To digress just a bit from our analysis of the March 4 numbers, the Limbaugh appeal brings into the open the motive and strategy that go a long way to explaining virtually all of the bizarre disparities and anomalies that have beset at least the Democratic side of the 2008 primary season. It has for some time been quite apparent that the goal of Republican strategists, finally exposed in Limbaugh’s rather desperate public exhortation, has been to make sure the Democratic nomination process is as drawn out, bitter, procedural, and ugly as possible, culminating in a brutal battle involving superdelegates and credentialing, one that will turn off (to say the least) the public and leave festering wounds in the party itself.

If the goal had been simply to have Clinton win, that could have been easily achieved through cross-over voting and/or rigging--remember that Obama won something like a dozen contests in a row, most of which could have been pushed far enough in Clinton's favor to give her a decisive delegate edge. This wasn't done. What was done instead was to revive Clinton's campaign (it appears by rigging) when she was on life-support in NH, keep her within striking distance on Super Tuesday, let Obama gain popularity and momentum, then revive Clinton again on March 4, just when the Democrats nationally were getting comfortable with Obama as their candidate (again see http://www.usatoday.com/news/politics/election2008/poll-tracker.htm [8]).

Now it will get really ugly and whoever emerges as the nominee will have been undermined enough--so the story will go, anyway--to manage to 'lose' to McCain; i.e., either Clinton or Obama will have accumulated plenty of plausible defeatability. And the story of Democratic 'civil war' (as the MSM is already gleefully framing it) and disarray may even be good enough to 'explain' how they failed to capitalize on the enormous structural and dynamic advantages they hold on the Congressional side, setting the stage for currently unimaginable Republican gains in Congress in November.

Perhaps crossover voting accounts for both the magnitude of the Clinton victory in Ohio and the miraculous reversal of Obama’s early voting margin in Texas. Or perhaps it was crossover voting and some computer voting (that is, voting by computers) for good measure. Is there any way to know? It does not take all that much imagination to see Clinton’s successive resuscitations as Karl Rove's specialty of the house, his apotheosis as conscienceless strategist: to go into 'retirement' and apparent seclusion, give scholarly and apparently appealing speeches from above it all on the lecture circuit, and meanwhile find the exact alchemic strategy to turn a pile of rusty Republican political scrap metal into gold. But strategy is one thing and rigging something else entirely. Or is it?

Of course a “fit” is not tantamount to proof. But when you have a multi-year parade of numerical anomalies combined with unexpected outcomes that a brilliant and apparently conscienceless strategist would bring about if he could, must it not at least shake the blind faith that Americans, cued by their opinion leaders, continue to place in the honesty of their black-box electoral system?

Isn’t it time to stop rolling for our nation’s future with blank dice?

Attachment	Size
BlankDiceElections.pdf [9]	119.65 KB

Little Stat Helper: A Guide to Sampling Statistics and Election Assessment

Little Stat Helper

By Jonathan Simon, Election Defense Alliance

The benefit of statistical sampling lies in the surprisingly strong power of a small part to predict the behavior of a large whole. Although we tend to accept the results of polls and other research based on sampling, most people if they really thought about it would find it quite a head-scratcher that you could predict with great accuracy the preferences of a nation of 300 million--whether in candidates, policies, or favorite kind of cheese--by questioning a mere 3000, or .001%, of them. This is nevertheless the case, providing that certain conditions apply and certain procedures are followed.

To get highly reliable results it is important that the sample be done as randomly as possible. If bias or convenience enter into the sampling process all bets are off and statistical process loses its "crispness," the cut-and-dried rule of simple equations. More on that in a moment, but the key stumbling block I have found for acceptance of statistical sampling is the mind's natural and intuitive protest that larger wholes require larger samples. Once we reach a whole of a certain size--i.e., the size we are dealing with in federal elections--though, larger wholes in fact do not require larger samples, however counterintuitive this may seem. Not only has this been established theoretically, it has also been demonstrated in thousands of experiments. It's just the way it is. High-school statistics, chapter 1.

Now, bearing this basic concept in mind, here are the terms that get thrown around, and that must be understood to have a rational discussion of any statistical protocols and proposals:

Margin of error (MOE) of a sample refers to the range in which we expect to find the discrepancy between the count of the sample and the count of the whole from which the sample is drawn. In most research a +/- x% MOE means that 95% of the time, or 19 out of 20 times, we expect the count of the sample to fall within x% of the count of the whole; that is, the "confidence level" for that MOE is 95%, the standard used in most scientific research. 5% of the time a random sample with an x% MOE will count up more than x% off from the whole.

