About this Blog

RealClearPolitics HorseRaceBlog

By Jay Cost

HorseRaceBlog Home Page --> Polling/Public Opinion

A Note on Gallup's Party Identification Map

Today, Gallup released its results of partisan identification in the 50 states. The results are, as usual, interesting.

Gallup Party Identification July 2010.jpg

This map does not correspond with the national presidential map terribly well, in that it underestimates Republican electoral strength. Why is this?

Part of the issue probably has to do with the evolution of American partisanship. You'll note that most of the "Republican" states are in the Great Plains: Kansas, Nebraska, South Dakota, Wyoming, Montana, Idaho, and Utah. These states have historically been Republican since they were brought into the Union. Actually, many of them were brought into the Union in 1889. The Republicans had control of the presidency and both chambers of Congress, and quickly added these states, which they believed would vote staunchly Republican.

They voted for Bryan in 1896 and Wilson in 1912 and 1916, but otherwise they were staunchly Republican up through the Great Depression. Kansas, Nebraska, and the Dakotas were the first Roosevelt states to peel away from FDR's coalition, voting Republican as early as 1940. And while Harry Truman did well in this part of the country in 1948, they have been pretty reliably Republican since 1952.

These states are thus but a handful that have moved very little in over 100 years in terms of party alignment. Most other states have moved from one side to the other - Vermont used to be the most Republican state and South Carolina used to be the most Democratic. Now, as Gallup finds, it is basically reversed. Even within states we often find major changes: Democrats used to do well in Western Pennsylvania and Southern Illinois while Republicans were strong in Eastern Pennsylvania and Northern Illinois, but now both intra-state trends are reversed.

Party loyalties can survive for many years on a state and local level even if voters have moved on the congressional and presidential level. So, it's fairly common for districts to be reliably Republican for national offices, but more amenable to the Democrats at lower levels. And vice-versa. For instance, Kentucky has been voting staunchly Republican on the presidential level since 2000, but Democrats outnumber Republicans in the lower state house by almost 2-to-1. New York is now a highly Democratic state on the presidential level, but it has a long history of Republicanism, which shows up in the Democrats' very narrow control of the state senate.

If you look carefully, you can find such vestiges of the old party alignments all over the country. They also show up in the Gallup map. States like Georgia, Louisiana, and Texas have become increasingly Republican on the presidential and congressional level in the last 30 years, but Gallup has them listed as "competitive" in no small part, I'm sure, because many functional Republicans called themselves Democrats when Gallup inquired about their preferences.

Interestingly, the reverse does not seem to hold true on the Gallup map. Old Republican states like California or Vermont show up as solidly Democratic. Why might that be? It might have to do with the divisiveness of the Bush presidency, which might have pushed a lot of "liberal Republicans" into the Democratic fold.

It might have to do with the depth of commitment that the South once had to the Democratic party, which vastly outstripped any region's Republicanism. Franklin Roosevelt won every state except Maine and Vermont in 1936. Yet if he had been a Republican, he would surely have lost the old Confederacy, whose loyalty to the Democratic party was put in jeopardy only when the Democrats ran Al Smith, a Catholic, for President in 1928. That's how committed the South once was to the Democratic party. Republicanism in the South is still a fairly new development, just 50 years old or thereabouts, and so perhaps a lot of nationally Republican states still have commitments to the Democratic party that manifest themselves on the Gallup map.

It might also have to do with the fact that the Gallup poll is of "national adults," which can favor the Democratic party. We see indications of this in polls of national adults on Obama's job approval, which tend to be friendlier to the President than polls of registered voters or especially of likely voters; ditto generic ballot tests. The Gallup poll's numbers on un-leaned partisanship tend to track the exit poll results on base party preference fairly well, but Democratic advantages can show up when Gallup asks Independents to which party they "lean." For all of 2004, for instance, Gallup found that, when leaners were counted, the Democrats had a 3 point advantage in party identification. But on Election Day, Republicans outnumbered Democrats and un-leaned Independents basically split between Kerry and Bush. In 2006, Gallup found an average Democratic advantage of 10 points when Independent leaners were included, while the Democrats won the House by 8 points. In 2008, the Democratic advantage when leaned Independents were included was 10 points (again) while Obama defeated McCain by 7.

These factors - the stickiness of old party alignments, the effect of the Bush presidency on the Republican brand in Democratic areas, the deep loyalty of the South to the Democratic party, and the Democratic tilt of a "national adults" sample - probably explain why the Gallup map bears very little resemblance to the red-blue divide we take for granted today. So, Gallup's results are interesting from a sociological/political science perspective, but I think they bear only little relevance to voting preferences in presidential elections or the "nationalized" congressional election we're set to hold in November.

-Jay Cost

Gallup's Bouncing Ball

This week, Gallup's "generic ballot" number - which asks people if they plan to support a generic Republican or Democrat in the upcoming congressional elections - found a big boost for the Democrats, who bounced out to a 6-point lead.

Dems Lead Generic by 6.gif

The jump was sufficient to merit comment from Allahpundit as well as other conservatives. And I'd like to toss my two cents in.

Polls tend to have house effects, and most of us tend to notice these effects when they result in a pollster falling on one side of an average or the other. Rasmussen, for instance, tends to have Obama's job approval on the low side while ABC News/Washington Post usually puts the President on the high side. But there are other types of house effects, and Gallup has an interesting one: it is kinda bouncy.

For instance, look again at that above graph. Over the course of the last two months, Gallup has shown as much as R+6 on the generic ballot and as much as D+6. In a small poll, that might be attributable to sampling error, but the Gallup sample is over 1,500 registered voters, which produces a margin of error of less than 3 points.

Is it really the case that there has been a 12 point swing in the last two months? I doubt it. Looking at the RCP generic ballot average, the only two pollsters who have done multiple generic ballot questions during this time are Fox News and Rasmussen, and neither found such substantial swings.

This is not the only example of Gallup's bounciness. If you are anything like me, you wind your way over to Gallup.com at 1 PM Eastern Standard Time, as that is when it releases it's latest numbers on the President's job approvals. Those numbers move around more than any others out there. On Sunday, for instance, the President's job approval number rose 3 points, and his job disapproval fell 3 points. That made for quite a swing, which is noteworthy because the poll is based on 1,500 total respondents over a three-day track. That means about 500 respondents every day. For his net approval to fall 6 points by cycling out one day and adding another must have meant a very substantial one day movement. If you watch Gallup every day like I do - you'll note that such swings are fairly common, much more so than Rasmussen, the other daily tracking poll.

This is not a recent phenomena with Gallup, either. Check out its trial heats for the later stages of the 2000 presidential election campaign. Note October 2000 in particular:

2000 Bush v Gore.gif

Holy cow! At the beginning of the month, Gore had a 12 point lead, but at the end of the month it was Bush who had an equally large lead. That's some bounciness. And I would note that ABC News poll had a much more stable track through the month of October, and none of the other pollsters showed such outsized movement.

As for the generic ballot itself, it is historically a difficult metric. For instance, in 1998 the Republicans won a solid if uninspiring 2-point victory in the House popular vote, but Gallup's November generic ballot showed the Democrats up 7. In 2002, the October generic ballot had the Democrats up by 6 points, but the GOP went on to win the House popular vote by nearly 5 points. In 2006, Gallup's post-Labor Day generic ballot was tied among likely voters, but a month later the Democrats were out to a 13-point lead. Its post-Labor Day poll found something similar in 2008, leading Gallup to declare that "The Battle for Congress Suddenly Looks Competitive." It wasn't. In both 2006 and 2008, Republicans had a rough month of September, and in both cases the party's trouble really began after the Gallup polls. Still, in historical retrospect, I do not think that the "actual" generic ballot numbers were ever as tight as Gallup found, or as far apart as it found just a few weeks later.

Now, don't get me wrong. The problems with the generic ballot question are due to the limitations of the question itself, not how Gallup handles it. Democrats have a generic edge in nationwide party identification, which often does not materialize on Election Day, so historically the generic ballot favors the Democrats. There's nothing that Gallup can do about that. Plus, there is nothing wrong with having a house effect. I'd be suspicious of a poll that had none. Gallup is a great, reliable pollster that has been doing good work since before most of us were born. It would not be in business today if it wasn't. And kudos to Gallup for offering up more and more of its historical data, which is not only interesting but also a real public service.

My point here is fairly modest: it's incumbent upon us, the consumers of polling data, to digest it properly - which means generally that we have to be aware of house effects, and specifically in the case of Gallup we should not get hung up on every little inflection point. Gallup bounces around quite a bit, and it has exhibited this quality for some time. The best approach in handling this is to average its results across a few weeks.

When we do that for the Gallup 2000 numbers on Bush v. Gore, we find Bush holding a modest 4 point lead, 47 to 43, which should make intuitive sense given the dynamics of that race in its final stages. Similarly, the average generic ballot since Memorial Day 2010 shows the two parties essentially tied, which sounds about right to me for a measure of registered voters several months before Election Day.

-Jay Cost

ObamaCare is Politically Vulnerable

Liberal commentators are comparing the passage of ObamaCare to other landmark pieces of legislation - like Social Security and Medicare. I agree that in the provision of social welfare, this bill ranks nearly as high. But when you examine how the welfare is provided - it is strikingly inferior. Franklin Roosevelt and Lyndon Johnson made use of an ingenious social insurance system - promoting the idea that we all pay in today to take out tomorrow. It was consistent with American individualism. It was simple. It was intuitive. It was bipartisan.

Obama's new system has none of those virtues. It's an impenetrable labyrinth of new taxes, benefits, and regulations, passed on the narrowest of possible majorities with more than 10% of the Democratic caucus joining every Republican. Even Wile E. Coyote would be embarrassed by its inefficiencies.

Still, the thought among its proponents at the moment is that the legislation, once enacted, cannot be repealed. It will have the benefit of our system's strong "status quo bias." Accordingly, expect yesterday's critics of the filibuster to become its valiant defenders should push come to shove.

The status quo bias is a very real thing, and it makes the Republican efforts to modify or repeal challenging. The GOP must control the entire government by January, 2013 to enact major changes to the legislation. By then, the thinking goes among proponents, those with a personal stake in preserving the legislation will be in place to protect it, just as seniors have been on guard against raids on Social Security.

Yet it's not that simple. The Democrats crammed a $2 trillion bill into a $1 trillion package by delaying the distribution of most benefits for four years, until 2014. This creates two major political vulnerabilities for ObamaCare.

The first is an imbalance between winners and losers through the next two elections. Harold Lasswell defined politics as who gets what, when, and how. By this metric, ObamaCare is bad politics for the foreseeable future. Like any major piece of legislation, this bill assigns winners and losers. The winners will be those who today are uninsured, but who will (eventually) acquire insurance. But there will not be a major reduction in the uninsured until 2014. So, the actual winners are going to be pretty few in number for some time.

Meanwhile, the losers begin to feel the effects immediately. Between now and the next presidential election, ObamaCare is going to pay out virtually zero dollars in benefits, but it will take billions out of Medicare. This is bad for seniors. They have an incentive to oppose portions of this bill (while supporting others, like the closing of the "Doughnut Hole," which Republicans will never repeal). While the Democrats will claim that this reduction in benefits will have no effect on the quality of their care, CBO is much less certain:

Under the legislation, CBO expects that Medicare spending would increase significantly more slowly during the next two decades than it has increased during the past decades (per beneficiary, after adjusting for inflation). It is unclear whether such a reduction in the growth rate of spending could be achieved, and if so, whether it would be accomplished through greater efficiencies in the delivery of health care or through reductions in access to care or the quality of care. (Emphasis mine)

The italicized sentence is an enormous political problem for the Democratic Party. After decades of developing a reputation for defending the interests of senior citizens, the Democrats have put it in serious jeopardy with this legislation. And they've done so right at the moment when demographic shifts are making the senior population more powerful than ever.

Why create such an imbalance between winners and losers? The Democrats are not fools. Why would they do this?

The answer is pretty simple: to hide the true cost of the bill. They don't want to push a $2 trillion program now because this country is facing the greatest deficit crisis it's seen in decades - and such a price tag does not make for good politics these days.

These budgetary gimmicks enabled them to pass the bill, winning over enough self-described "deficit hawks" in the Blue Dog wing of the party to limp to 219 in the House last night. Yet their smoke and mirrors can only mask, not alter the reality, which is this: at a time when the country is facing an enormous deficit problem, the Democrats have created another significant financial obligation for Uncle Sam. This is the second major political vulnerability of ObamaCare.

It's easy to forget these days, seeing as how we've been on a 15-year break from the politics of deficit reduction, just how brutal it tends to be. If you want to know why the parties have become so polarized in the last 30 years, the deficit is a big part of the answer. When Reagan indexed the tax code and stopped runaway inflation, governmental bean counters couldn't depend on bracket creep to solve future imbalances between taxes and spending - and so the lines between the two parties were drawn starkly and clearly.

