Outlook: Five myths about midterm elections

Aug 16, 2010

Alan Abramowitz, professor of political science at Emory University, will be online Monday, Aug. 16, at 11 a.m. ET to discuss his Outlook article titled "Five myths about midterm elections."

Hi, Alan Abramowitz here to discuss my Outlook article from yesterday on midterm elections.

Republicans are crowing that they lead in the generic congressional ballot preference by 4-6 percent and also claim it actually overstates Democratic strength. How well does this metric predict congressional election outcomes? It seems to have been accurate in '06 and '08. If the current trend doesn't change, should Democrats plan on being in the minority in the House next year?

The generic ballot has proven to be a fairly accurate predictor of the national outcome of the House elections.  Less so for Senate elections of course.  The recent results have ranged from a small Democratic lead to a small Republican lead, averaging close to a tie.  That would predict significant GOP gains in November.  But a better prediction is possible by using other factors along with the generic ballot such as seats held going into the election, presidential approval, and just the fact that it's a Democratic midterm.  In general, all of these point to large Republican gains--possibly large enough to regain control of the House, although probably not the Senate where 10 seats is a big hill to climb.

To what extent is success in the midterms about winning independents, centrists, or what have you, and to what extent is it about turning out the people who are already on your side?

Some of both.  Turnout is important and Republicans are typically more reliable voters.  Democrats need to get more of their base out to hold down their losses. 

"But even when voters seem very unhappy, the vast majority of incumbents in both parties are reelected. Despite Congress' low approval ratings this year, only a handful of incumbents have lost their primaries, and there were peculiar reasons for several of those defeats." Left out of your analysis is how many congressional districts have been gerrymandered, in occasionally preposterous fashions, to eliminate any hint of actual political competition. We routinely hear about "safe districts" in political wonkery; there should never be such a thing. In some districts, you could run a murderer, child molester or dead crab, and s/he would win as long as it has the proper letter after its name on the ballot. It would be nice to think that the "Tea Party" and other anti-incumbency movements could move past such thinking this time around, but history shows the odds are against them no matter how many congress-persons are under ethics investigations, indictments, or criminal proceedings. I mean, we're in a city that elected a convicted felon back to the mayor's office!

There are a lot of safe House districts but that's mostly due to political geography not gerrymandering.  Most large urban areas are heavily Democratic.  Many parts of the South and Mountain West are heavily Republican.  Even states have become less competitive overall.

What are the current Gallup congressional poll numbers and what do they tell us about the midterms"?

Over the past four weeks they have ranged from a 5 point Democratic lead to a 6 point Republican lead.  Longer term, the average has been close to a tie.  That's good news for Republicans though because their supporters are more likely vote. 

In Illinois, we've managed to find public housing for most of our previous governors (we'll see today about Blago). Despite record high convictions, it seems as though the powers that be manage to remain in power. We don't throw out the villains, we keep incumbents in power, and we keep voting for those approved by the party leaders. What could it take to actually push some of the incumbents out? War? We've occupied Afghanistan far longer than Vietnam. The Economy? Jimmy Carter's recession was deeper, tougher, and coupled with gas-free weekdays. Corruption? The House Post Office scandal touched a few congresscritters, but most skated. The question is this: Is there any current national issue that you believe will have a major adverse impact on this election, for either party?

The economy is going to be the big issue this year and Democrats, as the party in power, are going to be hurt by voter discontent with economic conditions.  It may not be fair, but that's the way it works.  The incumbents who lose in November will be almost exclusively Democrats--just as the ones who lost in 2006 were exclusively Republicans.

What's your take on the mid-term impact of the Tea Pary? Helpful to GOP in raising anger/enthusiasm; helpful to Dems in giving victory to fringe GOP primary candidates; or maybe both? Also seems this could be part of 2010 narrative that is ultimately helpful to Obama in 2012, making nomination of more centrist GOP pres candidate less likely.

