“Kohut: In the interview that we did, right after the Mexican presidential election, I asked Warren [Mitofsky] why he thought he did so well in calling Mexico. He said it was because they were more rigorous in the way they had the interviewers sample people coming out of the voting booths. Faulty sampling was a key to the bias that was apparent in the exit poll in 2004. What are you doing to address the issue of more rigorous sampling of the individual voter?”
“[Lenski] Yes, the evidence did support Warren’s contention in that regard. At the least, it showed that when the interviewing rate was less — when you were interviewing 1 out of every 10 people rather than 1 out of 2 people — we tended to have a higher survey bias in 2004.”
Now both Kohut and Lenski do not consider an alternative hypothesis: that the exit polling was accurate – it showed Kerry won – and was reversed on uncheckable machines that is, by fraud (see the hundred page House Judiciary Committee Democratic report, “Preserving Democracy: What Went Wrong in Ohio in 2004,” written by John Conyers, on errors in the Ohio election, all pointing in the direction of George W. Bush.). Further, they offer no real evidence that the initial exit poll in Ohio was wrong. For initial exit polling got Bush’s victory and Kerry margin of defeat right to three-tenths of a point in Utah, Kerry’s victory in New York and California and Bush defeat to three-tenths of a point. The same accuracy was true in 41 states. But that exit polling used the same frequency of sampling in these other states, as in the three it supposedly got bizarrely wrong. Those three are Ohio, Florida, and Pennsylvania where the final “results” deviated by 5 to 7.8 points. To make their case plausible, they would have to show how exit polling, designed by Mitofsky and Lenski – the most seasoned exit pollers – was right so extensively and in so many states in 2004 as well as in many previous elections, but off by so much – not conceivably a result of a purported special “sampling error” – in three states. To be reassuring even in a non-Edison monopoly, non-secrecy poisoned environment, they would have to answer this more detailed question and produce real figures. Not a chance. But leaving this failure aside, what Lenski names as problems could be remedied quite easily by the first reform which I suggested at the outset of this essay.
In addition for at least all Presidential, Senatorial, Gubernatorial and Congressional elections, a company independent of Edison or better yet, an independent group of academic experts should be hired or funded by the US government to do a separate and simple exit poll questionaire, asking at different locations, only who people voted for. They should sample at adequate intensity: 1 in 2 or 1 in 5 voters, if there is evidence, though 1 in 10 seems adequate. That exit poll could serve both as a test on the fairness of the election and a step toward a free market – or at least a competitive – check on the current Edison group/Media Consortium misuse of exit “polling.” In the Raw Story column, Joshua Holland emphasizes Lenski’s experience abroad in “Azerbaijan, Georgia and Venezuela.” Holland does not mention that these very exit polls were funded by and Lenski himself probably paid to do this by US AID (the State Department). Even at Edison, exit polling abroad is not simply the operation of a private firm, but often works for or with the U.S. Government. On one level, both Lenski and Holland do not speak of this Government connection in order to skirt questions from citizens. But either overtly and/or perhaps covertly, the U.S. State Department uses Lenski, Edison Research and other polling organizations abroad to do this work.
Now according to Eric Bjornlund and Glenn Cowan, Vote Count Verification: a Guide for Users, Implementers and Stakeholders, 2011, p. 52-54, written for US AID, initial exit polling was funded by the State Department since 2000 in a minimum of 14 cases.But it is doubtful that AID funds this polling at very high expense, as Lenski suggests. It is even more doubtful that Edison is paid less by the Media Consortium for trying to figure out an American election than, say, the American government pays for a “transitional” election in the Ukraine or Venezuela. And listening to Lenski’s previous comment: does Edison really charge less, in the United States, for its ostensibly more complicated polls? Is Lenski underpaid? But if Edison’s exit polls are so underfunded that the Company cannot get the range – the margin of error – in which a winner can be declared or the race announced too close to call, surely the public should demand of the 6 corporation-Media Consortium that they invest enough money and hire enough poll workers to make American elections “transparent,” too.
To The Fix at the Washington Post, Lenski is proud of the winners – Trump, Sanders in Michigan – Edison initial exit polling has identified. Phillip Bump is the author of The Fix – an unintentionally ironic name for a column about exit polls and the New York Democratic Primary – in the Washington Post. On April 22, three days after that Primary, Bump also interviewed Lenski for “How exit polls work, explained.” Note: Bump never speaks of the actual discrepancy from the initial exit poll of 52 for Clinton – 47.6 for Sanders to the final 57.95 to 42.05 “rout” For that discrepancy is vast – 11.9 points, 6.2% beyond the margin of error. Yet asking precisely about that discrepancy would be what a reporter would do in an article of this title. A reporter: one whose purpose is not to downplay the magnitude of these discrepancies nor simply to promote Edison’s successive adjustment of ostensible exit “polls” to match machine recorded “results”…
To Lenski, Bump is obsequious:
“Exit polls get a bad rap. I don’t have a dog in this fight, but it’s true.”
“Exit polls provide remarkably quick data on elections that comes with a smaller and smaller margin of error as an election night passes, telling us who came out to vote for which candidate and why. “
Here without a hitch, Bump adopts Lenski’s description of Edison’s doctored exit “polling” in the United States, and ignores the goal of ensuring fairness in elections. And in the Holland interview at Raw Story, Bump also adopts the idea that knowing the demographics of voting rather than ensuring its fairness or transparency is the goal. Bump is a dog in this fight.Further, Bump’s introduction of Lenski dances around the issue of why a single corporation has been given a monopoly of exit polling. Bump says breezily: “now it’s centralized through Edison.”
“To figure out why, I reached out to Joe Lenski, executive vice president of Edison Media Research. Since 2003, Edison has been conducting polling for the National Election Pool (NEP), a group of six media organizations that includes Fox, CNN, ABC, CBS, NBC and the Associated Press. Once upon a time, each of those organizations would have run their own exit polling; now, it’s centralized through Edison.”
All these corporate initials might be Very Impressive until one notices: this is just a monopoly, speaking solely through one person. This monopoly fails to make evidence available to citizens with questions. And as Bump blurts out, it rejects initial exit polling – as the State Department and Europe use it – as a test for fairness in favor of getting “a smaller and smaller margin of error as an election night passes.”But with encouragement from Bump, Lenski hits on the idea of a mistaken “cluster sampling” of young voters to explain the very controversial, vast initial exit poll/machine “results” discrepancy in New York. Lenski refers to an inadequate sample which contains a cluster of responders not repeated in other samples (Nate Silver exaggerates this occasional error in his sometimes bizarre objections to exit polling ).
Worse yet, the unmentioned initial exit poll/machine “result” discrepancy was 11.9%, 6.2% outside the margin of error. But no such vast error was made in sampling young people elsewhere in the state. So why this wild, distorting burst for the young among samplers in the City? But neither Bump nor Lenski adduce any numbers (the claim would be laughable, if they did). For even supposing that this claim about oversampling is reasonable, it could not account for the huge discrepancy in favor of Clinton. Now, Lenski and Mitofsky as designers, and Edison Research, have long experience in doing exit polls in the U.S.. So in fact, Lenski says, poll takers quickly correct for overemphasis on younger voters. But that observation removes any interest from his and Bump’s claim about how the initial exit poll was, ostensibly, distorted. It wasn’t… And that leaves only Lenski’s (and Bump’s) deceptive description of successive exit “polls” which accord with the machine “results.” That is not a correction of a sampling error on the initial exit poll; it is a corruption of what exit polling is supposed to do. It is thus, contra Lenski, a huge change from the design of exit polling by Mitofsky starting in 1957. It provides cover for an uncheckable “result.” It travesties both the impression and the reality of fairness produced by a reasonable use of initial exit polls.
Note a further ambiguity here. On the one hand, Lenski says that Edison initial exit polls actually forecast the winner (he does not speak of “transparency” except in “transitional democracies”…). For instance, he boasts that Edison’s exit polling is entirely right for Trump, and surprisingly, for Sanders in Michigan. On the other hand, however, Lenski says below that Edison hastily incorporates actual precinct counts taken on machines into what are then no longer exit “polls.” Instead, they are adjustments by Edison or electoral officials, not based on talking with voters. Edison does this to report the “winner” in the sense of whatever shows up on the machines. In the latter case, however, initial exit polling is no longer a test or check of fairness, but so-called exit polling is a camouflage for whatever the machines show. Unless one assumes vast error in initial exit polling – and there is no reasonable basis to do so given wide international and national experience as well as Lenski’s own comments about Trump in New York and Sanders in Michigan, one must doubt Lenski’s adjustment. For the grotesque assumption in the corporate media is that American elections are, by definition, above reproach. Even when only 19.7% of eligible New Yorkers vote in the Primary, the second lowest total to Louisiana. Even when 123,000 voters are removed from the rolls in Brooklyn and two election officials immediately fired. Even when there are widespread reports of very unusual 25% disenfranchisment of Democrats at polling places in the Bronx and Long Island, Even when some very large number of registered voters were forced to cast uncounted “affidavit ballots” because their registrations had “disappeared”.In the corporate media, there are no memories even of Tammany Hall or Mayor Daley….