Deficit reducers always have to choose between two undesirable alternatives: cut spending or raise taxes. The problem with both tactics is that somebody loses while nobody really wins. The benefits of a reduced deficit are diffused across the population and are but weakly felt. Tax increases or spending cuts are felt directly and intensely. Typically, to balance the budget, somebody has to be made worse off tomorrow than they are today.

But not when it comes to ObamaCare, at least not prior to 2014. The benefits could be altered to ease the deficit burden without making anybody worse off tomorrow than they are today. Of course, the beneficiaries of the subsidies would not be as well off tomorrow as they expect to be, but that's different from being made worse off. That could be an important distinction if the politics of deficit reduction are as fiercely zero-sum as they have been in decades past. If it comes down to a choice between a new tax on the middle class or scaling back the unimplemented provisions of ObamaCare, guess what the policymakers in Washington, D.C. will choose.

We're definitely heading toward some kind of hard choice about the deficit. If we weren't, the Democrats wouldn't have employed all those gimmicks to claim that the bill costs less than $1 trillion. They know people are worried about this issue.

Last week, President Obama said again and again that the time for talk is over. Yet this week he's going on the road to defend his new bill. This is why. ObamaCare is politically vulnerable. It lacks the bipartisan support that created and protected new entitlements in decades past. The public does not have confidence in it. Worst of all, it creates an imbalance between winners and losers for four years, and it amounts to a staggeringly expensive new entitlement at a time when the country has to think hard about how to trim its sails.

Follow me now on Twitter!

-Jay Cost

Another Look at Obama's Job Approval

If you are looking for a good snapshot of where President Obama's job approval is right now, you cannot do better than the RealClearPolitics average. It's intuitive, straightforward, and indispensable.

Another way to look at Obama's job approval is to examine the trend line for each pollster. This can offer a way to control for their "house effects." The following chart does that by looking at the monthly average of eight major media pollsters (Fox, CBS, CNN, Ipsos, Pew, NBC, ABC, and AP).

Obama Job Approval.jpg

A few observations are in order:

1. By separating the pollsters from one another, we can see the various house effects. For instance, CBS and ABC are the most favorable polls to Obama while Fox and NBC tend to be the least. AP and CNN are the "bounciest." Some months, they are above the average. Other months, they are below.

2. Obama's job approval slid precipitously from July through August. This coincides with the heating up of the health care debate. This trajectory is consistent across all eight pollsters.

3. The President rebounded a bit from his August/September lows, but he is now at or near his lowest point in all of the polls except the (bouncy) AP poll, which had him much lower in September than the other polls. This chart makes that clear:

Obama High and Low.jpg

4. The polls generally find Obama's overall job approval higher than his approval on various issues. For instance, these are the results of the latest ABC News/WaPo poll:

ABC News:WaPo Issues.jpg

One can't help but wonder if a legislative success on the health care package will result in a further decline in the President's job approval rating.

5. What will be interesting to watch next year is whether the President's job approval slides further as the campaign begins in earnest. Will the Republican argument against Obama and the Democrats - once it hits the airwaves - damage the President's standing further? It is possible. Ronald Reagan and Bill Clinton both suffered about 8-point declines in their job approval ratings from January to November of their first midterm years. [Obama is about where both Presidents were at this point in their terms, a little behind Reagan and a bit ahead of Clinton.] George W. Bush's net approval dropped 36 points in 2002; of course, it was very high after 9/11. Also regarding Bush 43, when the Democratic campaign against him heated up in early 2004, his net job approval slid 13 points from the first of the year to the beginning of the summer.

Will the Republican argument against Obama push some voters who disapprove of Obama on specific issues into overall disapproval? Will it push some of those marginal approvers into disapprove/don't know?

Follow me now on Twitter!

-Jay Cost

The Public Option in the NBC News/Wall Street Journal Poll

Last week I argued that question wording might be influencing polling outcomes on the public option - generally skewing the results closer to the Democratic side of the ledger because of contested buzzwords like "choice," "competition," and "option."

I noted at the time that the best way to test this theory was via an apples-to-apples scenario in which we can hold the pollster, the methodology, and the time of the poll constant. That's why I thought the Rasmussen results were significant: Rasmussen changed the wording of questions on the public option and found markedly different results.

The new NBC News/Wall Street Journal poll, conducted by Hart/McInturff, gives us another such opportunity. They split their sample into two groups (A and B), and ask each subsample a different version of a public option question.

Here's the first version, asked of subsample A.

NBC:WSJ 1.jpg

This is your typically tilted question. The idea of a "choice" is referenced - again, Republicans would hotly dispute this. In this specific wording, respondents are asked how they feel about "(giving) people a choice," forcing opponents of the public option to play the part of Ebenezer Scrooge. Unsurprisingly, this wording produces some good results for public option advocates. Another potential factor driving these results: opponents of the public option might not have a category to register their opposition here. Can they say "not at all important?" Perhaps, but does that accurately reflect their views? A lot of opponents of the public option think it is quite an important issue.

Here's the second version of the public option question, asked of subsample B.

NBC:WSJ 2.jpg

This one is less tilted to the Democratic side, although Republicans would still dispute the idea that a health care marketplace with a public option will actually generate competition. Still, the removal of the highly loaded phrase "(giving) people a choice" makes this less tilted overall - also, this time people have an opportunity to register support or opposition. And notice the big change. A majority of respondents are either uncertain or in opposition.

So, this is another apples-to-apples comparison. As with Rasmussen, NBC/WSJ finds that changes in question wording on the public option can produce big changes in the poll results.

-Jay Cost

Does the Public Want a Public Option?

Progressives in the blogosphere and the halls of Congress are pushing for the so-called "public option." One of their major arguments is that the public wants it.

But does it?

From a certain perspective, the public option polls very well. Let's look at some of the polls on this in the current RCP average of Obama's job approval, being careful to note question wording.

Here's ABC News/Washington Post:

ABC News:WaPo.jpg

Here's Marist:


Here's CBS News/New York Times:

CBS News.jpg

Here's CNN:


Case closed, right?

Not exactly.

In the aggregate, the polls present a very mixed picture. These numbers are good for reform efforts, but other numbers are bad. For instance:

-Respondents don't generally approve of the reform bills. In some polls, a majority disapproves.

- Respondents give mixed marks to Obama for his handling of the issue.

-Respondents strongly disapprove of the job Congress is doing with health care.

-Only a small portion of respondents believes they will actually be helped by the health care reform proposals.

-All in all, since the health care debate really heated up in July, Obama's job approval has dropped in the RCP Average from about 59% to 52%. His disapproval rating has gone from about 34% to 43%

How can we reconcile these gloomy numbers with the sunny results on the public option?

It might be due to the public's lack of information. I'm sure that the average polling respondent is paying some attention to the health care debate, but she is paying much less attention than political junkies. This will limit the amount of information she actually has in her mental filing cabinet. So, the crucial question is: even if she has absorbed some pro- and anti-reform arguments, does she have enough information to relate them to specific reform proposals? Color me skeptical on that one. I think your average respondent - even with some general opinions on reform - will have a hard time using those broad considerations to evaluate items like the individual mandate, guaranteed issue, community rating, and...wait for it!...the public option.

So, asking about specific proposals might be taking the conversation too far into the woods for the average respondent - and she is going to have a hard time recalling a relevant piece of information upon which to base a response. Instead, she might use the question itself as a basis for her answer. It follows that the information or perspective given in the question could make her more or less partial to the proposal under consideration.

And the pollsters are frequently providing information that is partial to the Democratic side of the ledger. As Kellyanne Conway argues:

Asking an under-informed public in a poll about "public option" is incomplete. It calls for a response to feel-good phraseology rather than a probing of underlying ideology. "Public option" in health care is not so different from "campaign finance reform," "Violence Against Women's Act," "revenue enhancements" or for that matter, "world peace' and "no rain this Saturday."

The pollsters are using plenty of "feel-good phraseology." ABC News/WaPo presents the idea that the government insurance plan would "compete" with private insurance plans. This is a contested notion, as Republicans think that the public option will drive private insurance away.

Marist uses the phrase "public option," which has become the conventional term for this insurance reform - but is nevertheless an intentionally constructed phrase designed to garner maximum public support. "Government-run health care" is foreboding, but "public option" is inviting.

CNN uses the phrases "public health insurance option" and "compete."

CBS News/NY Times specifically relates the public option to Medicare, a program that is so popular that Democrats are now thinking about reframing their pitch for the public option as merely an extension of Medicare to all. I wonder if they got that idea from CBS News/NY Times!

If the theory that question wording is playing a role is correct, then altering the wording should induce a change in the results. So, what happens when information less partial to the Democratic side is introduced? To start answering this question, let's consider the Gallup results, which are decidedly less bullish on the public option:


Like ABC News/WaPo, Gallup uses the Democratic buzzword "compete." However, Gallup also uses a Republican buzzword: "government-run." This is opposed to the weaker formulation - "government administered" - offered by CBS News/New York Times and CNN. With this more balanced choice of words, Gallup finds a roughly even split. I would not call this definitive evidence, but it suggests that we might be on the right track.

Let's take a look at Rasmussen. He has offered a series of really interesting questions on health care. First, he gives a basic version of the question that ABC News/WaPo, CBS News/NY Times, Marist, and CNN asked:

Would you favor or oppose the creation of a government-sponsored non-profit health insurance option that people could choose instead of a private health insurance plan?

That gets strong approval, as per usual when people hear words like "choose," "compete," and "option."

Then Rasmussen asks this follow up:

Suppose that the creation of a government-sponsored non-profit health insurance option encouraged companies to drop private health insurance coverage for their workers. Workers would then be covered by the government option. Would you favor or oppose the creation of a government-sponsored non-profit health insurance option if it encouraged companies to drop private health insurance coverage for their workers?

What happens when this Republican argument is substituted for the Democratic argument? Support for the public option plummets dramatically. Nearly 3/5ths of all respondents voiced opposition to the public option when it was phrased in this way.

Additionally, Rasmussen asked whether respondents thought the public option would save taxpayers money (they didn't), whether they thought it would offer better health insurance than private insurance (again, no), and whether people preferred to have a public option or a guarantee that nobody will lose their current coverage (the guarantee won in a landslide).

These results are very consequential. After all, Rasmussen is holding a lot of factors constant, enabling us to observe: same poll + same methodology + different frame for the question = different answer. That strongly suggests that the frame used for the public option question goes a long way in determining the answer the public gives.

So, does this mean that the public is actually against the public option? I'd say no. Instead, I would suggest that the public lacks sufficient information about that specific item to deliver a firm opinion. Accordingly, its opinion varies depending upon question wording, priming effects, the ebbs and flows of the news cycle, and so on.

-Jay Cost

Polling on Health Care Reform

Over at Pollster, Charles Franklin performs some fascinating analysis on public opinion on health care. He puts together a series of trend lines based upon different "smoothing" techniques, which cut down on statistical noise to varying degrees. Despite all these different methods, he still finds the same basic trendline:


This shows that the country is now about evenly divided on the various health care proposals working their way through Congress. Support for the bill dipped during the summer, but has risen to pull about even with opposition.

This is not a great result for proponents of the current reforms bills. The trick is that it has to pass through the United States House of Representatives and the Senate. It's an inferential fallacy to assume that because a bare majority of respondents support the proposals (supposing they do), a bare majority of members of Congress would, too.

To appreciate this, consider the following histogram. It outlines the distribution of Obama's share of the 2008 vote by congressional district.

Obama Vote by Congressional District.jpg

Obviously, congressional districts are far from uniform! The modal category here is actually soft support for McCain, where Obama won between 40% and 50% of the vote. Yet the political battle over health care will inevitably be fought in those districts that softly supported Obama. According to Franklin's analysis, health care reform is polling slightly under Obama's vote share in 2008. So, those districts where Obama won narrowly, not decisively, are probably where the main political battle will occur. It's reasonable to assume that if the nation is now evenly divided on the reform measures, those districts taken all together are divided, too. Many of them should be divided internally as well.

This highlights a core problem the Democrats have in the Congress. They win a lot of districts by blowout margins. This makes them safe for the party, but it means that their voters are packed into relatively few districts, suggesting that to pass large-scale policy reforms such as the one being debated now, the Democrats have to find support in districts where Republicans do well, even in bad years for the GOP like 2008.

This problem becomes all the more salient when we consider the practical playing field - namely, that the bills working their way through the Congress are unlikely to get any Republican support. If the Democrats plan to pass it all by themselves, there is going to be quite a bit of pressure on many members.

