August 6, 2006 - August 12, 2006
Apologies for slow posting since Wednesday - it's been a busy week. Taking a step back from the Connecticut primary, two new national polls were released this week, and both show slightly lower job ratings for George Bush than most of the other national surveys released in late July or early August. Do these indicate a new downturn for Bush? Our friend Charles Franklin has crunched and graphed the numbers and says no.
The most recent Fox News/Opinion Dynamics poll (story, results) puts the Bush job rating at 36% approve, 56% disapprove. However, as Franklin points out, the approval rating on this latest survey is exactly the same as their last poll fielded July 11-12, but the disapproval rating is three points higher. His post goes into detail on the
The poll getting more attention today comes from AP/IPSOS and shows a drop in Bush's approval from 36% to 33%, "matching his low in May," as the AP story puts it. Franklin takes a close look and concludes the latest AP result is a statistical outlier from the most recent poll trend. That is, other national polls have shown Bush's job rating holding steady or slightly rising, while the AP result is a sharp departure. He continues:
It is possible that the AP result signals a sharp break from the past. Given the track record of outliers in these data (over 1100 polls in all) that is not likely. Far more likely is that new polls will confirm that the trend has changed by modest amounts, either up or down, and that the next poll will be closer to 38.7% (both above and below) than to 33%.
This is not to say that the trend cannot change. We have seen three very clear examples of reversals in aproval trend since January 2005: in November 2005, February 2006 and May 2006. At some point approval may again trend down (or more sharply up, for that matter.) But it would not be a statistically good bet that the AP poll is where approval really stands right now
Read it all.
The nearly final 3.6% margin by which Ned Lamont defeated Joe Lieberman in yesterday's Connecticut Senate Primary (with 98% of precincts counted, Lamont leads 51.8% to 48.2%) was certainly closer than the margins on the final public polls. The two polls conducted last week by Quinnipiac and Research 2000 had Lamont ahead by six and ten percentage points respectively. An earlier survey by Quinnipiac had Lamont ahead by 13, and a mid-July survey by Rasmussen had Lamont ahead by 10.
The Quinnipiac pollsters will rightly point out that their final results fell within the 3.5% margin of error of their final poll, no small feat given the degree of difficulty in polling this unusual off-year primary election. But consider that the turnout of more than 282,000 voters or nearly 41% exceeded nearly everyone's expectations and far surpassed the 25% historic average for primaries featuring a governor's race. This leaves a few open questions, at least from my perspective:
- Did Lieberman narrow the gap in the campaign's final ten days, as suggested but not quite confirmed by the last two Quinnipiac polls?
- Or was Lieberman consistently closer in Lamont's rear view mirror during the final weeks than the public polls made it appear? Did the sampling methodologies and likely voter models of the public polls consistently exaggerate Lamont's during the campaign's final weeks?
It may be possible to answer some of these questions with the poll and vote return data, although we do not have access to all of the relevant data. For example, both campaigns conducted internal tracking polls that provide an independent assessment of the trend over the campaign's final week. Do those surveys confirm a late Lieberman rebound, or was the second-to-last Quinnipiac survey, the one that showed Lamont ahead by 13 points, an outlier?
Campaigns are often willing to disclose internal data once the campaign has ended and the dust cleared. Unfortunately, since the Lieberman-Lamont contest continues, such a release from either camp appears unlikely.
The geographic turnout patterns are also relevant. Charles Franklin has already posted an amazingly thorough (and graphical) turnout analysis of the turnout showing that Lieberman did better in the larger towns and cities, while Lamont did better in less urban areas. He also confirms the so-called "Volvo/donut" turnout pattern suggested yesterday by Hotline On-Call, that turnout was higher in the smaller towns where Lamont had an advantage, lower in the larger towns where Lieberman did better (see also Hotline's follow-up analysis this morning).
This pattern leads to two questions: First, how well did the various polls do in modeling the geographic distribution of the vote? What portion of their completed likely voter samples came from the smaller towns where Lamont did better? How did that distribution compare to reality and did that distribution vary significantly from poll to poll?