No wonder the Media Consortium is so cautious about releasing initial exit poll figures on their “place for politics” shows. For as Lenski notes, until the polls close, they mainly release demographic questions about voters or particular questions about whom voters trust. Yet even after the polls close, an exact figure for immediate exit polls – though known – is not routinely announced by reporters for the Media corporation. For in the United States, testing for fairness is not what Edison, the Media Consortium and ES&S and other machine manufacturers are about. On CNN, when the New York polls closed, Wolf Blitzer announced the initial exit poll of Clinton 52-Sanders 48 because the first results coming in would plainly close. That’s what Blitzer said. And Peter Jennings on MSNBC made a similar announcement. But inside the United States, as Blitzer’s announcement at 9 pm Eastern of a close Democratic race in New York on CNN shows, initial exit polling is dangerous for Edison, for corporate media mythology about the accuracy and fairness of elections, for the elite, and for those who falsely gain – there are some – electoral “victory”…It can confirm the impact of widespread, illegal curtailment of voting, clash with machine reported “results.”
Lenski does not give us the evidence and argument for the huge discrepancy in New York in detail or explicitly, because there is no justification for Edison’s inversion of what exit polling is.Listen now carefully to Bump’s and Lenski’s language about the so-called oversampling young people:
“FIX: So the New York issue, I can’t help but notice that you said that younger people are more likely to fill out the surveys and then also that Sanders was over-represented in the initial estimate. Do you think there’s a link there?
LENSKI: Oh, yes, there’s definitely a link there. We’re adjusting for that throughout the day. As I said, we know the response rate in our 35 precincts. We know that younger voters are more likely to choose to fill out the questionnaire than older voters, and that’s typically the case so we’re already making those adjustments.
Obviously in this case, that was even more than normal. As soon as we started getting sample precinct returns, we made that adjustment even more so that we’d match the actual results.”
Inside the United States, Lenski, echoed by the Washington Post, here inverts the purpose of exit “polling” “VW-style”: “As soon as we started getting sample precinct returns, we made that adjustment even more so that we’d match the actual results.”
Lenski attempts to breeze on:
“There are other issues here. We’re trying to make our best estimates on turnout by region of the state. In New York City, our exit poll reporting of how many people had voted in our precincts showed New York City making up 45 percent of the vote in the Democratic primary. It ended up making up 51 or 52 percent. So that’s another adjustment we’re making after poll closing, and doing the best we can with incomplete data before the polls close.”
Now it could be true that pollers underestimated the vote in New York City, as he suggests, though it would be interesting to know how different that would have made an initial exit poll. For the enthusiasm of huge Sanders crowds in the South Bronx, Manhattan and Brooklyn all suggest, if Lenski’s claim were right, that it might not have affected that exit poll much. In any case, Lenski’s claim sounds plausible, except that the initial exit polling – something that produces an accurate range – has already been mixed, that is, made deceptive, with machine-based “precinct results.” Lenski then crows:
“In Michigan, we actually had exit polling all day showing Bernie Sanders up by two points though every pre-election poll had Hillary Clinton up by 10 points or more — so we’re sitting out there on a limb the other way. In that case, the exit poll was right and the pre-election polls were wrong. It happens both ways.”
Now Hatlem also put in a phone call and sent emails to Joe Lenski, Executive Vice-President of Edison, to ask how Sanders could have lost 33 votes between his first exit poll at 9 o’clock and the second at 11. He also sought to press Lenski on the very statistically implausible 11.9% discrepancy between exit polls/machine results.No one answered. Hatlem left a message. He promised to put up Lenski’s response if he called back. Needless to say, refusal to talk to those who have questions like the ones I posed at the beginning – to talk only to “journalists” reporting for the corporate media – is not transparent.In addition, contradicting Lenski’s odd claim about pre-election polls, US AID has never funded pre-election polls to test the fairness of elections. Instead, the State Department funds the use of initial exit polls. Since one talks randomly to people who just voted, real exit polling provides a far more precise picture of what happens than pre-election polling. In addition, the discrepancy in New York city and in the “results” for Hillary are too great. Nothing Lenski says can reasonably explain this gap. Instead, there is lot of fluff and error– again, unchallenged by Bump – in Lenski’s comments.Lenski continues:
“’While everyone is talking about the Democratic side, we went out at 9 o’clock saying that Donald Trump was going to get 58 percent of the vote. [Note: this exit poll result was not announced by Wolf Blizer on CNN or by Peter Jennings at MSNBC…] He got just about 60 percent of the vote. Everything we did on the Republican side hit the mark. I understand when the data moves as much as the Democratic data moved between 9 o’clock and 9:45, that causes a lot of consternation out there. But there are plenty of other states where we’ve been right on.”
Yes, initial exit polling is usually right. Edison does not make large mistakes in doing it. But if the election is flagrantly changed compared to the exit polls, the obvious suggestion is that something unfair occurred. Further, Lenski has worked with AID abroad. He was intimate with Mitofsky. He knows all this. What he says is successive exit “polling” is deception. Still, Lenski correctly invokes Trump all evening and “plenty of other states where we’ve been right on.” But this true claim contradicts point-blank his lame, unchallenged comments above, to Joshua Holland at Raw Story, about how American initial exit polls do not pick a range of fair results for winners and losers. Lenski adds more fluff:
“It’s a survey like any other survey. There are sampling issues, there are non-response issues, etc. We’re making the adjustments we can to make this data as accurate as we can with the incomplete information that we have.”