Obama Vote by Congressional District Democratic Districts Only.jpg

As we can see, there are a lot of Democrats in McCain-voting districts. So, if it is the case that the McCain voting districts are opposed to the health care bills, the Democrats are going to need at least a few representatives to vote against their constituents to get the bill through the House. That is a huge request to make, especially considering how salient this issue is. It's never a good idea to vote against your district on an issue that your constituents are paying close attention to.

Up to this point in the analysis, we've assumed that support/opposition to the bills mirrors the 2008 vote. It likely does not follow the 2008 vote perfectly, and there are probably at least a few notable deviations. The problem is, we just do not know how support breaks down by district. We lack reliable polling on this front. Importantly, many members probably lack such knowledge as well. Polls are expensive to contract and polling by congressional district is problematic. So, many members likely do not have a systematic read on their districts, the kind of knowledge that can be acquired via scientific surveying. They could, of course, rely on methodologically questionable analyses that "find" that certain reform measures are overwhelmingly popular, but I would not suggest that.

Instead, they have to rely on other metrics - like telephone calls, emails, attendance at town halls, and so on. This is why - even if the August town halls did not move public opinion against the bills - they were probably still quite consequential, as they gave members a sense of how their districts were feeling about the reform measures. Because turnout in congressional midterms is always less than presidential elections, even if those town hall outbursts represented a minority position in a district, it still cannot be taken lightly. After all, in a midterm election a member can be tossed from office by an opposition bloc that, during a presidential year, would constitute a minority.

While Senators have better access to polling, and therefore they probably have a more systematic perspective on their districts, the health care bills still face many of the same challenges in the Senate. This is the distribution of Obama's vote by state.

Obama Vote By State.jpg

Again, we can conclude here that the main locus of debate will be in states that went for Obama softly. And when we look at states with Democratic Senators - we see basically the same thing as we did when we looked at the House.

Obama State Dems.jpg

Again, for a bill to become law, Democrats are going to need some members from McCain states to support it, unless they can pull in some Republicans. This again suggests that an even split in support for health care is more of a hindrance than a help in getting a bill through.

-Jay Cost

Obama's Worst Poll Number

Gallup's breakdown of Obama's job approval by age was illuminating.


First off, note Obama's drop-off among young people. Young people were supposed to be a critical component of the new Democratic majority. Granted their approval is still slightly higher than the other groups, but it has far and away been the most volatile, dropping more than any other. This should not come as a huge surprise. Baby Boomers were partial to McGovern in 1972, but swung around to Reagan in the 80s. Young people's political dispositions are still being formed.

Yet, Obama's worst poll number here is actually his share among seniors. I'm guessing it relates to the health care debate. The White House should be very concerned, and for one simple reason: seniors vote.

Here are some empirics on that claim. I looked at states that featured hotly contested midterm Senate elections in 2006. I counted ten: Maryland, Missouri, Montana, New Jersey, New Mexico, Ohio, Pennsylvania, Rhode Island, Tennessee, and Virginia. For each of these, I pulled out the share of the electorate that was 65 and over for President in 2008, Senate in 2006, and President in 2004.


First off, there was not a noticeable drop-off among senior voters from 2004 to 2008. Only Ohio shows a significant change, and it has an increase. About half have a slight increase and half have a slight decrease. That's consistent with national polls, which have seniors contributing 16% of the total electorate in 2004 and 2008.

Second, notice 2006. In seven of the ten states, seniors accounted for a larger share of the electorate during the midterm. In several of them, the differences were substantial. At least in the hotly contested Senate elections, the 2006 electorate was noticeably older. This corresponds with national data as well. The national House exit poll in 2006 found 19% of the electorate was 65 or older, compared to 16% in the 2004 and 2008 presidential elections.

One reason for this might be that there is a lot of stimulation to vote in a presidential election - especially the last two matchups, which were hotly contested - but that stimulation drops off for the midterms. Thus, you're left with an electorate voting more out of habit, rather than being drawn to participate by the excitement of the spectacle. That could give seniors an advantage.

If Obama's numbers with seniors stay in the cellar, this could mean midterm problems next year for the Democrats. The silver lining here for the White House is that most of the drop-off occurred recently, which suggests that Obama might be able to win at least some of these people back. If he can improve his overall standing on the health care issue, he'll probably pick up with seniors.

-Jay Cost

More on the Recent Changes in Party Identification

Recently, Pew published an interesting graphic on historical party identification that enables us to continue our discussion of partisan affiliation through time. This is the picture that Pew presents.

Pew Partisanship.gif

It's important to note that prior to 1989, the party identification data is from Gallup. After 1989, it is from Pew. This is an interesting picture - and it generally squares with what we know about the partisan battles over the last 75 years. Importantly, it does not track ideology, which makes a huge difference. In particular, 1946 to 1964 is a period of Democratic dominance, but not necessarily liberal dominance - certainly not in the Congress, where a coalition of Republicans and conservative Southern Democrats was often able to thwart northern liberals. I've discussed this before, and my general sense of the post-war period is that it is best understood as one of ideological balance, with discrete, short-lived periods of liberal "breakthroughs," like 1964-1966.

Now, let's compare this data to the exit poll data we reviewed last week. We'll also include Gallup's partisanship data from 1989 forward. All of this is contained in the following chart:

Gallup, Pew and Exit Party ID, 1972-2008.jpg

There are three salient points about this graph:

Changes in Gallup and Pew Generally Track Exit Poll Changes. All three tick upwards during fat times for the Republican Party, most notably around Reagan's reelection in 1984. Lean times show a tick back downwards. This is as should be expected.

Gallup and Pew "Over-Dramatize" The Exit Poll. Slight changes in the exit polls tend to be more dramatic in the Gallup and Pew samples. This is especially apparent if we look at 1984-2004. The exit polls show only modest changes - with Republican self-identifiers ticking down to 35% from 36%, then up to 37% in 2004. Meanwhile, Gallup and Pew show a great deal more variability, each moving about 5% over the period. This is consistent with a point I made last week. GOP self-identification might have dropped 10+ point in recent media polls, but only 5 points in the last exit poll. This is not necessarily good news for the GOP, as fewer changes in exit poll party identification indicate that the actual electorate has a more stable partisan orientation. This means that ground lost might not easily be ground regained.

Gallup and Pew Consistently Underestimate Republican Identifiers Relative to the Exit Poll. Pew tends to be more pessimistic about the GOP's standing than Gallup, but both always show fewer Republicans than the exit polls. The difference is typically 5 to 7 points. What could account for this? I can think of two explanations.

(a) The exit polls are a snapshot of party identification on a single day, while the Gallup and Pew numbers are an average of the whole year. In 2008, both Gallup and Pew showed the GOP at its strongest point shortly before Election Day. Gallup generally showed the GOP stronger in the fall of 2004 than in the Spring or Summer of that year. [Unfortunately, I was unable to locate monthly or quarterly party identification numbers for prior presidential election years.] Why might this be? Some subset of "natural" Republican partisans might only return to their political home when the campaign begins in earnest, around Labor Day. If so, an annual average of party identification - or one that looks at out-years - might systematically underestimate GOP strength relative to where it is on Election Day.

(b) Non-voters are less likely to identify themselves as Republicans than voters, and they are included in the Gallup and Pew numbers. In fact, recent turnout - which is at its highest in some time - is still less than 60% of the voting age population, which means that about 40% of the Gallup and Pew samples in recent years should be non-voters. In a year like 1996, non-voters will constitute more than half of these samples. According to the National Election Study, non-voters are not as inclined to see themselves as Republicans as voters (on a five point partisanship scale: strong Democrat, weak Democrat, Independent, weak Republican, strong Republican). In fact, from 1972 to 2004, the average difference in Republican identification between non-voters and voters was fourteen points. This trend is muted on the Democratic side, as a good portion of non-voters are inclined to see themselves as "weak" Democrats.

I think the take home point from all of this is fairly clear. The Gallup, Pew, and other media pollsters tracking party identification offer data that is of real value - but it has to be interpreted with care. There are big, consistent differences between media polling data on partisanship throughout the year versus the Exit Poll, which is a better metric for partisanship on the day that it matters, Election Day.

Just as Gallup, Pew, and others "over-dramatize" changes in party identification - I think the recent meme on the decline and fall of the contemporary GOP has been "oversold." That's not to say that the party is not in a rough spot at the moment, but just that the analysis by many pundits is like a good steak that's been cooked just a bit too long.

-Jay Cost

Are American Voters Ideologically Polarized?

I've written recently on this blog about political polarization, noting several factors that point to its rise: (a) there has been an increase in ideological sorting among the parties, with conservatives being more closely identified with the Republican party and liberals more closely identified with the Democratic party; (b) partisan identification is a better predictor of vote choice than it was several decades ago; (c) partisan sentiment for the other side has been turning negative; (d) job approval of recent presidents has fallen more starkly along partisan lines in recent years.

The implication of these considerations is that the electorate is becoming more deeply polarized. However, we have to be careful with how far we take this idea, as recent scholarly work has shown that there might be limits to the polarization hypothesis. The evidence is mixed - and there are competing camps among political scientists. Some, notably Alan Abramowitz of Emory, have argued that polarization along issue and ideological lines has been on the rise. But others, notably Morris Fiorina of Stanford and Samuel Abrams of Harvard, have cautioned against this conclusion, asserting that while the electorate might be closely divided, it is not deeply divided.

Obviously, this is not the forum to arbitrate between these claims, and I am far from qualified to be the final judge. Instead, what I want to do today is highlight the big reasons reasons why some argue against ideological polarization in the mass public. This is the view to which I am partial - and ultimately I wish to show how this position is consistent with the polarization we have seen in voting and presidential approval.

First off, we have to stipulate the following: there is a broad consensus that political elites have become more polarized. Evidence for this abounds. Consider, for instance, the following chart. It uses congressional voting to track the ideology of the median legislator by party and region.

Ideology Since 1893.jpg

Positive scores (0 to 1) imply conservatism, negative scores (-1 to 0) imply liberalism. 0 is moderate, and I've highlighted the 0-line on both charts. So, the farther above zero you see the line rise, the more conservative the legislator is. The lower, the more liberal.

With the onset of the Great Depression, the Republican Party moderated across all four regions of the country. However, beginning around the Great Society, there was a shift to the right in all regions but the Northeast. Republican legislators in the Northeast moved rightward starting with Reagan. On the Democratic side, note that the party outside the South was quite liberal, especially after 1958. The South was the conservative faction for many decades; however, as African Americans finally won the right to vote free of suppression, Southern whites drifted to the GOP, and the "New South" began to attract different types of voters, the Southern part of the caucus has become more liberal, too. Today, thanks to what Vanderbilt's Marc Hetherington calls "the Big Sort," we have two highly polarized congressional parties: big inter-party differences and small intra-party differences.

Meanwhile, candidates for office tend to take polarized positions during the electoral campaign. Systematic evidence of issue-positioning is sparse - but the data that is out there shows that candidates for Congress typically adopt positions more in line with the party, rather than the middle of the electorate, which is where the classic economic theory of voting predicts candidates will converge. Many causes have been hypothesized: income inequality, redistricting, the increasing ideological nature of the congressional electorate, the increase of party power in the legislature, even the closeness of the party division in Congress has been offered as an explanation for legislative polarization. Regardless of the cause, there is consensus that elites have become more polarized.

But those are the elites. What about average voters? Have they become more polarized? Again, if we go only by their voting - they have. But that does not necessarily mean that they have become more ideologically polarized. Let's continue to use the same definition of polarization - tight clustering around two distinct ideological or issue-based poles. Has the public polarized in this fashion?

The critics assert that they have not. They make several arguments.

(1) It's easy to overstate the relationship between social groups and political groups. This has been a hot idea, what with the micro-targeting system for GOTV popularized by Karl Rove and Mark Penn. However, the relationship between the two is far from perfect - and it varies over time. For instance, white Christian evangelicals gave President Obama 26% of the vote in 2008. John Kerry won 21% of this group in 2004, which further indicates that this segment of the population is not uniform in its voting over time.

At issue here is whether social groupings imply political positions in the mass public. Even for a group like evangelical Christians, which is often assumed to have an almost determinative set of political positions based on social identification, the correlation is limited. So, for other groups - like gun owners, who gave Kerry 36% and Obama 37% of the vote - you're bound to find even weaker correlations between social group and political preference.

(2) Ideological self-identification has not changed much, with most voters still in the middle. The National Election Study has been asking respondents since 1972 where they place themselves on a 7-point ideological scale (from extremely liberal to extremely conservative). An analysis conducted by Fiorina and Abrams for the 2008 Annual Review of Political Science found little to no change in the percentage of each subset over the last 30 years. Respondents still cluster in the middle.