That's just the way it is. You don't get 100% certainty. But what is certain is that if you ran that sample a billion times, the number of times it missed the whole by more than x% would approach 50 million (5%) very closely. That's why computer simulations are so helpful, because you can actually do this and check the results.

A confidence level of 99% (or better), which I've recommended for votecount checking, would tell you to expect a result within the MOE 99 out of 100 times (or better).

The size of the whole numerically has virtually zero impact on the size of the sample needed (once you get above a whole of, say, 50,000; although there is a simple formula, irrelevant here, used for adjusting for such small targets).

As I've said, probably the most difficult and counter-intuitive thing to swallow about sampling is that you don't need to increase your sample size when the size of the target whole jumps from, say, 1,000,000 to 250,000,000; your 30,000 ballots would work about as well as a sample of the whole country as of Rhode Island. Hard to accept but it's true and very elementary statistics.

Given a random sample of a large (numerous) whole, the MOE and Confidence Level as defined above can be calculated quite easily. For a competitive election (60%-40% or closer), the magic formula boils down to: MOE at 95% Confidence = 1/square root of the number of ballots sampled (generally referred to as "N").

So, to plug in a few numbers: If you sample 10,000 ballots, then 1/sqrtN = 1/100 = 1%, and you'd say that your MOE is +/-1%. You would expect the sample results to differ from the total tabulated results by less than 1% in 19 out of 20 such elections. If you looked at, say, 1000 such elections, you'd find that the sample/whole difference was less than 1% in just about 950 of them. The more elections you ran, the more exact that 19 out of 20 would become.
If you sampled only 400 ballots, then 1/sqrtN = 1/20 = 5%, and you'd say the MOE is +/-5%, and you'd expect the sample results to differ from the total tabulated results by less than 5% in 19 out of 20 such elections.
If you sampled 30,000 ballots then 1/sqrtN = 1/173.2 = 0.58%, and you'd say your MOE is +/- 0.58%, and you'd expect the sample results to differ from the total tabulated results by less than 0.58% in 19 out of 20 elections.

Now all those examples presumed a Confidence Level of 95%, the standard for most research. But what that would mean is that if the MOE were used as a trigger for full hand counts or any other relatively drastic check of the results, elections officials would be obliged to proceed to such a step once in every 20 elections or races in the absence of mistabulation, in essence because we set the trigger at an MOE that's only expected to "work" 19 out of 20 times. The Confidence Level that is standard for most research would probably be seen as inadequate for checking on elections.

Fortunately, given a sample size N, a MOE can be easily calculated for any Confidence Level. To find the MOE at a 99% Confidence Level, for example, just take the MOE numbers above and multiply by 1.29: 10,000 ballots would give you a MOE of +/-1.29%; 400 ballots would give you a MOE of +/-6.45%; 30,000 ballots would give you a MOE of +/-0.75%; all at 99% confidence. This would mean that in only one out 100 elections would the difference between sample and whole exceed the MOE in the absence of mistabulation. We believe that one such "false positive" per 100 races would be tolerable to most BOEs (especially since the sample can be run again--that is, resampled--after such a result, rather than proceeding directly to a full hand count).

By the way, the "magic number" of randomly sampled ballots needed for a +/-1% MOE at a 99% Confidence Level is about 16,500, as can be checked on a nice website -- http://www.raosoft.com/samplesize.html [10] -- very helpful in such calculations.

And, to illustrate an earlier point about the irrelevance of the size of the whole: For a state with 5,000,000 votes, you'd need 16,533 ballots, for a state with 10,000,000 votes, you'd need 16,560, and for a country with 100,000,000 votes, you'd need 16,585. To boost the confidence level to 99.9%, so that you could tell a BOE that they'd have to deal with a "false alarm" only once in 1000 elections, the magic number would be 27,000 ballots.

For a given venue (be it state or Congressional District) of known or predictable size (i.e., number of votes expected to be cast) coming up with sample sizes is child's play, just a question of plugging in a few numbers on a calculating website such as that given above. What's left to tackle is randomness.

There are several factors that can get in the way of randomness in sampling, but in the context of elections they all boil down to convenience or bias. And, given the proper protocol, they all can be avoided. Bias generally crops up when interviews are necessary, as with polling and exit polls. Interviewers may select respondents they "like" rather than say every 7th person to walk through the door; they may frame questions in a leading manner; they may hear what they want to hear in the response and mark it accordingly. Respondents, for their part, may be more likely to participate with an interviewer they like, or may give the interviewer answers the respondent thinks the interviewer wants to hear.

All of these possibilities create higher potential for error and are very difficult to quantify. Convenience can take the form of trying to capture your respondents in "bunches," such as at a few precincts (no exit pollster, for example, has the resources to send interviewers to every precinct, so they pick a few precincts carefully based on their likelihood to reflecting the whole), or at a certain time of day, or from the top of a big stack of ballots. Here too error is increased, in a way that is very difficult to quantify, turning statistics from crisp to soggy, straight science to a science-art hybrid.

A well-designed and administered hand-count sampling of ballots avoids all of these pitfalls, and is indeed "crisp" (and in this way very different from exit polling and targeted audits). Since we're counting ballots rather than interviewing, the bias pitfalls are eliminated. Since we propose counting a fixed proportion of the ballots at every precinct (rather than counting all or some of the ballots at selected precincts, as some have proposed), we avoid the principal convenience pitfall of a "clustered" sample.

All that remains is to insure that the ballots to be sampled at each precinct are, in effect, "shuffled" and a good random sample drawn. This can be achieved by literally shuffling the ballots after they are retrieved from their bin and then selecting ballots from the pile according to a predetermined choosing scale: say, every 15th ballot or every 87th ballot, depending on the overall number needed from the venue. With a modicum of observation and supervision, a random sample can be guaranteed.

Questions will surely arise as to cheating, attempting to rig the handcount sample as well as the machine count. The best answer is in the purpose of the handcount protocol and what cheating would in fact achieve. Since, unlike that of the machine count, the purpose of the handcount is not to get as many votes as possible for your guy, but to match the machine count within the MOE and thereby avoid a full handcount, the incentive for "stuffing" the handcount with extra ballots for your guy wherever possible vanishes.

In fact election officials' goal becomes to do the handcount sample as accurately as possible in order to avoid triggering a full hand count. Even granting that a given official or group of officials knew that the machine count had been rigged to add an extra 5% to "their guy," consider how difficult it would be for them to add the necessary number of handcount ballots to hit that rigged number within the MOE, given that it would have to be done in dozens if not hundreds of precincts in view of both partisan and neutral observers. Effectively impossible.

Now let's turn to a few more concrete numbers. In a Congressional District (CD), a competitive race draws between 200,000 and 250,000 voters. Picking the lower bound, we can achieve a +/-1% MOE at 99% Confidence Level by sampling about 15,000 ballots, or 7.5% of the total cast. Given an average precinct size of 500, that would work out to an average of just about 40 ballots to count at each precinct. Not very labor intensive.

Looking at a medium sized state such as Ohio, with 5,000,000 voters and 11,000 precincts, we'd need about 16,500 ballots, or 0.33% of the total cast. This works out to an average of just 1.5 ballots per precinct. This is so easy that it leads naturally to the idea of a bigger sample, so that the Confidence Level can be improved even further. And indeed it turns out that you can reach a 99.99% Confidence Level (that is, one false alarm in every 10,000 elections!) with a +/-1% MOE by sampling 37,500 ballots or an average of less than four (4) per precinct. Woo hoo!

In the state of California a Confidence Level of 99.99% requires those same 37,500 ballots, which boils down to an average of less than two ballots per precinct. In such a large venue, therefore, we can do even better: we could go to a MOE of +/- 0.5% at 99.99% confidence by taking 150,000 ballots, or about eight ballots per precinct. Such a MOE would sound the alarm on outcome altering mistabulation of any race decided by greater than one half of one percent. Woo hoo times two!

In conclusion, we propose a uniform, omni-precinct, proportional handcount sampling of ballots -- the Universal Ballot Sample [11] method* -- be used as the most reliable check mechanism of machine counts, where full hand counting of paper ballots is not yet an acceptable alternative. Such a protocol obeys "crisp" laws of statistics and is highly reliable, with little or no incentive for gaming or practical way to do so.

It can be implemented where paper ballots are in use, whether with opscan systems or, somewhat more problematically, where DREs are fitted with a paper ballot printer. The labor involved at the precinct level is reasonable and within the capacities of virtually every local BOE. The consultant work to generate the parameters is also minimal.