I think it both helps and hurts the Republican Party.  Clearly having a fired-up base is beneficial.  But some of the Tea Party candidates nominated in Republican primaries could cost the GOP winnable seats.  The Nevada Senate seat is a prime example of that.  There are several others as well in the Senate and House.  As for 2012, certainly we can expect to see the Republicans nominate a strongly conservative candidate--what I like to call the "anti-Obama."  The outcome will depend of course on the condition of the economy and President Obama's standing with the electorate two years from now--which can't be predicted at this time.  But if the GOP nominee is perceived as too far to the right that could certainly hurt the Party's chances of retaking the White House. 

I'd like to hear your thoughts about how different scholars conceptualize polarization in the mass electorate--and where you come down on thinking about it. It seems to me that there is consensus out there about elite polarization, but how does that work causally with what we see in the mass electorate--is it elite polarization-->mass polarization or vice versa? Aren't there still a bunch of moderate independents out there keeping us sane?

You are correct.  There is consensus about polarization among elites.  The controversy involves the degree of polarization among the electorate.  My own research indicates that polarization has increased considerably over the past few decades among the public as well as among elites--and these two trends are related of course.  Moreover, among the public the most politically engaged (intersted, informed, and active) citizens are the most polarized and their opinions are weighed most heavily by candidates and elected officials. 

Good to have some perspective on midterms! I was wondering, professor - how did you and Mr. Ornstein come to collaborate on this article? And what was your process for working together?

I've known Norm for a long time and since he's a Congress specialist and I'm an elections specialist it was a natural collaboration.  We just went back and forth several times over the course of about a week before we got to the final version.  It was a pleasure working with Norm.

News organizations, politicians and political pundits are mesmerized with political polls, especially for contested elections. Reporters and pundits delight in reporting a tight horse race or abruptly changing poll results. However, political polls cannot predict future events reliably because they don't meet the prerequisites for valid statistical inference. For reliable error margins, statistical inference theory requires stable populations or sampling environments from which multiple samples are drawn over time. Also respondents must be selected randomly with each possible opinion in the population equally available to be sampled. It's doubtful these requirements exist in today's political polling. Political polling's dirty secrets would be revealed if poll designs and error margin calculation methods were published. Yet, news organizations, political parties and politicians spend millions on political polling when voter turn-out often is the decisive factor (which can be predicted spending far less). Other than producing dazzling graphics, is it a case of the blind following the blind or is there nothing better to do with political contributions and news budgets?

We have seen an amazing proliferation of pubic polls over the past few election cycles.  Some of this is a result of the use of new technology, especially the use of robo-dialed telephone polls by organizations like Rasmussen and Survey USA.  These polls are relatively cheap but they do cut a lot of corners.  Rasmussen in particular does his polls in one night with no callbacks and no calls to cell phones even though cell phone only households  now account for a large proportion of the electorate, especially among younger voters and minorities.  So there is really no way to calculate a margin of error for these polls.  But that doesn't reduce the level of curiosity about their results.  As long as media outlets think their viewers and readers want to know about the horse race, we're going to see more and more dubious polls.

What would your forecast be of Republican gains in the Congress and the governors races? Also; do you think the governors races are being somewhat overlooked?

Yes, there has been less attention to the governors races but there are 37 seats up this year.  I expect to see significant GOP gains there as well, probably in the neighborhood of 5-7 seats which would give them 28 or 29 of the 50 states--although Democrats might still control most of the big states.  Right now I'd say GOP gains likely to be in the 35-45 seat range in the House, 6-8 in the Senate.  That could change in the next few weeks although I don't expect to see any big changes.

Prof. Abramowitz, thanks for taking questions today. We keep hearing that this administration has actually accomplished a great deal legislatively during the first half of Obama's term. Why isn't he -- and by extension the Democratic Party -- getting more credit for that? With the Republicans tied or even leading in the generic matchups, that doesn't seem to be reflected. Do voters have any expectations for what they hope will happen if the Republicans retake the House and perhaps the Senate, or don't they think that far ahead?

Great questions.  I think the voters are generally in a sour mood because of the condition of the economy.  It's difficult for Dem0ocrats to get credit for their legislative accomplishments, which have been considerable, when people are hurting due to high unemployment and falling incomes.  Voting Republican is mainly a way of expressing that general feeling of discontent rather than an endorsement of specific GOP policies.  Most voters would probably want to see Republicans work cooperatively with the White House after the election if they get control of one or both chambers, but that's not likely to happen--the ideological divide between the parties is just too wide.