(3) The same goes for issue positions. The difficulty for measuring shifts in issue positions is that alterations in question wording can induce apparent changes in opinion, even if no real changes have actually occurred. Additionally, salient issues come and go, which makes it difficult to track slow-moving trends. So, researchers have to utilize a relatively small database of questions that have been asked again and again. Different scholars have drawn different conclusions using the same data. Abramowitz and Colorado State's Kyle Saunders have found increased polarization on issue positions - but, in response, Fiorina and Abrams argue that the American electorate is, "a largely centrist public drifting slightly rightward on some issues, slightly leftward on others, but with only very small declines...in the number of moderates."

(4) There has been evidence of party sorting, but its extent is uncertain. Party sorting is where partisans adopt issue positions consistent with their party's platform. Party sorting has unequivocally occurred over the last few decades - especially among better-informed voters - and there is a strong consensus that it is a consequence of cues from polarized political elites.

However, the core debate is not over if, but how much. Is this sorting limited to the core activist base of both parties, or a broader segment of the population? Results are mixed. For instance, Pew has asked voters their positions on a set of issues for 20 years - between 1987 and 2007 - and the average difference between Republican and Democratic respondents has been slight. Relatedly, the 2004 National Election Study found that 33% of strong Democrats and 41% of strong Republicans are out of step with their party on abortion. That's a high number for an issue where party sorting is thought to be quite pronounced. However, other metrics - including those offered by Abramowitz and Saunders - have found greater sorting. Hetherington recently argued that 40% of the population exhibits "deep party sorting" with another 35% exhibiting at least some. Plus, one's degree of sorting strongly and positively correlates to one's degree of political knowledge, a common finding since the 1960s. All told, the critics accept the idea of party sorting, at least to some degree, but caution against concluding that the mass public is nearly as sorted as the party bases.

So, having reviewed these objections, we can ask: does this non-polarization view square with the highly polarized results we have seen recently for voting and job approval?

Here, it becomes highly consequential that elites are polarized. We have to remember that a vote in our two-party, candidate-centered system is a binary choice between two individuals, whereas a political preference is a multi-faceted opinion about an issue. This can make all the difference in the world. A moderate voter who must choose between two extreme candidates will inevitably make an "extreme" vote choice, if only because there is no moderate in the race. If all we had to go on was his vote choice - we might conclude that he, too, is an extremist, when in fact he is not.

The same goes for job approval. Again, to quote Fiorina and Abrams: "People express approval or disapproval of the president's performance not simply by looking at their own positions, but by comparing what the president has done with what they would have liked him to do." This means that a president who is behaving in an extreme way will have a polarized job approval rating, even if there is little ideological or issue-based polarization in the mass public.

As I noted earlier - there are differences of opinion on this subject. I'm partial to the views expressed principally by Fiorina and Abrams. But Abramowitz and Saunders, on the other hand, are top-notch scholars who disagree. So bear that in mind.

Writing this year in the British Journal of Political Science, Hetherington offers his take:

Contemporary American politics is probably best described as polarized on the elite level and increasingly well sorted in the electorate. In the 109th Congress, Republicans and Democrats in the House achieved complete ideological separating, and the distance between the average Republican and Democratic member reached its highest point in nearly a hundred years. On the mass level, the ideological distance between partisans is now larger, but their attitudes are not clustering towards the poles even if their evaluations of specific polarizing political leaders are often very far apart.

I think this is a sound conclusion.

-Jay Cost

The Fight Over the Economy Is Just Beginning

In my recent discussion with Ruy Teixeira, I argued that true ideologues constitute a relatively small percentage of the public. But that is not to say that the broad middle of the nation does not have a core set of values that guides its political decisions. Among other things, it believes firmly in the idea of economic growth, and it isn't hesitant to punish politicians for weak economies.

The relationship between the electorate and the politicians is akin to Darth Vader and his lieutenants in The Empire Strikes Back. When the underlings failed Vader, he impatiently struck them down without a second thought, moving on to the next in command. Similarly, when politicians fail to deliver growth, the judgment of the electorate is just as swift and almost as brutal.

A Gallup poll conducted in 1999 found that 71% of the country approved of George H.W. Bush's job as president. Yet Mr. Bush had the misfortune of presiding over a downswing in the business cycle. Though the economy had been growing for six straight quarters by Election Day, unemployment was above 7%. He won just 37% of the vote. That 1990/91 recession also hurt his successor. In the early Clinton years, the economy grew and unemployment fell, but growth in real per capita income was slow to rebound. By the midterm, just 43% of voters approved of Clinton's handling of the economy, and the Democrats lost 52 House seats.

How's that for brutality? One (relatively mild) recession, and the public delivers harsh punishments to both parties years after growth returned. "Apology accepted, Captain Needa."

There are three lessons for today's politics. First, the country is impatient about growth. Recessions are virtually immoral in this country - and if growth is slow to return, or if its effects are slow to be felt by the average voter, the public will not take it lightly. The top line GDP number is not enough. If other indicators - like unemployment and real income, metrics that speak to how people are experiencing the economy - are still weak, the public's response can be just as wrathful.

Second, the public's diagnosis of the economic problem need not be enlightened. Imagine you lost your keys on a dark street. You'll look for them under the nearest streetlight - not because that's where they are, but because that's where you can see. That's how the electorate makes judgments about complicated subjects like the economy. It focuses on what it understands, whether or not that gets to the real issues. Recall the political damage George H.W. Bush suffered because he hadn't seen a price scanner before. Somehow, this meant he was out of touch, and thus not suited to bring the economy to recovery.

Third, Walter Shaprio recently suggested that Republicans will not gain from any populist backlash. I wouldn't be so sure. Out parties can make substantial, recession-related midterm gains despite having been led by unpopular presidents. Perhaps the best example is 1938. Amidst the "Roosevelt Recession," the country turned to the party of the reviled Herbert Hoover, who still had a negative rating in 1944. FDR's majority in the subsequent Congress depended entirely upon the old Confederacy - meaning that the GOP was the country's first choice outside the one-party South.

This links into the second point. The public lacks economic expertise, yet it must still assign blame for the struggling economy. It is unsurprising that - regardless of whether he deserves it - the President is often the recipient. After all, he is the most visible politician in the country. Additionally, Presidents are quick to accept credit for a flourishing economy, so inevitably they take the blame for when it languishes. When you blame the President and want a change, the opposition party is the only viable option.

While the current focus on Timothy Geithner, the Treasury, and the financial markets is understandable - this will probably not be the script of the broader political battle over the next 20 months. Assuming that the financial system is brought under control, the political debate will focus relentlessly on recession and recovery. Though the Administration, the CBO and the Blue Chip forecasters project modest growth in 2010 (ranging from 1.9% to 3.0%), all of them expect high unemployment (7.9% to 9.1%) and an economy performing below peak capacity. If these predictions are true - the corresponding public dissatisfaction will define the campaign of 2010, and the legislative battles that precede it.

Both sides will struggle to pin blame for the weak economy on the other. Republicans will indict President Obama, arguing that his policies failed to improve things. President Obama will remind voters of the previous administration, arguing that congressional Republicans advocate the same policies that brought about the recession. The public lacks the technical expertise to arbitrate based on the merits - so the outcome will depend in part on how bad the economy actually is (the worse it is, the worse for President Obama), and which side shows the greatest political acumen.

If you find this to be a dispiriting commentary on democratic accountability, think of it this way. Electoral justice might be rough, but it's also consistent: bad economies mean electoral defeat for somebody. Thus, those who are still in office when the dust settles learn a valuable lesson: grow the economy, or next time it could be you. In the long run, the public gets what it wants - a government dedicated first and foremost to growth.

-Jay Cost

A Note on the "New Progressive America"

My thanks to Ruy Teixeira for commenting on my response to his report. Seeing as how he has put up the overwhelming share of the effort here, it is appropriate to allow him the last word on the substance of our debate.

I do want to offer a clarification on my broader position, which was outlined in detail with Sean Trende here. Teixeira writes:

Cost may choose to believe it's of no real significance that growing demographic groups and areas of the country are strengthening progressives, while conservatives are holding their own only where America is stagnant and declining. He may even be able to convince conservatives that he is right. In doing so, however, is he doing them a disservice? The longer conservatives believe nothing has really changed, the longer they will resist doing what they need to do: change their positions, soften their ideology and move toward the new progressive center of American politics. After all, it may be a progressive center, but it's still the center. And eventually conservatives are going to have to deal with that.

My position is not that "nothing has really changed," and by implication that the country is still a center-right country. Similarly, I would not counsel conservatives to do nothing.

Instead, my position is this: American electoral politics is full of change. That is the principal reason I am generally skeptical of arguments about enduring majorities - left or right. There's just so much change in our electoral politics that an enduring majority is a highly problematic category.

I draw this conclusion based on five observations about the last forty years:

(1) The parties typically share control of the institutions of government.

(2) The longest time any party has held the presidency and the Congress together was four years. In both instances, subsequent electoral defeats were decisive and humiliating.

(3) Voters typically have low levels of political information, which makes it difficult to develop genuine ideology.

(4) The electorate as a whole is known to vary its preference on non-ideological factors, like the performance of the economy and the President's management of the government.

(5) The purpose of the political party is to acquire the majority. Historically, this has meant that the parties respond to setbacks, often by redrafting elements of their message.

I would not argue that there are not stable features in our electoral politics. Most of the electorate is partisan, so it stays put from cycle to cycle. This also does not mean that there have not been realigning features - as one group switches from one side to the other and stays there. Urban ethnics did in 1928/32 and Southern whites have been moving since 1948. My initial response noted the shift of white Catholics and the white working class. So, long term forces matter.

The conclusion I draw from these observations is that while there is stability and realignment, we need to remember that non-realigning features account for much of the cycle-to-cycle swings in the balance of power, which never take us very far from 50/50 for very long. Too many analysts forget that, and explain the most recent electoral results via realignment. That is why, for most elections in the past, you can find somebody somewhere who claimed it was somehow realigning.

Generally, I think realignment is one of those concepts that over-promises and under-delivers. It might account for the political dynamics from the Civil War to the Great Depression, but I do not think it explains nearly as much about the contemporary scene. So, I do not buy Teixeira's "new progressive America" - but for the same reasons I didn't buy the arguments for a "new conservative America" four years ago. If some conservatives find solace in my argument today, it is accidental. Four years ago, progressives might have found it comforting.

Similarly, these observations imply a clear suggestion to Republicans: innovate, innovate, innovate! Conservatives who presumed that previous victories implied a long-term majority were simply wrong. Political fortunes have swung back and forth over the years in part because political parties, once removed from office, work hard to get back into the majority. This is what Democrats were busy doing while some Republicans were celebrating their "enduring majority" (How many times did they trumpet their victories in 97 of the 100 fastest-growing counties?). Now, it's time for Republicans to get to work.

What should they do? That's outside my area of expertise. I do know - and Sean Trende and I pointed out here and here - that, as Teixeira said, Hispanics moved to the Democratic Party between 2004 and 2008, and this was a contributing factor to the party's losses in several states. Far from asserting that this is of "no real significance," I would suggest that this is a great place to start.

-Jay Cost

New Progressive America? A Response to Ruy Teixeira

Ruy Teixeira has published a new report at the Center for American Progress (CAP), entitled "New Progressive America: Twenty Years of Demographic, Geographic, and Attitudinal Changes Across the Country Herald a New Progressive Majority."

Here is the gist of Teixeira's argument:

Obama's 53 percent of the popular vote is the largest share...received by any presidential candidate...[since] George H.W. Bush...So, separated by 20 years, we have two elections that are practically mirror images of one another...

How did conservatives do so well in one election but progressives so well in the other? The answer: In those intervening 20 years, a new progressive America has emerged with a new demography, a new geography, and a new agenda.

Let me say at the outset that I will not argue that Teixeira's thesis is wrong. Instead, my position is that it is underdetermined: the facts support alternative conclusions not consistent with his assertion.

Generally, I approach arguments like Teixeria's with a high burden of proof. Electoral history over the last forty years indicates a norm of divided government in which both parties share control. Furthermore, for the years when there was unified party government - the majority party lost power relatively quickly in a decisive, broad-based defeat.

Thus, even well constructed arguments about enduring majorities are going to have a tough time convincing me. Unfortunately, Teixeira's argument has problems in its execution. I count three, significant difficulties.

First. Teixeira makes use of long-term estimates of population change to argue for a "new demography" that will "reshape our country in a fundamentally progressive direction." He cites projections in growth among Hispanics, Asians, and African Americans so that by 2050, "the country will be 54 percent minority." It is outside my technical expertise to dispute those estimates - though I am generally skeptical of predictions about anything that is such a long way off.

I would dispute the political implications of these demographic projections. Teixiera's argument about future political demography assumes a static quality to American politics that is ahistorical.