The uniform, omni-precinct, proportional handcount sampling of ballots is a viable and practical protocol that can be rapidly implemented to serve as a check mechanism on computerized recording and tabulation of votes where full hand counts have not been adopted.

Download link for the UBS paper: http://www.electiondefensealliance.org/files/New_UBS_811Update_061707.pdf [11]

Attachment	Size
Little Stat Helper.pdf [12]	221.5 KB

Managing Electoral Dynamics Via Covert Vote-Count Manipulation

Published by EDA, May 21, 2008. Click to download a PDF copy of this article. [13]

---

 

The Democratic Primaries 2008:

Managing Electoral Dynamics Via Covert Vote-Count Manipulation

By Jonathan Simon and Bruce O’Dell, Election Defense Alliance

Summary Statement

We present evidence supporting the hypothesis that systematic attempts are being made to manipulate the results of the 2008 Democratic presidential nomination contest, through overt means such as crossover voting by non-Democrats, and through covert means targeted at the electronic vote tabulation process itself. The net effect has been to prolong the nomination battle and sharpen its negativity, thereby boosting the prospects of the Republican nominee and making more plausible his “victory” in November—either by an honest count, or through continued exploitation of the proven security vulnerabilities in American voting systems.

Introduction

Perhaps John McCain is, as Humphrey Bogart says of the young Bulgarian who wins the money for his family’s exit visa at the roulette table in Casablanca, “just a lucky guy.”

Lucky that the Democrats find themselves locked into a protracted primary season inexorable in its dynamics and increasingly destructive in its impact. Lucky that Hillary Clinton has been magically revived each time she has found herself on electoral life support, to assume a position just far enough behind Barack Obama to be induced to resort to desperate measures and increasingly-negative ads, yet not so far behind as to be forced to bow out.

Lucky that dynamics ostensibly out of McCain's control have combined to give him such material assistance. Perhaps.

But there is compelling evidence that something other than luck is at work. With 82% of Americans polled convinced the nation is on the wrong track, self-destruction by the Democratic party is the only remaining credible means by which, come 2008, the GOP could sustain the perpetual rule envisioned by Karl Rove. (Rove, of course, has hardly retired and is now working from home, beyond the reach of the mandatory email backup system installed at the White House just before he left to “spend time with his family.”)

The goal of Democratic party self-destruction in 2008 could most reliably be brought to pass by one progression of events, one choreography: if a candidate, Hilary Clinton, known for her sense of entitlement, lifelong ambition, tenacity—and willingness to go negative—could be placed and kept in a desperate but not quite hopeless position, the result would follow, quite predictably.

What the mainstream media have now set up and trumpeted as an epic “blood feud” in the Democratic Party, whether or not it actually undermines the party’s prospects in November, will certainly pre-establish a plausible “explanation” for the defeat of whoever the Democratic nominee turns out to be.

The same is true for US Senate and House races, where Democrats are heavily favored to expand their majorities, given the large number of open seats this November that were formerly held by Republican incumbents and a string of recent special election victories. But Democratic congressional candidates in both houses are arguably now facing the prospect of negative coattails.

By setting the stage for post-election “spin” for the Presidential and congressional races in November, any outcome-determinative electoral manipulations would become much less “shocking,” and that much less likely to trigger investigation and ultimate detection. This jaundiced overview of the Democratic primary season¹ is unfortunately supported by a body of evidence that goes well beyond the odd anomaly or two.

When we examine—as the media has steadfastly refused to do—the numbers and disparities discovered in a parade of key states that determined the path the Democratic contest has taken to date, we find a telling pattern. This pattern is consistent with a tactical manipulation of the primary election vote counts in the service of the strategic choreography alluded to above: Keeping Clinton in the race and desperate. __________________________________________________________________________________________________________________________________________ ¹ There were significant oddities on the Republican side as well, beyond the scope of our analysis here. ________________________________________________________________________________________________________________________________________

There are sound reasons why the Clinton campaign itself is not among the suspects: If Clinton’s campaign or supporters had the capacity and the will to alter election outcomes, it is reasonable to conclude that she would have won, or at least be ahead in, the race; and the ownership and operation of electronic voting equipment remains almost exclusively in the secretive hands of vendors (Diebold/Premier, ES&S, Hart, and Sequoia) with avowedly right-wing Republican political sympathies.