“And as time goes on, we have more and more complete information.”
No, in all cases, the exit polls are complete and set a range of fair results, once the pollers have spoken with randomly selected voters and the polls close. In their place, Edison wrongly substitutes machine “results,” often without a paper trail, and as noted above, often with sloppy arithmetic. What Edison and the media consortium do is not exit polling as a test; instead, what is misnames “polling” is a manipulation by officials to sanctify machine “results.” Against his Raw Story interview with Holland, Lenski says again to Bump that the purpose of exit polls: “is to project a winner.” He tries to soften this with a comment on seeking a demography of voters and a falsehood that “historically” only the latter – and not fairness or the winner – is what such polling “mainly” seeks.
Nonetheless, the contradiction stands out:
“’There are two important uses of the exit poll. One is to project a winner. But the main use of the exit poll that night and historically is to have the most accurate representation of the demographics of voters.’”
’In a state like New York, we had 35 exit poll precincts and 50 sample precincts in the state. The sample precincts include the exit poll precinct. We have a reporter in each polling location in the sample, and as soon as the results are posted at that polling place, they will call those results in.’
‘Shortly after poll closing, we can quickly compare what the exit poll of that precinct said the votes were and what the actual votes were. So that’s when you’ll see a fairly quick adjustment to the exit poll estimates after poll closing.’
Wolf Blitzer reported the actual exit poll at 9 o’clock. He knew – reporters do, even when they don’t say so, that exit polls are right and hence, the huge initial discrepancy in the recorded figures would come down as the evening went on.In contrast, Lenski says, Edison officials hastened to dummy up the exit poll with “precinct results” to get closer to the machine “report.”
Let us compare what US AID says about exit polling with what Lenski says. Bjornlund and Cowan’s Vote Count Verification: a User’s Guide for Funders, Implementers, and Stakeholders, prepared by Democracy International for US AID, states at p. 52:
In recent years, domestic and international organizations have increasingly turned to exit polls to verify the officially reported results in the transitional elections of emerging democracies. Outside observers have credited exit polls with playing a key role, for example, in exposing fraud in Serbia and Mexico in 2000, Georgia in 2003, and the Dominican Republic and Ukraine in 2004. U.S.- funded organizations have sponsored exit polls as part of democracy assistance programs in Macedonia (2002), Afghanistan (2004), Ukraine (2004), Azerbaijan (2005), the West Bank and Gaza Strip (2005), Lebanon (2005), Kazakhstan (2005), Kenya (2005, 2007), and Bangladesh (2009), among other places.”
“Exit polls have long been employed in developed countries to quickly predict the outcome of elections. If conducted in countries with a history of democratic elections and in which citizens have reasonable confidence in their own safety and security, then well-designed exit polls can serve as an effective method for projecting election results.”
That describes Europe, particularly Germany, which counts paper ballots over a day but releases an initial exit poll just after the polls close, establishing winner and losers uncontroversially except in a very close election. They retain the paper ballots…And even a close election, that is, one within the margin of error of the exit poll (usually 2 ½ points or less), the winner and the loser are, to the people, are transparent. Unless there are massive pre-voting violations, the result is fair. But that is not remotely a description of the United States. For in Lenski’s own words, cited uncritically by Bump in the Washington Post, initial exit polling is not to test transparency or fairness, but to accord, as quickly as possible, with machine “results.” Neither the appearance, nor, on questioning, the reality of what Lenski describes as exit “polling” accords with fairness. Elsewhere I have analyzed Edison’s Report for 2015 which spells out its use of exit polling in Iraq in 2014, analyzes who won from different parties, and announces the list of American primaries it will cover in 2016. That Report makes no mention of gearing successive exit “polls” to machine “results.” Lenski’s rationalizations about New York on April 19th – there, the discrepancy of the exit poll/machine “results” is too glaring and well-known – attempt to give some appearance that what he calls successive exit “polling” keeps a shred of the accuracy of initial exit polling and its implicit role as a test of “transparency.” It does not. Lenski has to contradict himself, slip and slide, incorporate precinct counts on machines and reknit so-called exit “polling” into the emperor’s new clothes: an ornamenting of the “results” with no physical way, no paper vote count, to check their accuracy…
This is morally corrupt. This is, intellectually, specious. That this is the best the Washington Post, Edison and the Consortium can do – blaming others for “weak-mindedness” – is an embarrassment. With his long, international experience, however, Joe Lenski is probably quite aware that what he stands for is a betrayal of fairness, democracy and the purpose of exit polling itself. He does not return the calls of skeptics, like Hatlem…