Second -- a question we can answer with data in the public domain -- how much did the "Volvo/donut" geographic pattern depart from past history? Vote turnout is typically lower in more urban areas. The question here is whether the urban-rural turnout gap was bigger or smaller than in previous elections. If so, "likely voter" models based on past experience may have been off as well. Of course, if the smaller towns contributed a larger than expected share of the total vote, it would imply that a poll sample based on past turnout geography would understate Lamont's vote, not overstate it.
It would be certainly be educational to try to answer some of these questions here, although most of the data we could use to try to answer these questions are in the hands of the pollsters.
PS: About that Exit Poll -- CBS News has posted a brief report online. UPDATE: At 2:00 EST CBS posted a second more in-depth report that includes results ot all questions. The complete report has much to chew on including cross-tabs by demographics (age, race, education, religion, income, union membership and self-reported ideology). The most relevant finding to current speculation about an independent candidacy is as follows:
Most Democratic primary voters would not support an Independent run for U.S. Senate by Lieberman this fall, even though the majority said they approve of the way in which he is doing his job . . . .
If Lieberman does decide to run as an Independent against Lamont and Schlesinger in November, he may find that many Democratic voters will choose their party’s candidate instead of him. In a hypothetical three-way race against Lamont and Schlesinger, Lamont would earn 49% of the votes of these Democratic primary voters, and Lieberman would receive 36%.
Among Lieberman voters, three out of four say they will support Lieberman again under those circumstances; 16% are not sure, and 6% say they will vote for Lamont. Lamont retains more of his voters; 88% of them say they would vote for him in November.
And oh yes -- the reason political junkies love exit polls -- the report also includes directly relevant to the central question I discussed above. The exit poll provides clear evidence that that Lieberman gained slightly over the last week of the campaign:
Three quarters of voters said they made their mind up about which candidate to support a while ago -- in the last month or even earlier, and those voters went for Lamont. But the race appears to have tightened in the last few days, and Lieberman ran ahead of Lamont among the 16% of voters who made their mind up in the last three days.
[Click the table for a slightly bigger, less fuzzy version]
Such a late trend does make a certain amount of sense: A disproportionate share of the late deciding voters -- probably including a lot of women, conservative/moderate Dems,
etc. -- flirted with the possibility of voting for Lamont but in the end
came returned to what was (for them) a safer, more familiar choice.
I noticed several hundred (at least) incoming Google searches on various combination of "Connecticut exit poll Lieberman Lamont," so let me pass on the official word regarding today's Connecticut primary from ABC's The Note: "There are no pooled network exit polls" [emphasis added].**
The Note also passes along this helpful information for those who plan to follow the returns:
The Connecticut Legislation and Elections website (LINK and LINK)
will be updated continually, as soon as information comes in. The state
is using a brand new system, which may or may not work. Should the
system not work, officials are prepared with a spreadsheet and the site
should be updated at roughly fifteen minute intervals with results.
Lever machines will dominate in today's primary, with a smattering of optical scan voting places. And here is a chart, courtesy of the excellent folks at the AP,
detailing the timing of the reporting of results on election night in
2004 to use as a guide for tonight.
First Reports from Counties: 8:35 pm ET
20% of Precincts Statewide by: 10:00 pm ET
90% of Precincts Statewide by: 1:01 am ET (Wednesday)
100% of Precincts Statewide by: 9:44 am ET (Wednesday)
Clarification: No "pooled" exit polls means no exit polls sponsored by the major networks, i.e. the National Election Pool (NEP) consortium. If anyone else conducts true exit polling in Connecticut, it's news to me.
Some organizations (such as the Associated Press and Zogby) have conducted telephone surveys on Election Day in recent elections
some places, and Zogby has taken to labeling his approach an "exit poll." I have no idea if any such efforts are underway in Connecticut today. I would have little faith in the ability of such a poll to project the outcome on Election Day, especially the Zogby model which apparently took three days to complete.
UPDATE (8:32 p.m.): Very alert reader Melanie reports in a comment below that she "filled out an exit poll for CBS news." Via email she adds:
It was an electronic device - the logo at the top said CBS News, with no
reference to our local CBS affiliate in Hartford. (It surprised me, which is
why I remember it.) There were a lot of questions - it took from 5-10 minutes.
She adds that most of the questions concerned the Lieberman-Lamont race, but for some reason "didn't ask about the race for the gubernatorial nomination." Further:
I just heard a reporter note on NECN (New England Cable News- not affiliated
with any of the majors, I don't think) that NY Times/CBS are doing exit
polling - I don't know if it's true, but that was what he said.
Obviously, the polls in Connecticut have now closed, so any results from this exit poll are likely to come directly from CBS News or the New York Times. ..
UPDATE II (8:48): Kathy Frankovic at CBS News confirms that they did in fact conduct an exit poll in Connecticut today in partnership with the Times, but will not be using it to project a winner in the Senate race. Stay tuned...
UPDATE III (9:30 a.m. 8/9): CBS has posted a brief report on their exit poll online.
For those waiting on the results of the today's Connecticut primary, here is a cautionary tale about the limits of polling in a primary election. Four years ago this week, I learned a lesson the hard way about what pre-election polls mean, and sometimes, what they don't.
Four years ago, I served as the campaign pollster for Rep. Lynn Rivers, then a four-term Democratic incumbent from Michigan who, by virtue of Republican redistricting, had been had been thrown into the same district with long time Democratic Congressman John Dingell. The August 2002 primary in which they faced off has some parallels to the Connecticut contest worth considering.
That primary. according to the Almanac of Politics, was "Michigan's most expensive House primary ever," with the two candidates spending a combined $4 million plus millions more spent by interest group independent expenditures on both sides. Dingell, who served in Congress for more than 45 years including years as a powerful committee chair had worked more recently with prominent Republicans to pass legislation. In the campaign, Dingell raised millions from favored interests, organized labor and prominent national figures in Washington. Rivers had, by all accounts, a far more liberal voting record and the backing of groups supporting the environment, abortion rights, gay rights and arms control and of course, the EMILY's List national donor network.
I hesitate to raise the Dingell-Rivers example because the differences with the Lieberman-Lamont race are in many ways as striking as the similarities. John Dingell, unlike Lieberman, was and remains a revered figure to his Democratic base. Unlike Lieberman, he maintained high favorable ratings from Democrats throughout the new district during the 2002 race. Moreover, both candidates had pledged to support the eventual winner, so there was never any talk of either running as an independent. While they differed on issues like the environment and abortion rights, Dingell and Rivers had very similar voting records on the issues of central concern to Democrats in 2002, such as the economy, health care and education. The race ultimately turned on contrasts of age, style and experience. And of course, the Iraq War -- the dominant issue of 2006 -- was still barely on the horizon in August 2002 (although two months after the primary both Dingell and Rivers voted against the resolution that authorized the war).
The valuable lesson from this story involves the limitations of polling. Dingell led in most of the early polls, but the race appeared to tighten as Rivers ran television advertising toward the end. The final polls conducted by the Detroit Free Press, EMILY's List and yours truly on behalf of the Rivers campaign all indicated either a dead heat or a slight but insignificant Rivers lead. My last poll had Rivers two points ahead. I never heard the exact numbers, but consultants for the Dingell campaign later told me that their internal polls showed essentially the same thing, except that they had Dingell ahead by a few insignificant points rather than behind. On Election Day, given the trend, they expected to lose. Yet when all the votes were counted, Dingell had won by a resounding eighteen-point landslide, 59% to 41%.
So what went so wrong? Other pollsters may disagree, but I believe we erred in our assumptions about turnout and the way we sampled "likely voters." Michigan, like Connecticut,
had recently switched to an held an August primary** and had not seen a seriously contested primary in years. Turnout for the August primary in the four previous off-year Democratic primaries had varied between thirty and fifty thousand voters. The most optimistic forecasters guessed turnout could go as high as sixty thousand voters. But the actual turnout stunned everyone: 98,952. That was nearly three quarters of the number that would later cast a ballot for Dingell against his Republican opponent in November.
This brings me to an important structural difference between the Michigan and Connecticut primaries. Unlike Connecticut, Michigan has no party registration and the 2002 primary was open to all voters regardless of their prior vote history. So independents and even Republicans were free to cast ballots but just showing up on Election Day. Moreover, most of the pollsters were sampling from lists of registered voters (rather than using a random digit dial or RDD sample to reach all telephone households) largely to match the gerrymandered district boundaries, but also to more efficiently identify those with a past history of primary voting. We will never know for sure, but my assumption is that all the pollsters erred by screening to tightly for Democrats and past primary voters, the kinds of voters who were more likely to support a liberal candidate like Rivers. I believe that our methodology led to an underestimate of the turnout surge in support of Dingell.
So what does this imply for today's Connecticut primary? Again, I am reluctant to draw too close a parallel because Connecticut's primary is open to registered Democrats only, and the public polls have been based on RDD sampling. The combination makes it less likely that the polls will miss as badly as in the Rivers-Dingell race.
Where the Dingell-Rivers race may be most instructive is regarding turnout. As in Michigan's 15th District four years ago, Connecticut voters have been bombarded with an unprecedented flow of information through all media. As Connecticut's Secretary of State Susan Bysiewicz put it yesterday, "You can't turn on the television, pick up the newspaper or listen to the radio and not know there's a primary," Add in door-knocking, phone banking, robo-calling and other activity and you have one inescapable message: There's a primary and it features a choice between two candidates with very different positions on the Iraq War. So Bysewicz has good reason to guess that, according Greenwich Time, "turnout for the primary could reach 40 percent, well above the 25 percent historic average for primaries featuring a governor's race on the ballot."
Some may argue that a larger turnout helps Joe Lieberman in the same way it helped John Dingell four years ago. I'm not so sure. This primary is open to only Connecticut's 696,823 registered Democrats. Any cross-over phenomenon is limited to the roughly 29,000 who registered for the first time or switched their registration to vote in this primary. Among the bulk of registered Democrats, I don't see much of an opening for an influx of enthusiastic, energized Joe Lieberman supporters who do not normally vote in primary elections. The potential for a big turnout among passionate anti-Bush, anti-war Democrats is another story entirely. The Quinnipiac poll completed in mid-July reported results for all registered Democrats, not just the "likely voter" subgroup. Ninety-one percent (91%) of the Democrats disapproved of President Bush's job performance, 93% disapproved of his performance on Iraq and only 13% agreed with Joe Lieberman's position on the war.
So here's my bottom line: Given the turnout puzzle means, I wouldn't be surprised to see tonight's result differ significantly -- in either direction -- from the recent poll results. Historically, polls in this sort of primary have not been the most reliable. However, if I had to guess, the energy and intensity appear to be all on Lamont's side and a large turnout probably works in his favor today. But that's just a guess. Time will tell.
**Correction: I obviously erred in the original version of this post in stating that Michigan had recently switched to an August primary in 2002. As multiple commenters point out, they have had August primaries for a long time. I'm not sure what memory lapse led to that goof, but I apologize for the error.
Quinnipiac has released another poll in Connecticut this morning that has the political web sites buzzing, perhaps a bit too much. While the latest survey shows Ned Lamont leading by closer margin than their last survey, the differences are small and evidence of a shift in momentum is sketchy at best.
This morning, the Political Wire tells us:
Sen. Joe Lieberman (D-CT) cut into challenger Ned Lamont's (D) lead the day before Connecticut's Democratic primary, according to a new Quinnipiac poll. The poll shows Lamont leading Lieberman, 51% to 45%.
Hotline on Call chimes in about Lamont’s "narrowing lead" with a "fairly simple" explanation:
Lieberman had a fairly good weekend of news coverage; virtually every Democratic voter in the state is paying attention.
Looking more closely at the numbers, MyDD’s Matt Stoller notices something important: "Among liberals, men, those with college degrees, and those making more than $50k a year, the race is basically the same." The difference between the two polls seems to be mostly among moderate to conservative Democrats.