For instance, consider that while John McCain lost the nationwide popular vote by seven points, he won the white Catholic vote by five points. From a historical perspective, this is remarkable. John Kennedy won 81% of non-Hispanic white Catholics, Lyndon Johnson 79%, and Hubert Humphrey (who lost in a three-way race) still won 55%. Forty years ago, any liberal analyst would have concluded that the white Catholic vote belongs to the Democrats. Yet today, we see the GOP holding white Catholics amidst a popular vote wipe out.

Similarly, who would have ever thought that the "white working class" - the backbone of the New Deal coalition for decades - would support the Republicans by 18-points as the nation supported the Democrats by 7? That is the most dramatic proof that voting coalitions are not static - and that we cannot extrapolate future alignments from current ones.

Electoral politics is not akin to Newtonian physics, where you derive your equations and then predict everything from here to eternity. Instead it's unpredictable. Why? One reason is the parties. They select issue positions and emphases to steal the other side's wavering voters and undermine its voting coalition. Again, recent electoral history has demonstrated that both parties are quite adept at this game. In light of that, how can we know whom Hispanics, Asians, "professionals," young voters, or anybody will support in 2048? I'd suggest we cannot. Using demographic estimates to predict long-range political preferences is an impossibly difficult task.

Second. Teixeira cites election data from the last 20 years without introducing the appropriate context. For instance, he writes:

[P]rogressives have been gaining strength among white college graduates. In 2008, Obama only lost white college graduates by four points, compared to an 11-point deficit for Kerry in 2004 and a 20-point deficit for Dukakis in 1988.

I do not dispute these numbers, but I disagree that this is evidence of a "new progressive America." There are many reasons voting results change from cycle to cycle. Though the concept has been vastly overworked, some of that change is due to "realignment." However, much of it is clearly not - and if the appropriate electoral context is excluded, you're bound to overestimate the extent to which realignment is a factor.

Return to the previous quotation. Democrats have gained 16 points in 20 years among white college graduates. Is this realignment, or the ebb and flow of non-realigning factors? I'd suggest the latter explains much of the difference. Consider the following chart. It tracks the average growth in the economy in the third and fourth quarters, the incumbent president's job approval in the pre-election Gallup poll, and the incumbent party's share of the presidential vote.

National Conditions and Incumbent Party Performance in Presidential Elections.jpg

Clearly, non-realigning factors play an extremely powerful role. This is why political scientists can build accurate predictive models of presidential elections based on a few simple variables.

Without this context, it is prohibitively difficult to tease out realigning factors between 1988 and 2008. The former is a year in which the economy was growing and the Republican President was popular. 2008 reflects the inverse. This makes it hard to conclude that realignment is accounting for the change among college educated voters, or any subgroup. I'm open to the idea that realignment is at work, because I think to an extent it is - but that means you must control for the non-realignment factors known to influence presidential elections. Teixeira does not do that, which means his statistics do not support his conclusion.

We can see this problem again when Teixeira argues about the increasing liberal tilt of the "growing areas of the country." He writes:

By and large, progressives received their strongest increases in support in the fast-growing, dynamic metropolitan areas of states, particularly the largest ones...The result is a political map with a distinct lean toward progressives, a lean that should increase in coming years.

Again, maybe. But is it not also plausible that the growing areas of the country were especially affected by the sudden, dramatic contraction in the economy that was occurring on Election Day, and that they responded (as the country as a whole typically has since 1840) by swinging to the out party in especially large numbers?

In other words, without the necessary context, every increase in the Democratic Party's share of the vote last cycle becomes a sign of growing liberal strength. This is an underdetermined inference.

Third. Teixeira argues that "the American people's views on what government can and should do" are changing. Again, I am not going to disagree with this per se. Instead, I'll suggest that the evidence Teixeira cites is unpersuasive.

Specifically, he references a series of polling questions (many done by CAP) whose wording is so vague it is impossible to infer any political implications. Here is a sample:

(1) "Religious faith should focus more on promoting tolerance, social justice, and peace in society, and less on opposing abortion or gay rights."

(2) "A positive image of America around the world is necessary to achieve our national security goals."

(3) "It is the responsibility of the government to take care of people who can't take care of themselves."

(4) "Government regulations are necessary to keep businesses in check and protect workers and consumers."

(5) The government should "invest in alternative energy like wind, solar and bio-fuels to create jobs, and reduce dependence on foreign oil."

Unsurprisingly, these questions elicited a high degree of public support, sometimes up to 75% in the CAP survey.

These sorts of questions would indeed damage the Republican Party...if it were a silly, oversimplified caricature of what it actually is. The problem is that these questions present a false choice, implying that one side is in favor of these items, and the other side opposed. Not really true - the actual political divides are much more subtle. The fact of the matter is that many Republicans are at least partially sympathetic to all of these assertions. Certainly, no mainstream Republican politician would suggest, for instance, that Social Security disability payments should be eliminated, or that workplace safety regulations should be done away with. I recall that John McCain talked frequently about biofuels on the campaign trail last cycle, and one major goal of the surge was to gain the trust of local Iraqi populations.

These questions do not directly capture the salient political cleavages in the country. Thus, they cannot be taken as evidence of a "new progressive America." Certainly, they hint at such cleavages, but they are worded in such a non-specific, general way that they do not reflect the actual debate, which is why they can garner such broad support.

Of course, there are answers to several questions that Teixeira cites that genuinely show liberal leanings on actual policy matters (like universal health care). Then again, as Teixeira notes, the CAP Study also found decidedly conservative responses to questions like this:

(1) "Limited government is always better than big government." (55% agree)

(2) "Free market solutions are better than government at creating jobs and economic growth." (57% agree)

(3) "Government spending is almost always wasteful and inefficient." (61% agree)

(4) "Government programs for the poor undermine individual initiative and responsibility." (48% agree, 21% neutral)

(5) "Social Security should be reformed to allow workers to invest some of their contributions in individual accounts." (57% agree)

Liberals can argue that there is an incipient liberalism in the data, and conservatives can argue that there is an incipient conservatism. However, I'd suggest another idea: the country is in many respects non-ideological.

One characteristic of ideology is issue constraint. An ideological thinker decides his positions based upon abstract principles, which therefore constrain the policy options he can endorse. For instance, an ideologue who thinks that the free market is better than the government would be constrained on his views of health care. To be ideological, he must oppose government managed care.

This data - as well as many surveys - suggests that a large segment of the country does not have such constraints on their views. The CAP study finds 55% of the public advocating limited government, and 65% advocating that the "government guarantee affordable health coverage for every American." There is substantial overlap here, which indicates that some subset of the population is answering in inconsistent ways, which implies non-ideological thinking.

In conclusion, I'll recapitulate my main points. Generally speaking, I am skeptical of arguments like this - even those that are well executed. Unfortunately, Teixeira's argument has problems in its execution. He uses oversimple assumptions about the links between future demographics and politics; he fails to separate realigning forces from the normal back-and-forth of electoral politics; and he relies on inconclusive polling data to argue that there has been an ideological shift in public opinion.

This is not to say that Teixeira is wrong. He may be right, and we may be entering 10, 20, even 40 years of a "new progressive America." My point is that he fails to make the case. Other theories - such as one that predicts continued, heated party competition amidst a substantially non-ideological public - are just as consistent with the evidence he cites.

-Jay Cost

How Much Does Rush Limbaugh Matter?

With Rush Limbaugh's CPAC speech, the radio talker is back in the news - and some conservatives are wondering whether he is good for the party. Rod Dreher says that Limbaugh's speech was "political crack" and that:

Anybody who challenges Limbavian orthodoxy is, ipso facto, the Enemy. If you suggest reform, even from the Right, you are a useful idiot for the Media, which are the Enemy, and can never be anything but the Enemy. Limbaughism sounds a lot like Leninism.

Meanwhile, Reihan Salam says that Limbaugh is trying to, "remake the Party of Lincoln as the Party of Limbaugh." He goes on to suggest that Limbaugh is over-stepping his boundaries:

What Limbaugh fails to understand is that any successful political movement is built of both true believers and evangelizers. True believers, like Limbaugh and Sean Hannity, fire up the troops. They tell their followers exactly what they want to hear, and they instinctively resist any compromise of their hallowed principles. As a general rule, true believers live and work and worship among other true believers, and they like it that way. [Snip] Every week Rush Limbaugh reaches an audience of over 13 million listeners--a staggering sum by any standard. Yet 13 million listeners plus their spouses, plus the family dog, plus a few dead aunts and uncles thrown in here or there, still doesn't add up to an electoral majority.

That's where the evangelizers come in. Evangelizers are in the business of making converts, and so they are obligated to make their way among people who are opposed--sometimes bitterly opposed--to their views. To succeed, evangelizers need to recognize the other side's strengths and to use its language. Just as missionaries would occasionally "go native" in foreign lands and abandon their original creed, there is a real risk that evangelizers will lose touch with their core beliefs. Yet other missionaries learned to adapt, to take the essentials of their faith and compromise it in such a way as to make it relevant and compelling to the locals.

Are we talking about politics or religion? It sounds an awful lot like religion - and I think we need to reframe the discussion.

Political participation is a mass phenomenon in this country. Aside from going to church and watching the Super Bowl, it might just be the only other activity that commands so many participants. After all, at least 131,370,793 Americans voted in the last presidential election. In comparison to that, just about everything else is niche entertainment.

Obviously, Salam is correct that Limbaugh's audience would be insufficient for a political majority, but a political majority in 2008 required the support of nearly 66 million voters. Ultimately, that undermines his broader point - for we are dealing with a scale so massive that we have little use for "evangelizers."

There are reasons political campaigns have taken the peculiar shape they have taken - that they do not conform to the ideal of "deliberative democracy." We can appreciate just how entrenched our less-than-erudite tradition of politicking is by reviewing one of the first elections that generated high levels of participation, the election of 1840 in which William Henry Harrison squared off against Martin van Buren. From historian Paul Boller:

The log-cabin-hard-cider campaign (of the Whigs) had to be seen to be believed. There was no dearth of spectators. Estimates of crowds assembled for Whig rallies ranged from one thousand to one hundred thousand...Hard cider was plentiful...Slogans, mottoes, nicknames, and catchwords abounded: "The Farmer's President"; "The Hero of Tippecanoe"; "Harrison, Two Dollars a Day and Roast Beef"; and, best of all..."Tippecanoe and Tyler Too!" There were also scores of log-cabin newspapers, log-cabin pamphlets and leaflets; and thousands of Tippecanoe badges, Tippecanoe handkerchiefs, and Tippecanoe products (including shaving cream) of all kinds. Whig songs...were energetic, exuberant, ecstatic, and endless.

Basically, the Whigs campaigned in 1840 by plying the public with booze, music, slogans, and merchandise. Electioneering has, of course, changed over the years - but this reminds me quite a bit of Obama's acceptance speech in Denver!

This is the sort of thing you need to do to engage the mass public. There's little need for "evangelizers" in this scheme, then or now. When appealing to a political audience as broad as the voting public, you are confronting a large majority of voters who pay relatively little attention and are essentially non-ideological in their political orientation. That means the idea of converting somebody from "liberalism" to "conservatism" as a precursor to getting his vote is simply not going to yield many votes. If it did, this is what candidates - who have the greatest interest in winning votes - would try to do. Instead, they speak in sound bytes and they have Stevie Wonder or Hank Williams, Jr. open their political rallies.

You're going to win these marginal voters in part by finding issue positions that solve practical problems they face - hence the reason candidates take so many positions these days. Marketing is a factor, too. But the biggest factor is whether the winds of public opinion favor you. That should remind us of the scale we're dealing with: when we talk about public opinion, we often use nature metaphors, which imply that it affects politicians, but politicians do not really affect it. Ultimately, we can make a pretty convincing argument that campaigns do not really alter the public's thinking - that the great mass of the public has preexisting opinions on the parties and the state of the union, that these translate pretty easily into vote choices, and the hundreds of millions of dollars both sides spends do little to alter the process. That's the scale we're dealing with.

I think that generates two conclusions for the subject at hand. First, from the perspective of electoral politics, Rush Limbaugh is not much of a factor. That's not to say he is unimportant in other ways. He influences lots of people and is certainly important from a cultural perspective. But we're talking about elections - where more than a hundred million people participate. That has to change our evaluation of his influence. Additionally, he might be a hot topic for a few news cycles, but news cycles are drops in the bucket from an electoral perspective.

Second, there is value in the discussion among conservatives about the future of their movement. But that does not mean that the payoff is going to be electoral. This is a discussion by political elites for elites. Electoral politics - at least the difference between winning and losing - is inevitably non-ideological and non-elite.