Our examination includes the Democratic primaries in the following key states: New Hampshire, Massachusetts, Rhode Island, Texas, Ohio, and Pennsylvania, each of which had surprising and unexpected results. In each of these critical elections there was a significant pro-Clinton disparity when comparing pre-election surveys and Election Day exit polls against the official vote counts.

1/8/08: New Hampshire

This was the first of the “must-wins” for Clinton. She went into New Hampshire on the heels of an embarrassing third-place finish in Iowa and a 20%+ defeat in Wyoming, had lost momentum, and was trailing by substantial margins in every pre-election poll in the Granite state (the range was from 5% to 13%, with both Obama’s and Clinton’s internal polling also showing a double-digit Obama margin).

Observers consistently reported Obama rallies that were far larger and more enthusiastic. There was no sign of a Clinton groundswell. Yet on Election Night the voters apparently changed their minds, and gave her a 3% victory. The media pundits scratched their collective heads and scrambled to explain this stunning reversal, which would have been remarkable enough if it had been a double-digit shift from a single reputable tracking poll, but was truly staggering when viewed against the backdrop of the entire phalanx of tracking polls.

There was palpable grasping at straws—but never even a hint that perhaps the polls had it right and something was wrong with the vote counts. Nor was there a mention that the first posted National Election Pool (NEP) exit poll had Obama ahead 39.4% to 38.1%, while earlier unposted NEP exit polls put Obama further ahead.

The first posted exit poll was already weighted to a carefully calibrated demographic profile of the electorate, and therefore as reliable an indicator of voter intent as is available. Indeed, that first-posted exit poll may already have been partially adjusted toward conformity with the incoming vote counts, thereby understating the apparent exit poll-vote count disparity.

That exit poll was largely spot-on for the other candidates; only Clinton and Obama's exit poll numbers shifted significantly as votes were tabulated. The mainstream media also did not mention the extraordinary disparity between votes that were counted by hand (Obama + 6.5% head-to-head with Clinton) and those tabulated by computerized optical scan devices (Clinton + 5.5% head-to-head with Obama).

Although the counting method (machine vs. hand) was not strictly homogeneously distributed throughout the state, neither was it clustered in such a way that would readily explain the huge statistical disparity in results. When considering benign reasons for such surprising and unexpected outcomes, conventional explanations all begin and end with the unquestioned belief that the computerized vote counts are valid. Quite an assumption in light of the parade of anomalies, disparities, and machine failures witnessed nationwide since the advent and proliferation of computerized vote counting.

Official election results are assumed valid, even though the votes are tabulated by secret software² concealed on memory cards immune to inspection and under the strict proprietary control of an outsourced corporate vendor; in New Hampshire, the vendor is LHS, about which unanswered questions abound. __________________________________________________________________________________________________________________________________________ ² Remarkably enough, we know with certainty that the precise model of optical scan voting equipment in use in New Hampshire, Diebold Accuvote OS Model 1.94W, is vulnerable to outcome-altering manipulation by insiders. A live demonstration on that very Diebold model was captured in the HBO documentary "Hacking Democracy". _________________________________________________________________________________________________________________________________________
In an on-going epilogue, the New Hampshire primary remains under scrutiny. Investigators are amassing detailed evidence of pervasive mistabulation, focused in certain counties. On the Democratic side, there were an alarming number of polling sites reporting more votes than voters. Recounting was rendered effectively useless by a nonexistent chain of custody, which permitted more than ample opportunity for ballot substitution and revision.

Memory cards were reported as having been erased and were never made available to investigators. Even something as basic as a reconciliation of the number of ballots delivered to number of ballots voted, spoiled, and uncast was lacking. Nor was there reconciliation of number of voters checked in at the polls to number of ballots cast.

At this first critical turning point in the Democratic contest an Obama victory would have, in the view of most analysts, effectively ended Clinton’s campaign. That victory—augured in pre-election polling, exit polls, and hand-counted ballots—vanished into the black box scanners provided by Diebold and programmed by LHS. Instead, Clinton was credited with a stunning comeback, given new life, and the nomination battle continued.

2/5/2008: Super Tuesday

Super Tuesday was essentially a standoff, each candidate doing what was necessary to remain viable. There were, however, several exit poll-vote count disparities far beyond the expected margin of error, each involving a shift toward Clinton.

In Massachusetts, another LHS state like New Hampshire, the shift was a whopping 15.5%, turning a projected narrow Obama victory into a 15% Clinton rout.