But let's take a step back for a moment. The last Quinnipiac poll sampled 890 likely Democratic primary voters between July 25 and July 31 -- that’s from two weeks ago Tuesday through last Monday. The latest poll interviewed 784 likely primary voters from July 31 to August 6 -- last Monday through Sunday. Yes, it included the weekend, but half the interviews were conducted mid-week. So attributing the change to a "fairly good weekend" of coverage is a stretch.
A second caution: The difference between the last two Quinnipiac polls is small, and not quite statistically significant. Lamont’s support went from 54% to 51% (-3%), Lieberman’s from 41% to 45% (+4). Each poll has a margin of error of a little over +/- 3 (+/- 3.3% on the first poll and +/- 3.5% on the second), but remember, that applies to each individual candidate estimate not to the margin or the difference between polls. The apparent change between the two Quinnipiac polls is not, strictly speaking, something about can be 95% confident about.
Note the very careful language of the Quinnipiac release that avoids characterizing the differences in the Senate results as a change. The release lacks the usual horse race rhetoric about Lieberman "cutting into Lamont’s lead" or "closing the gap." Quinnipiac simply tells us how the current result "compares to" the last poll.
Finally, consider the intriguing possibility that Quinnipiac sample composition changed from poll to poll. This potential is difficult to evaluate, because while their release provides cross-tabulations by gender, education, income and self-reported ideology, they do not tell us anything about the size of these subgroups for either poll.
In my last post, I took a close look at the big differences by ideology and did my own calculations to try to guess at the percentage liberal. My guesstimate is inexact at best -- because of rounding and the possibility that some respondents were excluded from cross-tab because they did not report an ideology -- so I hesitate to make precise comparisons. However, it looks to me as if this latest poll includes fewer liberals than the last one, perhaps as much as 6-8 percentage points less.
Now such a difference may be a real change in the "likely voter" population, at least as defined by Quinnipiac. As described here previously, Quinnipiac defines likely voters as those with the greatest interest in the election, interest in politics and intent to vote. So if Joe Lieberman had a particularly bad week two weeks ago, it is possible that some of his supporters -- who tend to be moderate to conservative Democrats -- felt a little less interested in politics and a little more reluctant to vote (see discussion of the Gallup likely voter model in 2004 for evidence of exactly this sort of shifting in general election polling).
Regardless of the reasons for the change, this latest Quinnipiac snapshot of the race is probably the more accurate read, especially given that the results are closer to those obtained by Research 2000 and (if we believe the rumor) the internal Lamont tracking polls as well.
What is the reality? The challenge of polling in these situations makes it hard to know for sure, but we know that every recent public poll has shown Lamont running ahead of Lieberman. The two public polls conducted last week showed Lamont ahead by an average margin of 52% to 44%, and all of the public polls -- including the mid-July Rasmussen survey -- all fell within +/- 3-4% of that average. And of course sampling error does not allow for the differences in the way the polls sample and select likely voters.
So Lamont was probably ahead last week among those who will cast ballots in tomorrow, but conclusions about last minute shifts in momentum are highly speculative. Political polling in this sort of race is a lot more art than science. If you live in Connecticut and care about the outcome, ignore the polls and go vote.
UPDATE: Charles Franklin has a post up that covers many of the same points as I did above. He has applied his usual graphic touch, including this chart that plots the Leiberman and Lamont results on the Quinnipiac polls by ideology. Note that the first poll in the series reported results for all registered Democrats rather than likely voters:
The gains for Lamont among liberals (the solid dark blue line) are, as Franklin notes, not suprising. However,
Lieberman's losses among his core supporters, the moderates, are less
extreme but politically devastating. From a 67-18 lead 4/30 he dropped
to a thin 49-45 margin 7/31, and has recovered a bit in the 8/6 poll to
53-43. (Note that what I said above about the lack of statistical
significance of the changes applies even more so here. These shifts are
great fun to look at and speculate about and interpret, just as I am
about to do!, but they aren't distinguishable from random noise in the
The graph indicates that aside from being indistinguishable from sampling error, the change in moderate to conserative Democrats (the dotted lines) over the last week is small compared to the change since May