Think of it this way. Suppose the Republican Party and the conservative movement fail to "reform" or "reimagine" themselves, but the country becomes highly dissatisfied with the governance of President Obama. What happens in 2010? I'll bet the farm that the GOP makes big gains in the House, ideological anemia aside. Now, suppose that the party and the movement do reinvigorate themselves, translate their principles into compelling policy solutions and generally begin an intellectual renaissance on the right - but the country is pleased with Obama and the Democrats. What happens? Again, I'll bet the farm that the Republicans make little or no gains.

When you get right down to it, elections are fought over the state of the union and the country's opinion on how the majority party has managed the government. The parties get to tinker at the margins, and ideology can be a part of this tinkering, but it's important not to make too much of it.

-Jay Cost

More on the Polls

On Friday, I noted that the differences among the national polls is large enough to suspect that something other than random variation is causing the disagreements.

I'd like to expand on this point by examining today's Pew poll, which pegs McCain's share of the vote at 38%, with a margin of error of 3.5%. That means that Pew predicts with 95% confidence that McCain's true share of the vote is somewhere between 34.5% and 41.5%.

While we don't know McCain's true share of the vote, we do have an estimate of it - the RCP average. Right now, it puts McCain at 43.6%. This figure is far outside Pew's 95% confidence range. So, if we use the RCP average as our estimate of McCain's true share of the vote, we would conclude that Pew is an outlier.

The question then becomes whether it is outlying due to random variation, or some non-random cause. We can never know for sure, but we can make a few points.

First, the level of disagreement between the Pew poll and the RCP average is great. Indeed, if we assume that the Pew poll has an accurate read on the electorate, the chance that McCain's true share of the vote is 43.6% is less than 0.5%. Given the number of polls that cycle in and out of the RCP average, we should expect at least a few outliers. However, it would be pretty rare to find one that disagrees with the RCP average by such a large amount.

Second, the previous Pew poll, which had McCain at 39% of the vote, was also an outlier when compared against the RCP average. So, Pew has twice in a row pegged McCain's number at significantly less than the RCP average. It is very unlikely to see this kind of result if random variation is the only cause.

Does this mean that Pew is wrong? No. We could only conclude that Pew is wrong if we know McCain's true share of the vote right now. We don't know that. Instead, what we can conclude is that the difference between Pew and the RCP average is likely produced by something other than random variation.

Pew is not the only poll behaving in this fashion. Today, the Gallup traditional model pegs McCain's number significantly higher than the RCP average. It has done this several times over the last three weeks - and every day since it began it has shown McCain doing better than the RCP average. It is unlikely that random variation would produce these effects. Today's Rasmussen poll shows McCain significantly higher than the RCP average, and it has consistently been higher than the RCP average for the last three weeks. IBD/TIPP frequently pegs Obama's number significantly lower than the RCP average, and it has shown him lower than the RCP average every day since it began. The GWU/Battleground poll has shown McCain consistently higher than the RCP average for 10 of the last 10 release dates, frequently at significant levels.

None of this is consistent with what we would expect from random statistical variation. These considerations reinforce the point I made on Friday. In all likelihood, something else is going on here. The pollsters have different "visions" of what the electorate is, and these visions are inducing such divergent results.

This is why I would urge caution when interpreting all this polling data. We're talking about disagreements among good pollsters. I take all of these firms seriously whenever they produce new numbers. They are disagreeing with one another in ways that can't be chalked up to statistical "noise." That gives me great pause.

-Jay Cost

A Note on the Polls

I've received several emails from people asking about the polls. The national polls do seem pretty variable, so I thought I would toss in my two cents on them.

First, we need a short primer on basic statistics. Real Clear Politics offers an unweighted average, or mean, of the polls. As long as there is more than one poll in the average, we can also calculate the standard deviation, which is one of the most important concepts in inferential statistics. The standard deviation simply tells us how much the polls are disagreeing with one another.

For instance, suppose we are testing the strength of Candidate A. We have 32 polls, which we can arrange graphically in what is called a histogram. Our horizontal axis shows the electoral strength of Candidate A. Our vertical axis shows how many polls we found with Candidate A pulling in that much of the vote. Let's say it looks like this:

Scenario 1.jpg

The average is 50%. The standard deviation is 1.6, which basically implies that the typical distance between a given poll and the average is 1.6. That's a pretty small number, and it squares with how concentrated the polls are around this average.

Now, suppose we have a distribution that looks like this.

Scenario 2.jpg

We get the same average, 50%. However, this time the observations are more dispersed around it. Here, the standard deviation is 3.0. That's higher, and we can see why. The individual polls vary more with one another. That's what the standard deviation shows us - how much the polls vary around the average.

A final point to get us ready. We might examine the spread between the two candidates - Obama is up 7 versus up 1 or what have you. This is certainly a valuable number to look at. Indeed, that's what we all care about! However, I am going to look at a candidate's share of the vote - not the spread. Ultimately, our analysis is going to rely upon each poll's reported margin of error. Those numbers do not refer to the spreads, but to each candidate's individual numbers. So, horse race polls actually have two margins of error - one for each candidate. Because the spread is the difference between them, it will be more variable than either candidate's individual numbers.

With this stuff in mind, let's focus on some hard numbers. As of this writing, Barack Obama's share of the vote in the RCP average is 50.3%. His standard deviation is 2.7. For McCain, whose average is 42.5%, the standard deviation is 2.3. For comparative purposes, I looked at the polls RCP was using from its 2004 averages. For roughly the same time in that cycle (10/17/04 to 10/24/04) Bush's standard deviation was 1.8; Kerry's was 1.7. This means that there is more disagreement among pollsters now than there was in 2004.

We can push this analysis further if we examine the distribution of each candidate's poll position. We'll first create a histogram of Obama's polling.


As we can see, most of the values cluster around the 49-54 range. However, there is a "tail" on the left-hand side. That's called a negative skew. That's a bit surprising. It's different from what we had in our stylized pictures for Candidate A.

Now, let's examine distribution of McCain's support.


There's no tail here, but the picture is still somewhat surprising. They are spread out fairly evenly across a broad range of values, with little clustering in the center.

Of course, a visual inspection can only take us so far. When we have only a few observations - and here we "only" have 15 - the true shape of the picture might not be clear. If we were to add another 5 or so polls, we might see something more like those stylized pictures presented above.

So, let's push the analysis a little bit further by looking at specific polls. We can test to see if the polls are separated from the average by a statistically significant amount. Again, since we're dealing with each candidate's individual poll positions - we'll test each candidate's number in an individual poll against the RCP average. To make sure we dot all our "i's" and cross all our "t's," we'll supplement the RCP average with a weighted average of the polls, which takes into account the number of observations when averaging the polls together.

Of the fifteen polls in the RCP average, four fall significantly outside the average for Obama and five do so for McCain. Meanwhile, three polls are right at the boundary of significance (one for Obama, two for McCain). The rules of statistics being what they are, we should expect a few polls here or there to fall outside the average by a statistically significant amount. But this is a lot. 40% of all our tests produced results around or outside the acceptable range.

So, we have made three observations: (a) relative to 2004, the standard deviation for Obama and McCain's polls are high, indicating more disagreement among pollsters at a similar point in this cycle; (b) the shape of the distribution of each candidate's poll position is not what we might expect; (c) multiple polls are separated from the RCP average by statistically significant differences.

Combined, these considerations suggest that this variation cannot be chalked up to typical statistical "noise." Instead, it is more likely that pollsters are disagreeing with each other in their sampling methodologies. In other words, different pollsters have different "visions" of what the electorate will look like on November 4th, and these visions are affecting their results.

Think of it this way. Suppose there is a bag of 130 million red and blue marbles that all the pollsters are sampling from. One pollster will pull a sample of 750 marbles, another a sample of 2,500, and so on. Oftentimes, they are going to pull different results from the bag. One pollster might pull 53% blue, another might pull 52%, and so on. However, as long as they are all pulling marbles from the same bag, the results will probably not differ too wildly. And after enough time, the distribution of those pulls should look something like those idealized pictures of Candidate A.

However, what if each pollster had a slightly different bag s/he was pulling from? In that situation, we should find more divergent results. That's basically what I'm suggesting here - that the bags the pollsters are pulling from are different. That's producing some of these larger-than-expected variations.

Now, I want to be clear: I am not making any claims about which pollster has the better sample of the electorate. I'm not singling anybody out for being right or wrong because frankly I do not know. I'm just pointing out that there seems to be disagreements among them that cannot be explained by random variation.

Importantly, there is one thing that the polls do not disagree on, the fact that Obama has a lead. All the polls show that. Also, we might begin to see convergence here soon. If pollsters have different methods for predicting what the electorate will look like, those methods might produce similar-looking "electorates" by the time we get to Election Day. At least for now, though, there is disagreement - not about who has the lead, but about how big that lead is.

-Jay Cost

On Gallup's Two Likely Voter Models

There have been reams of paper dedicated to reporting on the Obama campaign's voter mobilization efforts. This is what the Washington Post wrote on Sunday:

In 2004, Democrats watched as any chance of defeating President Bush slipped away in a wave of Republican turnout that exceeded even the goal-beating numbers that their own side had produced.

Four years later, Sen. Barack Obama's campaign intends to avoid a repeat by building an organization modeled in part on what Karl Rove used to engineer Bush's victory: a heavy reliance on local volunteers to pitch to their own neighbors, micro-targeting techniques to identify persuadable independents and Republicans using consumer data, and a focus on exurban and rural areas.

But in scale and ambition, the Obama organization goes beyond even what Rove built. The campaign has used its record-breaking fundraising to open more than 700 offices in more than a dozen battleground states, pay several thousand organizers and manage tens of thousands more volunteers.

What effect will this massive effort have at the ballot box?

Don't ask Gallup. On Sunday the polling outfit began offering its likely voter (LV) model (in addition to its registered voter (RV) model). But this year, there's a twist. Gallup is offering two LV models.

Obama's current advantage is slightly less when estimating the preferences of likely voters, which Gallup will begin reporting on a regular basis between now and the election. Gallup is providing two likely voter estimates to take into account different turnout scenarios.

The first likely voter model is based on Gallup's traditional likely voter assumptions, which determine respondents' likelihood to vote based on how they answer questions about their current voting intention and past voting behavior. According to this model, Obama's advantage over McCain is 50% to 46% in Oct. 9-11 tracking data.

The second likely voter estimate is a variation on the traditional model, but is only based on respondents' current voting intention. This model would take into account increased voter registration this year and possibly higher turnout among groups that are traditionally less likely to vote, such as young adults and racial minorities (Gallup will continue to monitor and report on turnout indicators by subgroup between now and the election). According to this second likely voter model, Obama has a 51% to 45% lead over McCain.

So, I guess it's up to us to decide which one is best. This puts us in a tricky spot - because the relationship between extra get out the vote (GOTV) efforts and extra votes on Election Day might be complicated.

In a 2002 article in the Journal of Politics, Charles Bullock, Keith Gaddie and Anders Ferrington investigated "voter falloff" in runoff primaries for the House of Representatives. Their interest was in what factors influence turnout in the second round of voting (which happens in a multicandidate field where nobody wins a majority of the vote). Unsurprisingly, they found that campaign spending is related to voter mobilization: the more dollars a candidate spends between the primary and the runoff, the better turnout the candidate enjoys at the ballot box. However, there's a twist.

They wrote,

The impact is nonlinear. If we assume $100,000 spent between the primary and the runoff, the net impact on the change in voter turnout is just 1.6 points; at $250,000 spent, the impact is an increase of 23.8 points; at $500,000, the impact is a net increase of 30.0 points, all other influences being constant. In a voting system that requires voters to turn out more than once, more campaign spending provides continuous stimulation, and apparently encourages participation, up to a point. With runoff spending averaging less than $100,000, it does little to spur turnout in a number of contests. Spending substantially affects turnout in the 26 runoffs in which more than $150,000 was spent. Diminishing returns from spending begin at about $950,000, and further spending is linked to falling rates of participation. [Emphases Mine]

This means that the relationship between spending and turnout might be a bit more complicated than some pundits have made it out to be. Of course, Bullock et al. looked at congressional runoffs, which are very different from presidential elections. So, we can't draw any inferences about the presidential election from this analysis.

However, this should induce some caution this year. The relationship between Obama's GOTV expenditures and his additional voters might be nonlinear, similar to what Bullock et al. find. That would be a situation in which some law of diminishing marginal returns conditions the relationship.

This makes some sense. If voting is positively related to social connectedness, money would have a decreasing marginal effect. After all, your first "$100k" will bring in people with greater social connections. They're probably paying more attention to political messages and maybe feel a greater social responsibility to vote. You'll get a good response from your GOTV efforts. But after those people come in, your next "$100k" will have to work on pulling people with fewer connections into the system. These people might be paying less attention, which means it will be more expensive to communicate with them, and they might feel a diminished sense of responsibility, which means that it might take more persuasion to get them to actually vote. It would therefore not be surprising that your second "$100k" pulls in fewer voters than your first. How many fewer depends on the precise nature of the law of diminishing marginal returns that governs the process.

I'm not saying that this relationship holds. I'm saying it might. If it does, you can't just look at how much money you're spending, you also have to know a thing or two about this law of diminishing marginal returns. This makes it difficult to estimate the effect of Obama's enhanced GOTV efforts. After all, those efforts are enhanced relative to Kerry's unprecedented efforts. So, that law of diminishing marginal returns, if it exists in this case, might be tamping down on the effect these extra resources have.

The operative word is "might." Contrary to what anybody might tell you, political outsiders can't answer this question - at least not right now. For all the discussion of Obama's GOTV efforts, it's all been about his campaign's inputs - the dollars spent, the organization created, the number of contacts made, and so on. There's no talk of what this is producing in terms of output. How could there be at this point? These contacted voters have not voted yet, so how can we know how efficacious this unprecedented effort will be?

This is where I find myself frustrated by Gallup's approach.

It is polling some 6,000 people per week. If the Obama campaign's unprecedented efforts were producing so many new voters that Gallup's old LV model will be rendered inoperable, we should begin to see some evidence of that in its data. After all, this is October. This would be the point at which Team Obama is really beginning to push these prospective voters into becoming actual voters. If its efforts ultimately prove successful - we should see begin to see that now.

In other words, the correct questions and a proper analysis, combined with a 6,000-person data set, should give us some insight into what kind of output we should expect from all this mobilization input. For instance, what about all those voters who are being excluded by the first LV screen but included by the second? Are they being contacted by the Obama campaign? If so, how frequently? In what way? What effect has this had on them? How has this influenced their thinking relative to voters who are not being contacted? With 6,000 respondents and a good empirical model, it should be possible to provide preliminary answers to these questions. That would give us some sense of which LV model is better.

Instead, Gallup has decided not to arbitrate between its models, leaving the question up to us. But I don't think we can answer it. We don't have the data to make a precise determination, and the relationship between mobilization efforts and new votes is too complicated to spitball.

-Jay Cost

Some Reflections on Polling in the Primaries

The polling has been bad this primary cycle. Last year's national polls were wrong - for both parties. The late polls in New Hampshire were wrong. Even in a state like Virginia, where McCain was supposed to win by a huge margin, he only won by a modest margin. What is more, pollsters have disagreed in state after state. One predicts California going to Romney. Another predicts it going to McCain.

What is going on?

There are surely many answers to the question. I'd like to suggest one I have not seen discussed in depth.

Let's start by taking note of an observation that political scientists have made since the 1950s. That is, average voters do not pay much attention to politics. This is a hard pill for political junkies to swallow, but swallow it we must. Indeed, I think most of the many inferential errors of inside-the-Beltway pundits can be chalked up to their false assumption that voters pay as much attention as they do. They need to get over this. It is just false.

We move from this starting point to a question: how do voters pick their candidate despite these low levels of information? Most researchers will tell you that they make use of cognitive heuristics - mental shortcuts that help them make decisions amidst uncertainty. Uncertainty is the consequence of inattention. Voters simply do not know that much about what is going on - but they nevertheless make a vote choice. Their mental shortcuts are tried-and-true ways of making good decisions despite this uncertainty.

Think of it this way. Political junkies might know the in's-and-out's of each candidate's health care proposals. They can thus make decisions about which they like best. However, the average voter does not have this kind of information. Yet, when he gets to the voting booth, he gets the same choice that the junkie does. The shortcut is his way through the uncertainty.

This leads us to another question - what serves as the shortcut? The answer for virtually the entire country is partisan identification. Upwards of 90% of the nation has some kind of party affiliation. This is despite the polls that identify the size of the independent vote as larger. It is not. There is, for sure, some portion of the country that tells the pollsters they are independent - but when we look carefully at them, we see that most of them lean to one party or the other.

Not only is party identification held by almost all of us - it is an incredibly precise predictor of vote choice. Republicans will almost always go 90/10 for the Republican candidate. Democrats will do the same for their candidate. Most of us have a party identification, and most of us rely on it quite heavily.

What does that mean in a general election? If partisanship is a near universal feature that is incredibly powerful, the preferences of most voters are anchored throughout the campaign - even though they are paying very little attention. They do not have to pay attention to know whom they will vote for. Accordingly, we will see the polls vary only a little bit throughout the campaign. Oftentimes, they will break in late October or even early November. However, the magnitude of the break will be relatively modest. This is not to say it will be inconsequential; just a few point swings in either direction could make a difference in many states. They might swing by +/- 10 points during the whole cycle - but this is paltry compared to some of the massive swings in this primary cycle.

A big reason for this stability is partisanship. As I said, it serves as an anchor. That is a good metaphor for it. Partisanship anchors preferences, keeping them from swaying, drifting, or listing wildly during the campaign.

In a primary campaign, voters must choose among candidates who are all of the same party. Partisanship therefore does not enter into their decisions. It is a non-factor. I think this might be inducing the wild swings in the polls. The polls are varying because the voters are; the voters are varying because their partisanship is not stabilizing their preferences.

Of course, primary voters tend to pay more attention to politics than general election voters. This probably makes them more able to make decisions without the use of their partisanship. Nevertheless, they still pay a price for not being able to use it. To say that primary voters are better informed than general election voters is not to say that they are well informed, or that they behave how the media implicitly assumes they do (i.e. carefully following every speech, parsing every sentence, keeping a constantly updated evaluation of the state of the race, etc). They do not do this.

It thus should be unsurprising that candidate personalities are so influential in voters' decision-making processes. How else do you make determinations when party distinctions are non-existent? Candidates often try to create clear contrasts, but these usually amount to making mountains out of molehills. The average voter is not really paying much attention, anyway. Thus, they have to go by their personal evaluations of the candidates.

This is not to say that vote choices are random. Clearly, there has been a regular pattern to the early contests. Certain types of voters obviously prefer certain types of candidates. The point is that without partisanship, personality is what makes the difference. And voters do not start to take careful note of personalities until late in the cycle. Consider that 48% of New Hampshire Democrats claimed to make their choice in the last week of the campaign. From a certain standpoint, that is incredible. If you think about all of the attention political junkies have paid to this race since last January - it is almost unbelievable to think that voters would not have decided months ago. But, if we put ourselves in the shoes of the average voters, and try to recreate their thought processes - it makes a lot of sense. Their partisanship cannot serve as a quick, easy guide. Thus, they have to take a good, long look at the candidates as people. Given their typical inattention to politics, the time when this happens is the last week or so.

This might explain the wide variability of the primary polling. Because they have not been anchored by partisanship - voter opinions have been unstable for most of the cycle, up until the very end when we are wont to see a massive break in one direction or another. The "error" in the polls might simply be a reflection of public indecision. For that matter, Clinton's massive lead through most of last year might have its origins here as well - without their partisanship, poll respondents had little to go on except their vague sense of the media's consensus view of the race. Predictably, they claimed to support Clinton. Finally, this might account for momentum. Voters take a close look at winners at precisely the moment they are basking in the glow of positive media coverage. Unsurprisingly, researchers have found that more informed voters are less susceptible to momentum effects.

These considerations have two implications. The first is good news for pollsters: life will get better for you! When we move into the general election - the polls will settle down and start agreeing with one another. I would note the stability in those head-to-head match-ups. They have barely budged an inch even as the race in both parties has been chaotic. Given that voters have their party identification to ground their general election responses to the pollsters - it makes sense they would be steady. The second is a warning to consumers of political news: continue to be wary of these primary polls. Without partisanship anchoring vote choices, they are still prone to large dramatic shifts at the last minute. That's what happens when voters try to make decisions without their use of most trusted cue, their partisanship.

-Jay Cost

On the ARG Poll

Anybody who checked Drudge today will have seen that there is a "shock poll" that puts Hillary Clinton 15 points in front of Barack Obama in Iowa. The polling company that produced the poll is ARG, and this is what it had to say about its results:

Hillary Clinton leads Barack Obama among women 38% to 21%, which is unchanged from a week ago (Clinton 36%, Obama 23% among women). Obama has lost ground among men to John Edwards and Clinton. Among men, Clinton is at 28%, Edwards is at 27%, Obama is at 16%, and Joe Biden is at 11%. A week ago, Obama was at 27% among men, followed by 21% for Clinton and 19% for Edwards.

This poll might indeed be a trend - the first sign of a swing back to Clinton among Iowa Democrats. Unfortunately, we will not be able to know for a few days - as polling companies presumably suspended operations over Christmas. I have a few caveats that I would put in place on this poll - just a few basic warnings about why we should not over-interpret these results.


1. ARG polled the weekend before Christmas, from 12/20 to 12/23. This might not be the best time to construct a sample of likely Iowa voters. No other poll I know of has come out with a sample taken from those days. This is a sign that most other pollsters were wary about Christmas weekend.

2. The ARG poll has Clinton up and Obama down by statistically significant amounts relatively to its last poll (12/16 to 12/20). On the Republican side, it has Mike Huckabee down and Ron Paul up by statistically significant amounts. This is a lot of movement - four candidates made statistically significant moves in the course of three days. Recall the last point, and note that these are three days when respondents probably were not thinking much about politics. December 20th to the 23rd are days usually filled-to-the-gills with last-minute holiday preparations. They are not great days for reflecting on the state of the presidential campaign. Thus, this movement might be due to the sampling effects mentioned in Point 1.

3. There are other elements of the poll that just don't scan with me. For instance, it shows Fred Thompson at 3% and Alan Keyes and Duncan Hunter both at 2%. ARG has shown Thompson low over the last few weeks - so this would not be a consequence of ARG's internal sampling method thrown off by the holiday weekend. But its last two samples estimated Thompson's support well below the rest of the Iowa polls. And 3% just does not pass the "smell test."

4. Mark Blumenthal has noted several interesting facts about ARG. First, they sample more heavily than any poll from first time Iowa caucus goers (on the Democratic side). This is probably why they usually have Edwards below where he is in the RCP Iowa Democratic average. Edwards is doing relatively well among previous caucus goers, but ARG is "diluting" their influence with the first-timers. Now, ARG's intuitions about first time caucus goers may be correct, but they are on the margins on this issue. Second, they did an extremely poor job of reporting their sampling methodology when Blumenthal requested it. They would not provide any information about respondent demographics, and they would not provide information about the number of long-time caucus goers in their Republican samples.

5. Just because differences between polls are statistically significant does not mean that they are necessarily caused by changes in the population. Clinton, Obama, Huckabee, and Paul have all made statistically significant moves - but some of these could still be statistical blips induced by the sample. This is as good a time as any to review exactly what statistical significance is.

The technical language that describes the margin of error usually reads something like this: "We are 95% confident that the true values are +/- 3%."

This is referring to Type I error, or the error of the false positive. It means that 95% of the time, when you take a poll and get 17%, the real world value will be between 14% and 20%. This also means that 5% of the time (or one time out of 20), it will be outside this range. This is the poll's tolerance of Type I error. The chances are 5% that you will have a false positive - you will believe that the real value is between 14 and 20 when in fact it is not.

But suppose you have 20 different statistics you are looking at. What are the chances that the real world value of at least one of them will be outside the margin of error simply due to sample effects? It is 64%!

This is something that is rarely noted when looking at poll trends - it is called the experiment-wise Type I error rate. When you look at the polls to divine trends, you are implicitly doing some form of statistical hypothesis testing. You are trying to determine whether changes are due to sampling error, or whether they are due to shifts in the population. To do this, you have to assume that sampling error will only explain so much variation in the polls. Usually (95% of the time), this assumption holds up. Occasionally (5% of the time), it does not. When it does not hold, you have committed Type I error. And the more polls you look at, the more likely it is that you have committed it.


Casual readers, please take note: I am not predicting that this is a blip. Contrary to what some have assumed, I do not make predictions about the ways the polls will move. That is a fool's errand. My point here is simply that it is possible that this movement is induced by sampling effects - and we should be careful not to over-interpret these results.

-Jay Cost

Interesting Internals in the ABC/WaPo Poll

A few days back, I saw that Clinton web video called "Caucusing Is Easy." You probably saw it, too. Anytime Bill does something, it gets noticed:

At the time, I thought it was a bit strange - insofar as it did not cohere with the conventional wisdom that Clinton and Edwards were the ones winning the support of veteran caucus voters, and Obama was winning over newer "voters." It seemed to me that this kind of video would be something to expect from Obama.

That last paragraph sports some important scare quotes - because newer voters often turn out not to be voters at all. Systematic survey evidence has picked this up - and it conforms with anecdotal accounts. Remember Nader's huge rallies in 2000? How about Dean's in 2004? This is why videos like this get made: young people are unreliable voters, and often need to be charmed into voting. Many pundits have speculated that the young supporters of Obama might be his Achilles' heel - as they are sufficiently motivated to come out and see him on a cold fall day, but not motivated to support him on caucus night. The former has some curiosity value. The latter? Not so much.

So, it surprised me that Clinton was the one creating a "hip" video about caucusing. Then I saw this in the Washington Post's write-up of its latest poll:

Overall, the poll points to some strategic gains for Obama. His support is up eight percentage points since July among voters 45 and older -- who accounted for two-thirds of Iowa caucus-goers in 2004. He also runs evenly with Clinton among women in Iowa, drawing 32 percent to her 31 percent, despite the fact that her campaign has built its effort around attracting female voters.

And despite widespread impressions that Obama is banking on unreliable first-time voters, Clinton depends on them heavily as well: About half of her supporters said they have never attended a caucus. Forty-three percent of Obama's backers and 24 percent of Edwards's would be first-time caucus-goers. Previous attendance is one of the strongest indicators of who will vote.

First off - WaPo does a good job of splashing some cold water on the statistical significance of the topline results. The margin of error on this poll is +/- 4%, so a 4 point lead for Obama is not statistically significant. However, statistical significance is conditioned by the number of observations. It has a lot in common with a simple computation of the standard deviation - which has the number of observations in the denominator. As the number of observations decreases, the standard deviation increases. It's the same basic premise for the margin of error.

So, when you are dealing with a subsample of the whole poll - say, voters 45 and older - the margin of error increases. I'd have to run a statistical test to confirm it (and the internals of the poll just do not provide the data to enable it) - but my intuition is that Obama's gains with older voters is not outside this increased margin of error. An 8 point difference between this poll and the July poll would be significant with the topline results, which are based on 500 observations - but probably not with a subsample of about 200 observations. WaPo was thus wrong to identify this as a "strategic gain" for Obama. It could very well be a sampling anomaly.

That being said - there are still some interesting inferences that we can make without running an undue risk of Type I error (i.e. wrongly concluding that something is significant when in fact it is not). The poll found Clinton and Obama relying on first time voters by about equal measure. The difference between their new supporters is not statistically significant - so the conventional wisdom about how Obama is relying on new voters more than Clinton does not hold. Accordingly, we have explained how Hillary's campaign could coax Bill onto a NordicTrack.

What probably does hold is the argument that Edwards is relying on new voters less than Clinton and Obama. The breakdown is basically 50/43/24. That is, (about) 50% of Clinton supporters are new, 43% of Obama supporters are new, and just 24% of Edwards supporters are new. Again, I would have to see more data than what ABC News/WaPo is providing - but my intuition is that the difference is statistically significant. Edwards is relying less on new voters.

This conforms with what Ana Marie Cox wrote last week in Time. Most of Edwards' supporters are reliable caucus goers. This might give us some clues about what to expect on caucus night. If attendance at the caucus is greater than what it has been in years past, that might bode well for Clinton and Obama. If it is equal to or less than what it has been, that might bode well for Edwards.

Another point on Clinton v. Obama. Clinton is splitting the female respondents evenly with Obama. She pulled in 31% of female respondents. He pulled in 32%. It is surprising to me that Clinton is not pulling in more females in this poll, given the tone of her campaign of late. These results make Mark Penn's promise to pull in Republican women seem like rhetoric designed to win over Democratic voters hungry for a victory.

For comparative purposes, I would note that the recent CBS News poll also found no statistically significant difference between Clinton and Obama among first time and long time caucus attendees. It also seems not to have found a statistically significant difference between them on levels of female support (although Clinton does lead this category by 12%, it is such a small sample that I suspect the lead is statistically insignificant), and levels of support among voters aged 45-65. However, it did find what appears to be a statistically significant difference on levels of support among voters 65 and older, though there is a tie between Clinton and Edwards among voters of that age. [Again - I'm "spitballing" these conclusions of significance because none of these polls ever give you the data you need to draw more assured conclusions.] So, all in all, I would say that the results of the ABC/WaPo poll roughly conform with the results of the CBS poll.

Now - it is important to remember that there are boundaries that we have to obey when drawing inferences from these Iowa polls. I wrote about this last week. Care is important because the Iowa Democratic caucus is a poor fit with the way polls are conducted. I am pretty sure that I have managed to color within these lines in this write-up - but we need to be careful.

-Jay Cost

What Moves the Polls?

Mark Mellman had a very excellent column in the Hill yesterday. The topic involved the validity of horse race polling.

This is what he had to say:

Regular readers have heard me rail against the inaccuracy of early polling, which often fails to presage ultimate electoral outcomes.

Yet I have also maintained that presidential elections are predictable based on the fundamentals -- incumbency, war/peace, prosperity and the like.

Recognizing the implicit contradiction, political scientists Gary King and Andrew Gelman asked, in one of academia's best-titled papers, "Why Are American Presidential Election Campaign Polls So Variable When Voters Are So Predictable?"

Their answer focuses on learning during the campaign, and while they may be right, I fear the problem runs deeper. [Snip]

Though it is heresy for a pollster to say it, the evidence also suggests people are only mediocre predictors of their own behavior. Responses to horserace questions a year out may be a special case of faulty prediction.

Hardly an original thought, it can be traced at least to Russian exile and founder of Harvard's sociology department Pitirim Sorokin, who titled his 1936 paper, "Can One Predict His Own Behavior 24 Hours In Advance?" His answer, based on a study of federal employees, was a resounding no: When asked how much time they would devote to various activities during the subsequent eight-hour workday, the average person was off by five hours.

My hat is off to Mellman - who actually relies upon the work of one of the best political scientists in the country, Gary King, to make an argument about politics. A very rare thing indeed! Usually, political scientists get no more "airtime" than the occasional self-evident quote that journalists integrate into their preconceived storylines.

Mellman's topic - the invalidity of early election polling - is one that I have discussed frequently on this blog. I'd like to extend this conversation because I think the problem with media polls gets to what I think are some serious failings in the way journalists and pundits analyze politics.

The best way to discuss this is simply to review the Gelman and King article - which, I should note at the outset, is an attempt to explain a problem with general election presidential polling. Their title indicates the question with which they tussle. Political scientists have developed models that do a very good job of predicting presidential elections based upon "fundamental" variables like incumbency, partisanship, and the state of the economy. All of these are available a long time before ballots are cast. Meanwhile, the polls run all over the place prior to Election Day. How to explain this?

Gelman and King offer what they call the "Enlightened Preference" Model. They assert that:
(1) Voters do not have full information throughout the campaign about the "fundamental variables" that ultimately drive vote choices.
(2) Voters do use all available information to make their decisions.
(3) Voters do not rationally account for uncertainty during the course of the campaign.

This explains how polls can vary so wildly, and yet final results can be so predictable. Voters base their election decisions on basic variables. Thus, their vote choices are quite predictable. But it is only at the end of the campaign that they have fully grasped the values of the variables. Additionally, they do not factor this lack of knowledge into their thought processes. And so, when pollsters dial them up - they rely on the data they have available, but give answers that are less certain than they realize.

Gelman and King write:

[W]ithout sufficient knowledge of their fundamental variables, and when asked to give an opinion anyway, most respondents act as they will in the voting booth on election day: they use the information at their disposal about their fundamental variables, and report a "likely" vote to the pollster. We believe that this report is sincere, but the survey response is still based on a different information set from that which will be available by the time of the election.

Note that this does not mean that the campaigns are useless. The campaign organizations are the agents that provide the information to the voters. If there is a rough organizational "balance" between the campaigns - then the relevant information will be communicated to the voters through them. The media can play a role here - by providing relevant information to the public so that their choices are as informed as possible. Of course, in the last several cycles, most of the media's work has been in covering the horse race.

Now, as I indicated, I think there is a lesson in this for all of us who produce and consume political news and analysis. Of course, bear in mind that this is not a certification of the Gelman/King theory. My specialties in political science are party and campaign organizations. It is not in public opinion or political psychology. I lack the credentials to certify this theory. My intuition is that they are pretty close to the truth - but I am not up on the latest scholarly research to tell you with certainty.

However, I do know enough about these subjects to know that popular accounts of voter psychology are seriously askew because they falsely assume both too much and too little of the average voter. Gelman and King summarize what they take to be the "journalistic model" of the political campaign. I think there is a lot of validity to the characterization:

Under this model, voters base their intended votes partly on fundamental variables, but considerably more on the day-to-day events of the presidential campaign. Voters are assumed to have very short memories, relying for their decisions disproportionately on the most recent campaign events and last piece of information they ran across. Candidates are thought to be able to easily "fool" voters by changing their policy stance during the campaign or causing the opposing candidate to say or do something foolish. [Snip]

Also according to the journalists' model, voters do not take their role in the process very seriously, have very little information of the campaign and the issues, and frequently do not vote on the basis of their own self-interest.

This, Gelman and King argue, is how journalists implicitly explain the day-to-day movement of the polls. I think that this is a valid explanation of the way journalists and pundits approach politics. And, like Gelman and King, I think it is completely wrong-headed.

I would note three salient features about this false view of the voters:

(1) It assumes too much of them. Implicit in this theory is the idea that average voters pay as much attention to politics as political junkies do. We saw a great example of this early this month when pundits and journalists started talking about how the last Democratic debate "changed" everything. Give me a break! 2.5 million people watched that debate. It didn't change a thing! To think that it did requires one to accept a false premise: that average voters follow politics like the junkies.

(2) It assumes too little of them. It assumes that they can be beguiled - that, for instance, they voted for Bush over Dukakis because the latter put a dumb-looking helmet on his head. Again - give me a break! There is an implicitly condescending attitude toward citizens in media theories about vote choice.

(3) It assumes that average voters are quite like journalists. They love the gamemanship of politics. The twists and turns of the daily political soap opera is what they find to be valuable. They care, for instance, that Hillary Clinton's latest wardrobe choice has a lower neckline. And they really care about the political strategy that influenced the decision.

If these three assumptions are false - then most of the media's horse race coverage is a skewed version of what is really going on. I have been arguing this point for quite a while on this blog. What the media assumes about the voters is simply not true - and therefore its analysis is based upon false premises, and thus is wrong-headed. To appreciate this - try the following experiment: take any standard issue pundit discussion that you see on the news, ask yourself whether they are falsely assuming these three things about the voters, and then ask yourself whether - if they stopped making these false assumptions - the analysis would be different.

What, then, drives general election poll numbers? Gelman and King argue that one of the things that is surely driving it is the partial information that has been collected to date. That is, voters are in the process of collecting the data they need to make an informed choice in November. When queried in, say, July - they have only acquired some of that data, which is what they use to make their selection.

I would argue that there is something else going on in these survey responses. Sitting in the background here is John Zaller, whose theory on public opinion merges very nicely with the theory on vote choices that Gelman and King proffer. His 1992 book, The Nature and Origins of Mass Opinions, is required reading on the graduate level - and I'd wager that a lot of undergrads are forced (much to their chagrin!) to slog through what is a very technical read. Zaller argues that one of the reasons public opinion varies is that respondents "receive" informational tidbits here and there from the dialogue of political elites. If those tidbits are essentially compatible with preconceived notions, they are "accepted" and mentally stored by the respondent - to be "sampled" from when the pollster comes calling.

Receive, accept, sample: the "RAS" Model of public opinion. So, in an ironic twist - the analyzers of public opinion actually are the ones creating some of it. It is not that the last Democratic debate had an independent effect on the polls (if there indeed was an effect at all); the effect was caused by the fact that analysts predicted that it would have an effect. Mass opinion is influenced by the elite conversation - so when elites talk about how an event shaped public opinion, they are in fact helping to shape public opinion in that way. But this kind of self-fulfilling prophesying is just a "game" that is quite separate from the way in which vote choices are formed. Public opinion can be tweaked by the elite dialogue - but insofar as this dialogue is not providing information on critical variables, it is not influencing vote choices. It's just shifting the numbers temporarily.

Now, let me reiterate that my specialty is not political psychology or public opinion. I have read and have understood the "great works" in the field, but I am not up-to-date on the current literature. And, as I said, Gelman and King's model is meant for general presidential elections. Nevertheless, I think that this explanation is probably not too far afield from reality. At the very least, I am certain that there is a disconnection between the way average voters really are, and the way journalists and pundits see them - and that this disconnection induces many, many errors in the way the latter examine politics. These errors ultimately make the media dialogue irrelevant for what really matters: who wins the election.

Or, to quote Gelman and King:

"Journalists should realize that they can report the polls all they want, and continue to make incorrect causal inferences about them, but they are not helping to predict or even influence the election. Journalists play a critical role in enabling voters to make decisions based upon the equivalent of explicitly enlightened preferences. Unfortunately, by focusing more on the polls and meaningless campaign events, the media are spending more and more time on "news" that has less and less of an effect."

-Jay Cost