In Arizona, site of some of the most dubious counting antics over the past several election cycles, the pro-Clinton shift was 11%, again reversing the outcome.

And in New Jersey, where machines are currently under high scrutiny supported by a court order, the shift was 8.6%.

Each of these shifts was well beyond the margin of error of the respective polls. Each resulted in shifts in delegate count to Obama’s detriment, as well as the loss of two victories that would have put a very different complexion on the outcome of Super Tuesday as a whole. The overall effect was, again, to maintain Clinton’s viability.

3/4/2008: Ohio, Rhode Island, Texas

In the weeks following Super Tuesday, Obama racked up a succession of impressive wins—including every caucus state, where vote counting is often face-to-face, and subject to greater scrutiny. As a result he pulled well ahead in the delegate count, and began to take on the mantle of inevitability.

Once more, pundits were calling the race all but over, and Texas and Ohio were often described as Clinton’s last stand. She needed wins in both states, it was flatly stated, to continue in the race. Even Clinton’s own campaign conceded as much. In the weeks before the election, Obama had closed an initial gap in both states and was running even or ahead in pre-election polling.

Ohio

In Ohio once again we are confronted with a discrepancy between exit polls and official tallies. The initial published exit poll, posted shortly after poll closing, showed a 3% Clinton margin (51.1% to 47.9%), while the final official vote count showed a 10% Clinton margin (54.3% to 44.0%). This disparity was well outside the exit poll’s margin of error. The official vote count was also a significant departure from a compendium of pre-election polls, which showed Obama gaining ground and approaching equality.³

Viewed in isolation, Ohio could be explained as a “late Clinton surge” that caught the pre-election pollsters by surprise. Primaries are indeed fluid and volatile, as elections go, and there were reports of organized attempts to encourage Republican crossover voting for Clinton, though the Republican crossover vote may have been less robust than initially reported.

It can also plausibly be viewed as another in a succession of “cover stories” (for example, the massive but phantom after-dinner Evangelical turnout offered up by Rove as a factor in reversing the outcome in 2004) that could well provide a relatively benign explanation for more nefarious operations. But instead there was a parade of contests in important states in the 2008 nomination battle in which a substantial exit poll-vote count disparity worked in Clinton’s favor—including New Hampshire, Massachusetts, Illinois, New Jersey, Arkansas, Arizona, California, and now Ohio and, as we will see, Texas and Rhode Island.

In contrast, we have observed to date no battleground state primary with a significant⁴ exit poll-vote count disparity in Obama's favor. Some have invoked the so-called “Bradley effect” to account for this string of disparities. According to this theory, some white voters who would not vote for a black candidate in the privacy of the voting booth are “shamed” into indicating to pollsters (i.e., in public) that they chose that candidate.

But research into the Bradley effect has established that it is, at best, an inconsistent and relatively rare phenomenon, very unlikely to account for such a pervasive pattern as identified above. It is only if one is unwilling to consider any possibility of computerized vote mistabulation that such superficially plausible theories as the Bradley effect take their place in the front of the line of explanations. __________________________________________________________________________________________________________________________________________ ³ See http://www.usatoday.com/news/politics/election2008/poll-tracker.htm [8] ⁴ In this case, significant means "larger than the exit poll margin of error." _________________________________________________________________________________________________________________________________________

Rhode Island

The exit poll-vote count disparity in Rhode Island was 14.1%; the exit poll posted after poll closing had Clinton up 4.1% (51.6% to 47.5%) over Obama, yet the official vote count had Clinton up 18.2% (58.8% to 40.6%). This is far outside the exit poll’s margin of error, and on a par with the similarly perplexing and bizarre 15.5% disparity favoring Clinton in Massachusetts on Super Tuesday. It is reasonable to ask, if exit polls are this far off, why bother exit polling? (Or perhaps just as reasonable to ask, if vote counts are this far off, why bother voting?)

Texas

In Texas there was a relatively modest 4% discrepancy between the first posted exit polls and official tallies—in the usual direction and larger than the margin of error, and also in this case, withheld from the public until more than an hour after poll closing.

While most primary exit polls are posted a few minutes after the polls close, an hour's delay enables ample opportunity for adjustment of exit polls toward conformity with the incoming vote count, and so the posted exit polls may understate the magnitude of the discrepancy.