How many voters who say they are independent really independent? My wife insists on calling herself independent but would never vote for anyone in the GOP. So how many true "fence sitters" are there?

You are correct.  Most independents are really "closet partisans" according to the survey data.  In 2008, for example, about 40% of eligible voters called themselves independents, but a lot of them didnt' vote and among those who did only 7% were true independents with no party preference.  Most independents are leaners--they usually favor one party and their beliefs and behavior are almost indistinguishable from those of other partisans.

So is there an answer? It's my belief that the more the country "balkanizes" into "safe" districts where there's no chance of a candidate with the wrong letter after the name being elected, the less the government will be held accountable or expect to be held accountable, which in the long run will increase the polarization even more. The country needs its politicians to live in fear of constituents being ready to hoist their heads on a pitchfork (metaphorically, not literally--though a little fear is always good), but the more we polarize, the less likely this becomes.

Actually polarization could increase accountability--voters know what the parties stand for now.  The problem is that our institutions don't match up with our polarized parties--especially the Senate. 

Isn't voter turnout often the decisive factor in elections so if you can predict turnout you can determine what is likely to occur more than polls 4-5 months out from the election? But to publicize poll results as if they are representative of what voters are likely to do “if elections were held today misrepresents what polls actually are -- the views of the 300-1000 individuals who happen to respond to questions many of which are unartful. On the other hand, how people vote behind curtains may be quite different than how they respond to pollsters. Otherwise, likely there would have been a California Governor Tom Bradley.

Despite their flaws, the polls are generally pretty accurate.  Especially the ones done near the end of the campaign.  In 2006 and 2008, the polls got almost every state race right, both for president and for senate seats.  So while I wouldn't pay too much attention to early polls, when we get to October I'd say the results will be a pretty good indication of how these races are going.  But I don't place too much weight on any one poll.  I look at the average of all of the reputable polls--excluding blatatly partisan ones.

Will this election really be a referendum on the Ground Zero Mosque? Is it fair to say that Sharia law is on the ballot now?

No and no.

Will the Citizens Unitedruling give either party a boost this year? if so; who benefits more?

That's a great question but we won't really know how that ruling will play out in campaigns for a while, maybe not for another election cycle.  The conventional wisdom is that this should help Republicans by allowing more corporate money into elections but it remains to be seen how much money corporations are willing to spend in support of political candidates.  My guess is that a lot of them won't want to be seen as taking sides so overtly--it could alienate customers as well as leaders of the other party. 

Do you see any U.S. House incumbents in Georgia vulnerable this year? Republicans are making noises about taking Jim Marshall's Macon-centered district (which backed McCain in 2008) and even the 47 percent black southwest Georgia 2nd District of Sanford Bishop

Right now I don't see Marshall or Bishop, or John Barrow, as vulnerable.  And those are the only potentially competitive districts in Georgia.  Marshall and Barrow have had close calls recently but their challengers this year don't look too threatening.  None of the Dem challengers have a chance. 

Why do all the articles about how incumbents hold their jobs ignore one of the basic facts? With divergence of political positions of the two parties, if you have a moderate to liberal Democratic congressperson, unless the incumbent loses the primary, you would have to radically shift what you believe in (assuming you were a Democrat) to vote for his opponent. For me, that is a far, far bigger factor than any other factor. I cannot imagine being upset enough with my Democratic congressman (who I like, but that is beside the point) to vote for the conservative Republican challenger. I assume the same is true for Republicans, so why isn't that factor considered the most important?

That is exactly right.  We're seeing a very high level of partisan voting in all types of elections from the presidency on down because the differences betweent he parties are so large.  That's why most House districts and many states are really safe for one party or the other.  Only a minority have a close enough partisan balance that they could switch sides from one election to the next.  But there are more than enough to switch control of the House or Senate. 

What's astonishing to me is that since these people get re-elected 98 percent of the time -- they are STILL so freaked out and afraid of losing their seats. Why is that? Common sense would say that they could basically do what they want (and they do - with regards to ethics violations) and still get re-elected. So one would hope that they would have a little more leadership to vote how they think they should vote, rather than voting how they think would help them get re-elected...it's absurdity at its best.