But the disparity in Texas between early voting results vs. Election Day in-precinct voting was of staggering proportions that seemed to defy explanation. The earliest returns posted on network websites showed a total of approximately 740,000 votes cast in the Democratic primary with 0% of precincts reporting. This was the early/absentee vote tally, which in some states is tabulated and available for release immediately upon poll closing. Obama’s vote at that point was 436,034 to 303,276 for Clinton, or 59% to 41%, an 18% margin.

But by the time the counting was done the next morning, Clinton had a 51% to 48% victory . . . a whopping 21% margin reversal. What was even more stunning, however, was that Clinton had caught up to Obama before even a quarter of the election day vote had been tallied: with 23% of election day precincts reporting and almost exactly as many at-precinct votes as early votes counted, the overall count stood at Obama 711,759, Clinton 711,183 (49%-49%), a dead heat.

To catch up so quickly and produce those numbers, Clinton had to win the at-precinct vote in that quarter of Texas precincts by 59% to 41%...an exact reversal of the early voting Obama landslide. What we saw in Texas were essentially equal and opposite landslides, as if we were observing two not only separate but radically divergent electorates, one that chose to vote early and one that chose to go to the polls.

The early voting period in Texas extends from 17 days to four days prior to the election. Ordinarily explanations for a divergence of such magnitude, particularly in intra-party contests, would be due to time-critical phenomena such as late-breaking gaffes, scandals, debate blowouts and the like. But there was no such occurrence.

During the early voting period the average of 13 pre-election polls showed Clinton 45.6%, Obama 46.7%. In the three days before the election, after the early voting period had ended, the average of eight polls was Clinton 46.8%, Obama 46.1%, a very modest change and certainly not the 21% mega-reversal displayed by the early voting and at-precinct vote counts.

While there is no obvious explanation for the pattern observed, one hypothesis worthy of investigation is that one set of counting equipment (either early-voting or at-precinct voting) was accessed by malicious insiders and manipulated. If the pattern of pro-Clinton shifts were to hold, the place to investigate first would of course be the at-precinct voting equipment and county central tabulators.

Having won Ohio and Texas, Clinton remained viable but still in dire straits, leading directly to the most polarizing and divisive phase of the nomination battle.

4/23/08: Pennsylvania

In the ‘quiet’ interval during the six weeks prior to the Pennsylvania primary, the effects of Clinton’s revived (but precarious) position had ample opportunity to play out. The Clinton campaign went on the offensive, with the type of personal, negative attacks that both campaigns had previously eschewed. Obama was relentlessly portrayed as elitist and out-of-touch by the Clinton campaign (and by Clinton herself), a depiction the mainstream media began to echo almost as relentlessly.

And, sure enough, incidents emerged that played into this depiction—most notably Reverend Wright’s sermons and Obama’s own quote that seemed to both pigeonhole and patronize the working-class voters of Pennsylvania. These were replayed by the mainstream media in an endless barrage of coverage, all keyed to the theme that Obama might be too out-of-touch, and too close to the radical black fringe, to be president. Obama appeared to successfully counter that round of negative attacks, and it appeared to have little or no impact in his polling support nationwide – nor, indeed, in Pennsylvania.

Obama went into the April 23 primary trailing Clinton by 5% or less in pre-election polls, with no late movement to Clinton detected. It was viewed as essential by mainstream media pundits that Clinton win “by double digits” to maintain her viability and pick up the momentum required to win decisive superdelegate support.

First-posted exit polls⁵ for Pennsylvania reflected pre-election expectations, with Clinton leading 51.6% to 47.8%, a 3.8% margin. By late in the evening, however, with the count mostly in, it was Clinton by 9.4%--close enough somehow for the morning papers, networks, and websites to lead with Clinton’s “double-digit” win.

As with New Hampshire, Ohio, and Texas, there was a wide range of irregularities, glitches, and vote suppression incidents reported. Again, an exit poll disparity beyond the margin of error. Again, a departure, in the familiar direction, from the range of pre-election polling. And once again the final result was that Clinton received just enough to sustain her campaign, her “double-digit” victory, courtesy of a generous round-off. __________________________________________________________________________________________________________________________________________ ⁵ Weighted, 1421 respondents, approximate margin of error +/- 3%. ______________________________________________________________________________________________________________________________________

The Upshot

Just as with a spaceship's carefully-calibrated mid-course corrections that make an ultimate difference of millions of miles, it does not take much to radically change the course of a multi-election political contest. A few quick bursts from the retrorockets at the right moment(s) will do the trick. Of course the dynamics of a campaign can change legitimately, as a result of the thrust and parry process, exposure of weaknesses, refutation of apparent inevitability, etc.

But the shift in dynamics of the 2008 Democratic nomination contest strongly correlated with a string of election results that raised serious red flags independent of their impact on the race. Glaring discrepancies far beyond the margin of error of exit polls and pre-election polls, and the confounding of the expected electoral dynamics, produced results that had the precise impact of prolonging and intensifying the nomination battle. Had the primary election results jibed with those independent measures and expectations, it would long since have been wrapped up.

Anyone actually in a position to take advantage of the vast array of security vulnerabilities in the computers that run our elections would have an obvious interest in remaining undetected. The safest path would be to take only what you need to achieve your bottom-line goal, and not one vote more. Anything beyond adds risk without reward.

Thus, in keeping with our hypothesis that the fundamental goal of primary contest electoral manipulation was to create “plausible defeatability” for the Democratic ticket in November, we would expect little additional manipulation in the last stages of the Democratic contest. It is apparent that an Obama defeat in November (and more extensive Democratic losses in down-ballot races) can be spun as a plausible consequence of the intra-party strife that has already been depicted as weakening the party and its nominee, and of apparent Obama weaknesses exposed in the course of the grueling nomination battle.

With such a cover story safely in place, even an against-the-odds Republican “victory” in November could be successfully spun and sold to the candidates, their parties, the media, and the voters.

The “Mystery Adjustment” Factor in Polling

One final observation concerning the pre-election polling that sets expectations for candidates, the mainstream media, and the voters themselves. We are deeply concerned that these polls too paint a false backdrop against which the signs of computerized electoral manipulation by insiders will appear diminished in magnitude over time, or even disappear.

The reason for this concern is obviously not that the fraternity of pollsters are knowingly acting to support or conceal systematic computerized electoral manipulation, but rather that pollsters simply cannot expect to stay in business if they consistently fail to predict the “actual” electoral results. The worst problem for a pollster is to be consistently “off” in the same direction.

Put another way, pollsters are not paid for achieving some abstract statistical purity but rather for accurate predictions—however achieved. If one places oneself in the position of a pollster who, time and again, is faced with results that are, say 6 – 8% more Republican than their predictions, or shifted in the direction the right wing would desire, it becomes clear that one would begin making a “mystery adjustment” to whatever data emerges from a clean survey methodology.

Such an adjustment can be easily generated by changes in demographic weighting that can at least in part be justified by reliance on data emerging from previous elections, themselves manipulated. Call it a fudge factor if you will, but it keeps the pollster in business, while failing to make such a correction would be professional suicide.

By way of corroboration of this phenomenon, in public dialogue with a major-party polling consultant, the following shocking admission was made: If the Democratic candidate is not leading by 10% going into the election in their internal polling, they expect the race to be a toss-up. This internal candidate polling is—unlike polls published for public consumption—intended to paint a ruthlessly accurate picture of contest dynamics to help the party prioritize expensive get-out-the-vote drives and last-minute media blitzes.

The fact that even major-party pollsters must adjust their own results to account for the “mystery swing” to the right is a strong indication that much the same distorting protocol is already being employed in public pre-election polling. When manipulated elections serve as the calibration tool for pre-election polling, we lose yet another independent check mechanism on the official computerized vote tabulation process. This only deepens the crisis.

Conclusion

Election theft is certainly hard to prove, with virtually all hard evidence withheld as proprietary; and even well-supported allegations by credible journalists, computer scientists, security professionals and election integrity activists are given a wide berth by both the mainstream media and the established political powers of both major parties.

Yet, even with the limited tools at our disposal, we keep discovering evidence—in pre-election polls, exit polls, and published election results–that is consistent with a pattern of widespread covert manipulation of vote counts.

We will continue to investigate and report these anomalies until a thorough and unblinking investigation of suspicious results is undertaken by those in position to collect the additional evidence needed to establish incontrovertible proof.

But since many of those in the best position to investigate election anomalies are themselves elected officials, our best hope may be to follow the recent example of Ireland and the Netherlands—dispense with voting computers, and simply count our own paper ballots by hand.

Attachment	Size
2008_Democratic_Primary_Manipulation_EDA.pdf [14]	188.45 KB