The ones from safe districts or states tend to vote along party lines very consistently.  Most of the defections come from the minority who represent districts or states that are evenly divided or lean towards the opposing party.  They're the ones who are vulnerable in a tough year.

Good morning, Dr. Abramowitz. Former student here (William and Mary, 1976). Democrats will be hammering away on the Republicans' lack of an agenda from now till November, and Republicans seem keen to oblige them by saying "no" to what Democrats want and little else. That might not be hurting the Republicans now, but is there a chance that it could hurt them (particularly with independents) by November?

Good to hear from you.  1976 was my first year out of grad school and my first year teaching.  I don't think the lack of an agenda is going to hurt the GOP very much.  A midterm election is mainly a referendum on the president's party--voters are expressing their satisfaction or dissatisfaction with their performance, not necessarily endorsing the opposing party's agenda.  In fact it's almost always a mistake for the opposition party to interpret the results as a mandate for their agenda--as we discuss in Myth #5. 

I love your articles on Larry Sabato's weekly newsletters; do you have a newsletter of your own? a Web site?

Thanks!  No, but I contribute to Sabato's Crystal Ball on a regular basis.  It allows me to play at being a pundit as well as an academic without the effort of publishing my own blog.  And it's a great source of political intelligence.

Care to mention a few blatatly partisan polls?

Ones that are sponsored by a candidate should be taken with a large grain of salt.  Some polls sponsored by parties or partisan organizations, like Democracy Corps, have solid reputations.  Right now I'm a little skeptical about some of the robo-dialed polls that are not including cell phones in their sampling frames.  Also, if they don't reveal how they screen for likely voters, I'd be suspicious.  The LV screen is really crucial in a midterm election.

Seriously - no one should be in Congress more than 10 years, if that.T he point is we shouldn't keep sending the same people back to Congress. We should continually be getting new people and fresh ideas. Otherwise, we get what we have now, which is congresspeople who have no idea how the rest of the country lives -- they keep passing more laws, which makes it more difficult to keep up and more difficult to exist without a lawyer and accountant, etc.,  on your staff. Which most of us can't afford...

I don't think term limits are the answer.  The experience with them in the states has not been very good.  If voters want to kick out an incumbent, there's nothing to stop them from doing so either in a primary or a general election. 

I am interested in your idea of the disappearing center. I am particulary interested in the generational split that I see, and if I am viewing this correctly. The younger voters tend to be more liberal, in fact the appear by far more "pro-government response to solving problems" in polls than other age groups. The older voters appear to be more anti-tax, less government. Thus, I see politics is often a battle between the more conservative older voters whose numbers are dwindling but they vote in higher percentages of their group than other age groups, versus more liberal younger voters whose numbers are increasing but they vote in smaller percentages of their group than do other age groups. How close or off are these observations?

Yes, there is a generational divide on these sorts of questions as well as on some social issues like gay marriage.  Part of the generational divide is due to the shifting racial composition of the electorate.  Younger voters include a much higher proportion of nonwhites.  In a midterm election, though, younger voters don't turn out as much as older voters--their share of the electorate will be down from 2008 and that's going to hurt Democrats.

There are politicians within the GOP (Sarah Palin, for example) who seem to celebrate ignorance or perhaps she and others could be said to be defaming the educated. Does this kind of politics make the educated GOP uncomfortable (people like Mitch Daniels comes to mind) or are they happy just to use that as a way to get votes?

I think there is something of a divide within the GOP between traditional economic conservatives and social conservatives, but the party overall is pretty unified this year.  They can agree that they don't like what Obama and the Democratic Congress are doing.  We may see internal GOP divisions open up more in 2012.

Do any of these have a better track record than the others?

Well I'm partial to Sabato--I think he has the best track record in recent elections.  But they're all pretty good and they're all pretty much in agreement right now about 2010.

In This Chat
Alan Abramowitz
Alan Abramowitz is the Alben W. Barkley professor of political science at Emory University and the author of "The Disappearing Center: Engaged Citizens, Polarization and American Democracy."
Recent Chats
  • Next: